A few weeks ago, I wrote an article for OSNews entitled “Update on Red Hat’s Limbo Progress.” It was to be a short article on how much Red Hat’s beta releases have impressed me – to share with everyone some of the changes a desktop user sees and maybe generate some additional interest in my choice, Linux. Little did I know, one of my comments nearly incited a riot- it would flood my Inbox, leave me feeling silly about something that I still think is true…it was just poorly stated.
So, let’s try a little experiment.Let’s all agree to step back, forget what we know, and pretend we’re seeing Linux for the first time. Let’s pretend we don’t know too much about anything besides Windows 98. Let’s visit a world where the computer is an intimidating machine that requires a highly skilled technician for basic tweaking. You with me? Okay, here we go…
We just booted up for the first time and logged in, so far so good. The desktop is pretty comfortable; a standard Windows user will be able, in virtually every distribution that configures X and defaults to it on login, to comprehend what they are seeing. Some of the recent changes, for the better, I’d add, have stripped the confusing names, like xSane, Toaster, Evolution, Mozilla and terms like Halt from their default menus and replaced them with much more logical and self-explanatory names like “Scanner”, “Burn CDs,” “E-mail,” and “Browse the Internet.” In fact, a very basic user, who is likely using Hotmail or Yahoo Mail or some other web-based e-mail could probably navigate around and get almost all of their day to day work done just fine.
Let’s take our user a little further. Let’s say she becomes comfortable and gets used to everyday apps like OpenOffice.org, grip, and XMMS. Then one day, our fairly standard user decides she needs a new application, maybe Mr Project, which we’ll just agree didn’t come with the distribution she has installed. Let’s even assume she talks to her buddy who tells her that the app is probably on freshmeat.net, a great repository for Linux apps. She finds it there easily enough and downloads it, and now she has an RPM in her home directory. She associates her home directory to her “My Documents” folder, maybe she knows it’s more like Windows 2000 than 98 in a sense. She double clicks it and nothing happens. A minute later, it asks her with what app she would like to open this mysterious RPM. A google search, or maybe, if she’s particularly smart, a newsgroup search, later and she has her answer. In this case, she must visit the terminal and type “rpm -Uvh packagename-1.0.i386.rpm“. She does it. It fails. She doesn’t have rights. After some more research, she finds out she must type “su” at the command line first, enter the root password, and then repeat the command. Once again, it fails, she’s missing some dependencies. She finds a similar package on sourceforge.net and downloads it (from her local mirror, of course). Now she has a compressed archive file, a “tarball,” she’s told it’s called, in her home directory. But even after she figures out how to uncompress it and untar it, she doesn’t know how to install it. Thank God someone had the forethought to tell the user in the README file to visit the command line, su to root, and run ./configure with the necessary flags. “Where should I install this?” she wonders? The README says, “for SuSE, here, for Red Hat, here, for Mandrake, here….” Frustrated, our persistent user spends some time looking for a solution. Since she’s not ready to download Debian just for the .deb files she’s told resolve dependencies, she uncovers the solution: URPMI…or APT4RPM …or GnoRPM. Let’s hope she doesn’t find out that one of them is deprecated, since she has no idea what that means. But wait! Now she’s found information about IRIS, Click n’ Run, Red Hat Network, and Red Carpet. “What am I supposed to do?” she asks herself. Her solution, despite her noble effort, is to boot back into Windows 98 where nice, comfortable .exe files are waiting and the answers are any old help desk away.
People don’t hate Linux, but the community’s arrogance when it comes to addressing legitimate user concerns are continually self-defeating. The acronym “RTFM” is a great example. Why do people even know what that means? Is it really that common that we tell someone to “Read the F*cking Manual!” that we need to abbreviate it?! It’s not that Linux users are bad people, it’s that there is that much about Linux that is not fairly obvious to a new user. Most flavors are getting a lot better at addressing these concerns. Most people are too. I understand there is a learning curve when learning a new Operating System, but shouldn’t we ask ourselves if we, as a community, are doing things the best way anyway?
There IS interest in Linux, but the simple things kill it quickly. Linux’s biggest weak spot, in my opinion, is software installation. Although there are numerous solutions, here are my thoughts, from scratch – on the ideal way to install software. Installing software over the internet cannot work all the time – it relies on your sources having the software as soon as you want it, often NOT the case. Therefore, I could propose something like the following:
First off, the definitions must come from a centralized committee, perhaps the LSB. A new format is declared, perhaps *.lif, for Linux Installation File. It needs to have an extension to indicate, fairly obnoxiously, “I AM THE INSTALL FILE FOR THIS APPLICATION.” The file, when double clicked, or run from the shell with the scary ./ prefix, invoke a menu that says “You have chosen to install new software. This must be done as the “Root” user with administrative permissions. Please enter the password in the box below” or something like that.
Next comes the fun part: we must either agree that all software is installed in the same place in each distribution, or else, we invoke a second menu, this one unique to your distribution, which recommends a place to install the software. Perhaps it says “You are trying to install a new program. Distribution X recommends new software be placed in the ‘/opt’ directory.” Of course, you could install it anywhere, but the menu should tell you where it thinks it should go. The system files would go where they should, just as a Windows application puts files in C:\Windows, but most of the software would be in one place.
Incidentally, acknowledging that some people prefer to compile their applications from source, this would most definitely be possible, these suggestions are purely supplementary.
Uninstalling software must be easy as well. Perhaps an option in the Gnome or KDE launch menu that says “Uninstall software.” Here, a list of all software is kept and the necessary way to uninstall it is tracked. I’m not the most technical person out there, but I would think GConf, a registry-like control for Gnome, would be a benefit to this. A user must have the option to uninstall, sometimes simply because they don’t like the clutter of too many applications, which, ironically, is another problem (too many apps by default – too many that do the same thing.)
Lastly, continuing in my “I’m not really qualified to make these sweeping suggestions but do anyway” piece, there must be an easy way to make shortcuts to these new apps. I understand the complexity in saying that I should be able to drag an item to the desktop holding Shift or Alt or Ctrl and get a shortcut, but…well…I should. If Microsoft can pull it off in Windows, why can’t we? Many people are vocal about the limitations of X Windows. I had the honor of exchanging some e-mails with a prominent player in the Linux world a week or two ago, and he explained that the non-multi-threadedness of X Windows was a killer that continued to plague Linux. With due respect, our desktop user, trying to make the switch doesn’t care, and frankly, the term multi-threaded just confirmed her doubt. Worse, she’ll probably tell someone she tried Linux and came back, and they’ll think, “Jeez…and she’s much more technical than I am! I guess I’d never be able to figure it out.”
APT and RPM, for those of you still with me, are not competing technologies, they’re not even in the same category. But at their basest level, they can partially manage packages, and when distributions ship with either their own proprietary update agent or in a format that is confusing and, for lack of a better term, “non-clickable,” it’s time to sit down and decide if we should cut our losses and go back to the drawing board.
At the beginning of this piece, I should have explained that everything I’m saying is predicated on the fact that we want new, less technical users in the Linux community and that we want Linux on the desktop – even if it’s just the corporate desktop. I believe that with numbers come commercial vendors, more commercial applications, and in time, a better, more complete community. And in the end, isn’t that what we all want?
About the Author
Adam Scheinberg is a Network Administrator for the Naval Sea Systems Command (NAVSEA). He uses NetWare 5.1 and Windows NT/2000 at work and is currently running Red Hat Limbo II, Windows 2000, and Windows .net Server RC1 at home.
Why should an installation file have a .lif ending? Oh… linux installation file… so obvious. Again, obvious if you know what it means… just like a lot of things in linux. So the author, despite attempts to make the system opaque and practical, still manages to drop something in there by expecting users to view/know file name extensions.
Hmm…
In the very beginning, Linux lacks hardware support, then the 2.4 kernel was.
Then, Linux lacks Desktop Managers, then KDE3 and Gnome2 were.
Then Linux lacks a complete browser. Mozilla was.
Then, Linux lacks Productivity Software. Open Office, Koffice (and soon Gobe Productive) were.
Now the problem is programs installation? Red Carpet and RpmDrake are beginning of answers(not to mention apt-get but it seems that CLI is not well suited for the average user).
It’s just another level of the scale and some people here on Earth are working hard to make these things possible.
So, what’s next? Games? In a few months, Doom III will be available for Linux… 😉
I’ve read a few comments about Gentoo saying that it’s too hard to use, and that you can’t compile from source on a desktop os … Why?? what does the average user care if it’s complied from source or not ?? so long as it installs what does it matter … in fact it’ll run faster if it’s compiled and not a pre-compiled 386 binary … so if someone wrote a gui for Portage that lets you browse the ports tree, sorted by app type, and shows you what you can have (maybe with a link to the home page of the project) and what you’ve got (including versions) and a button to install and uninstall it’s perfect … Personaly I’m happy with typing emerge … it’s a beautiful system and much easier to use than Windows, no clicking, no reading licences, no selling your soul …
it’s a GUI frontend to portage, works with KDE.
I’m a newbie to linux, but i’m a windows advanced user, and I chose Gentoo for portage and the mandatory learning that comes with Gentoo
Just like I said in a previous comment, I’ll stick with Gentoo because of the community.
For me, the problem about install/uninstall is settled: kportagemaster does it. I did not yet meet a dependency problem. Hope it wont happen!
For Joe User, Lycoris should do the trick with their “IRIS” stuff.
Brazilian Conectiva Linux distribution ported the Debian apt to works with rpm packages. To install a program “prog” you just have to type
# apt-get install prog
and all dependencies are also downloaded and installed. The only problem with Linux is that there are many distributions and therefore it is necessary to generate binaries to all of them. Another “advantage” of Windows to dummies is that they never compile a program; the installers only decompress the files. Try to distribute a Windows program only by source code. MS Windows Visual C++ is very expensive and it is not included on Windows.
And I think that Linux is not difficult. Everybody used MS-DOS and have to change to Windows 3.X and then to Win9X and then to XP. What is the difference ? Unix is even anterior to MS-DOS. Unix never was popular because RISC hardware is very expensive but now there are no excuses. The only problem is the Microsoft bain-wash philosophy.
I totally agree with Adam.I think one of the the key Objectives of the Linux Community is to grab the greater part(if not all)of the software market from Microsoft.If linux has to achieve this, then it has to beat the Redmond company at it’s key strenghts-EASY OF USE and EASY OF ADMINISTRATION.Windows is very easy to administer and to use and this is the fact that everyone has to acknowledge;Otherwise forget about linux on the desktop.
I have been working as an IT Support Engineer for more than 3 years now.But it still takes me ages to get something working in linux and yet in Windows it’s only a few clicks and under a minute I am done.
I LOOK FORWARD TO A DAY WHEN LINUX WILL BE SO EASY TO USE
Whatever about your binbox and judging by the innundation of comments here. OSNews is becoming a haven for people looking for Linux News or LinOSNews .. as there are many, many other sites dealing with these topics. I’d like to say a big Irish ‘feck off, ye linux goshites’
I’m totally with you on this editorial, but installation of extra software isn’t the only problem.
Back almost a year ago (Mandrake 8.1 timeframe, if I remember correctly) I gave Linux a more-than-fair try on my systems (P2 333 Dell laptop, dual P2 350 desktop, dual P3 666 desktop). Unfortunately for me, I was trying to use it as a desktop operating system instead of a headless server operating system.
Printing to my USB printer worked great, but only during the install; when I rebooted after the install, my system apparently had no USB devices of any kind attached, despite the fact that the USB modules had loaded properly… absolutely nothing on the bus could be seen. Requests for help or more of a clue on various newsgroups and forums were met with cries of “you must be an idiot” and “works for me”, the two most common answers to any Linux-related question.
OpenGL support was really flakey with my ATI Radeon card; rumour has it that nVidia’s “support” is better, but I’m not willing to buy hardware anymore specifically to try out an OS that may or may not meet my needs. Again, “you must be an idiot” and “works for me” rang out.
Since most of what I do is Internet stuff (Konquerer was OK; awfully slow, but OK), OpenGL programming (almost impossible with flakey drivers), and gaming (totally pointless on Linux unless you like playing Quake 3 all day), Linux wasn’t really for me… I was using it to get/read my email, and then rebooting into Windows 98SE (argh…) to play games.
The final straw was when my ISP went to SMTP AUTH for email; KMail didn’t support it at the time. At least questions about this were met with a more useful response… the KMail in KDE 3 would support SMTP AUTH, but that was still 3-4 months away.
Blew away Linux on all three boxes and install Windows 2000 Professional, a much, much better desktop operating system, which manages to just work 100% of the time. After having spent the previous five years in BeOS, and then about six months in Linux, it was like I’d left prison or something… suddenly my hardware all worked, I could find high-quality software that did what I wanted, and I could easily figure out everything I needed to do.
Note that I’m not some computer illiterate; I’ve used UNIXes since high school in the mid-80’s, I’ve worked for a UNIX-oriented operating system company (QNX), I’ve programmed loads of non-Windows platforms, I’m a shell scripting guru, I’ve had plenty of experience living in non-mainstream operating systems, and I build (and troubleshoot, unfortunately) my own computers. I’m a highly technical user, and Linux simply wasn’t what I wanted/needed on the desktop.
– chrish
Don Cox: Of course all of this contradicts the theory that greater ease of use
will increase market share. The Amiga’s market share is
currently almost zero.
It doesn’t proove otherwise either. Amiga would probably be a big player if wasn’t for that two greedy old man in charge of Commodore’s collapse. People won’t buy something which haven’t had update in years, no?
Charlie: So the author, despite attempts to make the system opaque and practical, still manages to drop something in there by expecting users to view/know file name extensions.
.lif was an suggestion, I’m sure Adam didn’t put much thought into the name. I think the point was in the way it should work.
Yann Klis: Then, Linux lacks Desktop Managers, then KDE3 and Gnome2 were.
Compared to those available to Windows and Mac OS, KDE nor GNOME is not there [i[yet[/i].
Yann Klis: Then, Linux lacks Productivity Software. Open Office, Koffice (and soon Gobe Productive) were.
A lot of Office users (including this sad one here) can’t use OpenOffice.org for lack of features. When i need the features, only Office have them. When I don’t need them, i rather use KOffice. Besides, KOffice and gobe have the feature base of MS Works/ AppleWorks.
It is coming close, but it isn’t there. Though if KOffice had at least half of OpenOffice.org’s developers, I bet kOffice would have more features than Office XP…
steve: I totally agree with Adam.I think one of the the key Objectives of the Linux Community is to grab the greater part(if not all)of the software market from Microsoft.
Microsoft makes more money out of Office :-p. Besides, even if Linux sits like a sitting duck and just makes sure it is the best open source OS out there, trust me, it would grab away a considerable amount of Microsoft’s market share with the same method Microsoft used to grab its market share: price.
paddyponchero: Whatever about your binbox and judging by the innundation of comments here. OSNews is becoming a haven for people looking for Linux News or LinOSNews .. as there are many, many other sites dealing with these topics. I’d like to say a big Irish ‘feck off, ye linux goshites’
Most of today’s news here aren’t related to Linux. Are you suggesting Eugenia to completely bar off Linux-related news? Probably the reason why Linux gets most of the news stories here is probably because there is stories….
You could always post your own news here. (Gosh, was it a month ago someone said OSNews could be renamed as BeNews?)
One of the problems with rpms (besides dependency hell) is you have to download the correct rpm for your distribution. A suse rpm probably won’t work correctly on Mandrake, why ? because they all install kde or some config in a different place. Here is what I propose to solve this problem.
Rpm need to be more generic. Each distribution would set an environment variable, like RPM_KDE_DIR=/opt/kde or RPM_CONFIG_DIR=/etd/sysconfig, RPM_BIN_DIR=/usr/bin, etc. Then the creator of the rpm could just say, install the kde apps in RPM_KDE_DIR and it would go to the correct place on every distribution ! Another thing that needs to be the same across distributions is the naming. It’s hard when a program depends on libkde on one distribution but it’s called libkde3 on the next.
Just my $.02
Line72
Don, I’ve never used an Amiga, I did use a Commodore very long time ago, I blurrily remember both platforms unrelated however (what OS was I using? Basic I guess). Anyway, I only wanted to express how beautiful is your Amiga loyalty. I was listening to Bill McEwen’s Amiga speech (on mp3) and striked me very funny something he said about Amiga users, something like “So what is Amiga? Is tenacious, Amiga is tenacity, if there were two Amiga users left in the world they would go like: Amiga is the best thing isn’t it?, yeah Amiga is the best thing ever, absolutely, it is the greatest…”
There has to be something very cool in Amigas to justify that tenacity. I do not understand very well their future business plan in PPC, they seem to have more of a stronghold in embedded systems.
“Zenja: – why the heck do we need to install small apps. Why cant I simply unzip a small app into a certain
directory and *presto*, its finished. This way I can move the app directory anywhere I want to, and remove
it when I want to.
I think you should be able to remove it when you want to, but place it where you want it to be? Naahhhh.
Just like a HiFi, you have a microphone port, and a headphone port. Could I put the microphone into the
headphone port and expect it to work as a microphone? ”
Zenja is right. A small program should work whatever directory you put
it in, and you should be able to drag it to another place any time.
One of the biggest problems with Linux is the absurd names of the
directories, and the use of the same names over and over again.
Why are there directories called “etc” and “usr”? What does it mean?
I have directories called “Sound”, “3D”, “Video”, “Text”, etc.
Partitions get names like “Programs”, “Data.1”, “Audio.1”, “Browser”,
etc. Anyone can name their partitions and directories in ways that
make sense to them. Why these arcane 3-letter codes?
>everything I’m saying is predicated on the fact that we want
>new, less technical users in the Linux community
Do we? Really? I certanly dont.
And Linux is by far not ready for that, not on its technical skills, but because of the lack of technical skills of the user/newbies.
Let those that love linux for what is it continue to do so,
and dont give the impression everyone wantes everybody to switch to Linux. I certanly dont want that, whats the point of
making a perfect hacker OS into a widespread bulletproof newbie safe OS?
And with widespread use comes commercializing. Bigtime. Ever thought how that might affect Linux?
http://www.freesoftware.fsf.org/kportage/screenshots/
In the past years Linux has met many difficulties and crossed over all them. Some months ago there was no desktop and no office for Linux. Now OpenOffice, Gnome, Kde and others are shining.
I see a pattern here and the current most wished thing is a decent installer.
I’m sure Linux will have a very good installer 12 month from now. My bet are for kportage but I was wrong many times in the past.
If you really want your hacker OS, get a BSD. You have plenty of choice: NetBSD is great! FreeBSD is so great! OpenBSD is extremly great!
Linux can and will be used on the desktop. There are MAJOR drawbacks as of now to using it as your desktop, and the installer issue is one, but they will find solutions as time goes by.
When you look from where Linux comes to where it is, you just can guess where it’s headed. And for the desktop the future is bright.
About commercialisation, what is wrong with it ? How do you want Lycoris or Mandrake to affect Linux ? It can only go in a good direction: funding KDE related projects, easing filesharing, X configuration (nightmare), driver support…
Applications repositories, if standardized will grow larger with thousands of mirrors (eliminate the suse mdk rh/else bins and rpms and debs by only ONE file + src => bandwidth available X5 or 6)!!!!
If user base increases to the point of creating commercial interests, then you will see ports of great windows apps, and who knows, game companies could pull or fund a replacement to X for games. With OpenGL supported by Id, you could see many games ported!!!
Damn, I look forward to the day Linux will have the games I want to play, and will be as straight forward as Windows is. I just love Windows2k, especially since I skinned it to the point you cant recognize it anymore, but its functionnalities are still, it’s stable, consistent, full of apps, games, and it provides me with a much better user experience than Linux does.
Note: why do I switch ? Windows costs just too much. And EULAs… well… I will NOT give up root on my computer, whatever they say. And I dont want to fall in the proprietary trap too. I use Gentoo to learn, with KDE.
First off, Adam talks about the “arrogance” of the Linux community as if it were a monolith, and further writes as if Linux users invented the abbreviation “RTFM”. Ironically, Adam contradicts himself by bringing up GUI installers like IRIS, Click n’ Run, and Red Carpet, which were formed by members of the Linux community in an attempt to address legitimate user concerns that Adam claims that the community dodges with its arrogance. Now you may claim that these GUI installers are incomplete, constraining, or not easy enough to use, and I’d agree that you are at least partly right. That, though, is a different issue. Arrogance has little to do with the state of software installation under Linux. The loudest louts tend not to be the ones writing the software.
Adam also brings up some incredibly specious points. For example, his sample tarball README says “for SuSE, [install] here, for Red Hat, here, for Mandrake, here….” If this were a non-standard tarball that didn’t use autoconf, such a README *might* be plausible. However, this tarball is specifically said to use a ./configure script–standard issue GNU autoconf stuff. Autoconf scripts automatically take care of platform differences not only across different Linux distros, but even across different Unices. The likelihood that the tarball README would advise different locations for different distros is about nil. Linux has enough real problems without Adam making up phantom ones. Please.
Many of the real problems Adam addresses are already being dealt with by the LSB. The LSB’s purpose is to make it so that binary software packages can be distro agnostic, and it already deals with things such as default install locations; that’s one of the purposes of the FHS. Also, the LSB already has declared an installation format: RPM. No need for an .lif package.
As for easy ways to make shortcuts to apps, that is already quite doable. Drag from the menu to a panel. This works in KDE and GNOME.
As for GUI ways of uninstalling software, that is already covered by the assorted RPM/DEB front ends. Kpackage and Red Carpet already deal with package deinstallation.
What you seem to want, Adam, is a streamlined package manager that works with a minimum of fuss and frustration. That is fine, but you could have done that without making specious comments on arrogance, bad examples, and calls to invent standards that already in the works.
James Joyce was arrogant. “Ulysses” is not for your average reader. I expect he wrote the way he did because he thought it was the right way. He must have been fully aware of the fact that “Ulysses” would never be a best-seller. But he wrote the book the way he thought it should be, for himself and a few readers willing to take it on.
Linux software is written by people who are trying to get it right. They are doing it their way. Most of them are doing it for free. They are under no obligation whatsoever to make it easy for Joe Q. Public to use, and people should stop whining about that. The main target group of Linux developers is Linux developers. This is natural and it is their right.
If you don’t like the way they do it, don’t whine about it. Write it the way you think it should be done, or, if you lack the abilities to do so, pay someone to write it the way you want it.
It works great for every kind of soft which use it…
q3
Unreal
Kylix
CoffeCup
and a lot more…………..
maybe it could be possible to make this tool act as a frontend to rpm and whenever people double click to a rpm
Loki’s tool takes place and propose to install it and if peolpe wants to change the path of the install, the frontend do a rpm –badrelocs stuff and voilà….
the main things would be to force people to make rpm easily relocable..
never try to change the path of a rpm ? lol….
Djamé
In January of 2000 I committed myself to Microsoft independence. Eight months into my Linux experiement I switched from Red Hat to Debian solely because I saw my brother using dselect on his laptop. He had just started using it out of the O’Reilly book that came out a while back. Ease of package management matters a lot to me. I’m strongly skeptical that anyone does it better than Debian but I’m pretty entrenched at this point and don’t care to even try anything else.
Two weeks ago I needed more disk space so I formatted the old Windows 98 partitions on the other hard drive. I wasn’t using them.
>why do I switch ? Windows costs just too much. And EULAs… well…
One of my point. If linux gets lots of commercializing, do you really think all of them will be nice and community friendly like RH,SuSE,MDK, etc.? You’ll end up with forks, non-free drivers, licence fees, etc.(Not everything is GPL).
Look at Lycoris and Xandros, already developing non-open software,
I think before you even get into the installer for new software, you have to bring a little more sense to how Linux is organized on your hard drive in the first place. I think part of the problem is that people have a folder on their hard drive named /usr/lib/dev/… and so on. Maybe this is integral to linux at the moment, but the organization has to made simpler. Drivers should go in a /linux/drivers folder. I think Windows has a good system with the Program Files folder (directory, whichever you prefer).
I think part of the problem for Linux is trying to work from the top (the gui) down instead of having a logical user interface come from a more accessible, organized base structure. Either a great deal of effort has to be made to cover this confusing base up (which may be the case with OS X) or it needs to be thrown out and rethought as an evolutionary step in the Linux platform. Remember, the beauty of Linux is its open-source nature, not the steep learning curve that keeps neophytes out.
Aren’t Installation Manager and Packaging Manager are 2 different things?
rpm is a packaging manager which acts upon the *.rpm files. Whereas an installation manager can be shipped with the product, like setup.exe or product.msi under Windows.
You run setup.exe or product.msi and it presents you the wizard screens which walk you through the installation process. Optionally, it comes with silent mode for remote or autopilot installation. The silent mode reads the given config file for default settings and does not present any GUI. Moreover, the installation should allow user to install the product remotely onto multiple machines without having to manually log onto each of them and install by hand. Windows can do this, why can’t Linux? You don’t want to manually install a product on 100+ machines.
Installation is going to be around us for many years. Do some research and make it right and simple to use.
—
Quang
[quote]
I totally agree with Adam.I think one of the the key Objectives of the Linux Community is to grab the greater part(if not all)of the software market from Microsoft.If linux has to achieve this, then it has to beat the Redmond company at it’s key strenghts-EASY OF USE and EASY OF ADMINISTRATION.Windows is very easy to administer and to use and this is the fact that everyone has to acknowledge;Otherwise forget about linux on the desktop.
I have been working as an IT Support Engineer for more than 3 years now.But it still takes me ages to get something working in linux and yet in Windows it’s only a few clicks and under a minute I am done.
I LOOK FORWARD TO A DAY WHEN LINUX WILL BE SO EASY TO USE
[/quote]
You are joking right?
Windows easy to administer? dear o dear where to start on this.
In my last job I administered a mix between linux and NT servers – the linux machines just worked while the NT machines continually needed fixing, often because they suddenly decided that their configurations needed changing.
I find personally windows to not be easy to use once you get past the real basics, whilst the more you use linux the more you can do with it.
I have lost count of the number of times where “design” features of windows or office have forced me onto google or MS support to do someting that could easily have been done on my linux box.
Well written and I agree 100%. Folks saying they don’t need to make shortcuts or click and run apps are livin in la la land. Linux needs all this and more to succeed and these suggestions hit the nail right on the head.
If you want linux to make it on the desktop, you need ease of use, and that is still what the OS is missing. I hope we always keep the underlying technical layer…but ontop we need fluff and bs.
However Gentoo is AWESOME for those of us that know what we are doing. WOO WOO! http://www.gentoo.org/
Adam, you stated in your post that your mail-box was flooded when you posted “Update on Red Hat’s Limbo Progress”. I don’t know if your mailbox is flooded but the forum sure is with 124 responses. Just a bit of irony perhaps. The count is probably still going to go up.
I agree with this editorial 100%. …the shortcoming in a lot of the community is the fact that people frankly don’t want to be presented with options, they want things to work, work fast, and work all the time. Fighting the OS to install things is not what the user wants to spend their time doing…
After reading this statement and a nimiety of similar ones I just have to ask why does somebody in fact try Linux as opposed to sticking with Windows? The immediately obvious answer is that people are not content with Windows and want something better. However, when they get over to that “greener pasture” they begin complaining that it doesn’t do everything exactly like Windows and they have to *gasp* learn something. I just can’t fathom the logic.
The author of this article says that Linux must get a clickable file that is in every way like clicking a setup file under Windows. Then he says he wants to have easy uninstalling like Windows. Nevermind the fact that Windows doesn’t actually uninstall ALL the software most of the time. It only removes enough to keep it from showing up in your Start | Programs menu. The registry, and often times directory structure, continues to be littered with the carcases of “removed” programs.
Why does Linux need double-clickable setup.exe files? How would that make installing software easier under Linux.
To illustrate my point, here is the process you must undergo for Windows to install something:
1) drive to the store and pick up, or download from the net, some software.
2) pay for it.
3) insert the CD or unzip a zip file.
4) go to the directory where the setup.exe file is located
5) double click the setup.exe file.
6) answer questions during setup such as you name, company name, where you want the program installed, how much of the program should be installed, etc.
7) once setup is complete, more often than not you get the fond privilage of rebooting your computer.
To uninstall under Windows you must:
1) click on Start | Settings | Control Panel
2) click on Add/Remove Programs
3) locate your program
4) click the remove button
Under Debian linux, for example, I would:
1) open a command window
2) type “apt-get install <program name>”
Apt does the rest and I don’t have to reboot when I’m finished.
To uninstall, I must:
1) open a command window
2) type “apt-get remove <program name>
And I’m done.
For people who like graphical tools, there is a program called synaptic that is a nice graphical interface for browsing available software and installing and removing things.
Now considering the above, how is hunting for and pecking on a setup.exe-like file easier or better than what is already implemented under Linux.
I wish people would stop running away from MS trash only to land on the Linux crowd and demand they do everything the retarded MS way. It’s unfortunate that some companies, like Xandros, are listening (like adding a C: drive to Linux. Absolute absurdity).
>They are doing it their way
so why do they whine about Linux being ready for the desktop?
Usually you download a compressed file (.sit or .zip), double click on it to uncompress it. This creates a .dmg file which is a disk image (is that similar to an RPM file?). Once the disk image is mounted, usually all I have to do is drag the application into my applications directory. If it is a system update, all I have to do is double click on the installer app. Nothing more. For the home user or end-user, this kind of simplicity is key. I would suppose that if Apple can do it, the Linux community can too.
I agree with you that Apple’s way of installing things is very easy. Apples way of installing software is far superior to the Windows way, but do you know how many times I’ve heard Windows users say, “Apple sucks. How do you install anything”? Once you show them, they still don’t like it. The fact is that people seem to hate learning and complain whenever they have to do it.
>why does somebody in fact try Linux as opposed to
>sticking with Windows? The immediately obvious answer
>is that people are not content with Windows and
>want something better. However, when they get
>over to that “greener pasture” they begin
>complaining that it doesn’t do everything
>exactly like Windows…
Actually they just don’t want to use MICROSOFT Windows. If there were a “Red Hat Windows,” I’m sure people would flock to it.
>Under Debian linux, for example, I would:
>1) open a command window
>2) type “apt-get install <program name>”
I hope you know the package name. If not, good luck. Hell, I tried Debian for the first time not too long ago in the form of Libranet 2.0, and when I wanted PHP, apt-get install php didn’t work. How was I to know it’s called “php4?” Will it understand what I mean when I type apt-get install “cd-burner software” or apt-get install “scanner software”?
<p>
I have read the many, many post criticising Linux and open source software. I just wanted to bring up a point; did you _PAY_ for the software you are criticising?
</p><p>
The author took the time (in most cases considerble time) to write the software and then made it available to you _FREE_ of charge. Yet you still expect more of the authors’ time to make it easy to install?
</p><p>
If you had paid for the software, I could understand your expectation that it be easy. But you get the software for free, and yet you are not willing to spend a couple of hours figuring out how to install it when the author has spent many, many hours writing it in the first place?
</p><p>
BTW, most mature oss _IS_ easy to install either through a packaging system like RPM or apt, or by the old open source 3 step ./configure, make, make install.
</p>
If i want to drive a car, I have to take lessons and read a manual.
If i want to use change the message on the voicemail of my mobile, i have to read the manual.
If i want to use my microwave i have to read the manual
If i want to program my video recorder, i have to read the manual
If i want to program my tv i have to read the manual.
If i want to use the alarm in the office i have to read the manual.
If i want to make diner, i read the manual: a cookbook.
If i want to make music, i take lessons and read book, watch video’s.
Most things take effort to learn: do it, it’s worth it!
Nothing is intuitive: my girlfriend couldn’t install software on a window machine, even if she wanted to. IT IS NOT INTUITIVE. I remember having to explain her the cut and past concept. Even that is dificult.
WINDOWS IS NOT INTUITIVE, MACINTOSH IS NOT INTUITIVE!
I think before you even get into the installer for new software, you have to bring a little more sense to how Linux is organized on your hard drive in the first place
I agree here, I think the file system is a total mess. Because Linux is so large, and worked on by so many people and projects, you could never realistically change something like this. Instead you have to try to make excuses for leaving it the way it is. OSS is good, but it can also prevent real innovation. The longer you wait to change something like this, the more difficult it will be to change. Being tied down on such a large scale will hinder its ability to evolve.
I would rather pay for a good operating system than use a bad one for free. (BTW I do give money to Mandrake)
>If i want to drive a car, I have to take lessons and read a manual.
I learned from my Dad. He told me what to do.
>If i want to use change the message on the voicemail of my mobile, i have to read the manual.
I fool with it until I figure it out.
>If i want to use my microwave i have to read the manual
I expect that it should make sense. I’ve never read a microwave manual, but somehow, my food is still cooked.
>If i want to program my video recorder, i have to read the manual
Not me. I fool with the remote until I get it.
>If i want to program my tv i have to read the manual.
See above.
>If i want to use the alarm in the office i have to read the manual.
My boss showed me how to use ours.
>If i want to make diner, i read the manual: a cookbook.
I cook a lot. I experiment mostly. I rarely use recipes.
>If i want to make music, i take lessons and read book, watch video’s.
I play with my guitar, piano, drums, and bass. I taught myself drums and piano and I picked up guitar in college.
I can and actively do all of those things. I’ve never read a manual. Most of these things have succeeded because people have been able to figure them out. Not because there was a long instructional process.
It must be repeated – average users are not particularly wanting everything else to be like Windows. It is simply that Apple and Microsoft (basically) have established the standards and it is the standards that the average user expects. That is all there is to it. As long as Linux falls short of supplying these standards (in its own fashion), there is no Linux for average users. It has nothing to do with Windows per se.
Again, it must be repeated, the vast majority of people think in a visual manner. That is why the GUI started the personal computing explosion. Example: MS-DOS. Many people got to the point where they could use DOS in the most elementary way – knowing what command to use at the command prompt to start Word Perfect or whatever application. But, beyond that, DOS was impossible for them to understand. Visual thinkers cannot easily “see” things like hierarchical filesystems and command lines in their minds. Abstract thought is very hard for them – and they are the vast majority of computer users.
In order for Linux to succeed on the desktop, doing everything, from installing to uninstalling and everything else must be so easy as to be pitiful. That is the way things are. You cannot get past that.
It is up to the programmers, developers and distributors to do this job. You are the only ones who can do it. You have to start thinking like the Apple and Be and even finally the MS people think – to think like the average user who thinks visually and not abstractly. It is a huge task. That is one reason it was so revolutionary. The geeks began to focus on how the rest of the people think, not on how other abstract thinking geeks think.
And finally, the reason to do all of this is not to bring down the monopoly, not to make big shots rich, but because Linux is a great OS and it should be an option for the average user. We know what’s under the hood. The average person just wants to know how to use the system for their work and play. There will always be people who will want to learn more and I agree with teaching and showing people the amazing things they can do that they don’t know about and teaching them, in a basic manner, how things work. But there are only a certain percentage of people who are even willing and wanting to do that.
As I said before, for our era – our time and place – all this was decided almost twenty years ago and Linux is still behind the curve as far as ease of use. I know Linux is special because of OSS, but it does not change the reality that Linux faces if we want it as a true desktop OS for average corporate and home users.
Did you really need to read a manual for your microwave?
Windows and Mac are way more intuitive than Linux. Of course there will always be people who won’t get it no matter how easy you make it, but we are we are talking about what’s intuitive for an average computer user.
I hope you know the package name. If not, good luck. Hell, I tried Debian for the first time not too long ago in the form of Libranet 2.0, and when I wanted PHP, apt-get install php didn’t work. How was I to know it’s called “php4?” Will it understand what I mean when I type apt-get install “cd-burner software” or apt-get install “scanner software”?
The same thing is true of Windows… If I want to find software for windows, I go to Google and type in “scanner software windows”. If I want to find it for linux, I go to Google and type in “scanner software linux”. The exact same process. The only difference is that once I know the name under linux, I can bring up the terminal and type apt-get install “name_of_package”.
Nearly everyone in this discussion assumes that usability necessarily means dumbing down the system and limiting the user, which is in itself a misunderstanding of the principles of UI design. Real usability and good design are about empowering the user. For an example, take a serious look at OS X. There’s a Unix system that is easy to use but not intrusive, limiting, or “complex and cluttered.” That should be your standard for judging an interface, not Windows XP.
If I may make a suggestion, there ought to be (if there isn’t one already) a Linux development group with a thorough understanding of UI design that begins with a plan for a usable interface and designs around that.
I am currently using MDK 9.0 b3, all the RPM’s that came with the distro work, if it did not come with the distro it does not work. That is easy. I think OOo and Netscape 6 are the only 2 things I have installed on the box that were not installed from the Mandrake CD’s. Arguing backwards compatibility and dependency problems is beating a dead horse, but to say there is nothing wrong with the system in its current state is just wrong. I should also point out that LSB is backing RPM. There are other install methods, but hardly any distros are using them.
If everyone here has such strong opinions, then shut your mouths and implement your ideas. The time you spend bitching about problems would be better spent coding. And if you don’t code, then don’t be a backseat driver to all the people out here who code for free.
If the problem was lack of code, it would have been solved a long time ago, the problem is misdirection. Something communication is actually useful for.
There are only a few words to say from this article.
It’s so damned right ! I agree with it for 100%.
I think before you even get into the installer for new software, you have to bring a little more sense to how Linux is organized on your hard drive in the first place
—————–
Don’t shoot your mouth off if you don’t know what you’re talking about. Take a look at http://www.pathname.com/fhs/pub/fhs-2.2.pdf first. Read it and understand the rationale behind the organization.
Now, once you’re finished that – yes Linux does need (in general) to improve more in terms of placement. A GREAT first start would be if distros rigidly adhered to the FHS. Suse is one of the distros that follow it well. Redhat – not so well.
Realistically, users don’t care where the application is installed on their system. They care about their home (My Documents) folder – thats it. They want to double click an app and have it install somewhere – wherever. As long as they can go to once place and uninstall it, happiness is reached. Its the lack of ****folllowing**** the standard that is causing part of the problem with install/uninstall.
Actually the tradition is to throw all the binaries into /usr/bin and the libraries, configs etc. elsewhere on the disk. FHS does not really address desktop Linux.
Here is the FHS spec for /usr/bin
http://www.pathname.com/fhs/2.2/fhs-4.5.html
Does that look like it was written with GUI desktop applications in mind?
Fact: reason or no reason, the Linux file system is a mess.
You do know the fist step is admitting you have a problem
I know its often hard to get past the denial stage… J
Don’t shoot my mouth off? Listen, I was just participating in a conversation. Save the “not shooting my mouth off” tough guy stuff for when I crack wise about your Mom or something similar to that.
As to your link, I skimmed some of it, then I realized you had just pointed me to a FORTY-ONE PAGE document explaining the logic behind the file structure of linux and I stopped. That basically makes my argument right there. I’m sure there is a very reasonable, forty-one page reason why Linux is arranged as it is, but if the goal is to increase the Linux userbase, then it needs to make more sense from the bottom up. The end result has to be an intuitive, powerful GUI experience. As long as Linux distros are adjusting their GUI’s to apologize for the sloppiness under the hood, the GUI is always going to be second rate.
>If i want to drive a car, I have to take lessons and read a manual.
I learned from my Dad. He told me what to do.
and so on…
Heh. Yep. That’s roughly how I do things too. I’ve found, typically I don’t need to read the manual for ALOT of things. I usually only glance over them really fast for warnings and stuff like that. (I like to make sure I don’t accidently blow anything up. It’s a little late to worry about it after the fact.) Typical use of alot of appliances is ultra easy. (I’m not saying everything is that easy or should be)
If everyone here has such strong opinions, then shut your mouths and implement your ideas. The time you spend bitching about problems would be better spent coding. And if you don’t code, then don’t be a backseat driver to all the people out here who code for free.
Ha! No offense intended, but some people DO code, but are already working on projects. I suppose we’re supposed to drop those right now, because you can’t be bothered? I’ll keep that in mind if you ever make a suggestion about any programs anyone else does. I’ll tell you to do it.
Honestly, if you like to program and actually like your project(s), you want to improve it/them. Not just tell everyone how great it is and tell people to stuff it, when they have a suggestion. And if you sit in your ivory tower, only implementing your own ideas then things are eventually going to get stale and go bad. That’s how a number of projects I know end up going down the tubes. Eventually everyone else leaves, because the “almighty programmers/company/whatever” told everyone else to stuff it. And so the users stuff it and leave; and the makers are left to sit in the dark and talk to themselves about how cool they are.
Computers are complicated .. just like automobiles are complicated. The fact that the controls to a car are easy to use diguises a cars complexity; but it is still there.
As an end user the part of the file system you need to know about is /home/yourdirectory, /mnt/floppy and /mnt/cdrom.
Windows is just as complex under the hood … just take a look at the Registry.
All of the things that you said linux needs is already contained in a nice os with a Unix/java backend and an Aqua/GUI front end. This os is called. OS X. It has graphical,installers, disk images, administrative privelege notifications, etc. You want it you name it. It has .app files instead of .exe files. And guess what. They are virus free. And when you perform an install you don’t have to worry about it installing DLL’s (Damn loser libraries) into 50 different locations.
/home/you – good idea
/mnt/floppy – good idea
/mnt/cdrom – good idea
/ – good idea
It is a mess in general, not every portion of it is a bad idea. Whatever about your car, I don’t care. I know quite a few people who have no problem with Reading The FM’s often but still use windows as a desktop because Linux still sucks on the desktop. Distributing and managing software for Linux is a chore.
just take a look at the Registry.
Windows registry, I rarely use it, but Linux still has /etc which is sort of a registry (and that I use). To write it off and say “It’s fine if this sucks because that sucks too” changes nothing, because in the end it still sucks. The fact that the CLI actually gets used in Linux is, if anything, is more reason to have a better organized file system.
BTW, I notice this discussion has strong resemblance to a discussion about religion. Everybody states their strong opinions on how thing are or should be, but, in the end, nobody is ever really persuaded in either direction by the discussion. To participate in demonstrates poor judgment on my part, and lack of consideration for those that are not on broadband lines and have better things to spend bandwidth on. Sorry.
Computers are complicated .. just like automobiles are complicated. The fact that the controls to a car are easy to use diguises a cars complexity; but it is still there.
As an end user the part of the file system you need to know about is /home/yourdirectory, /mnt/floppy and /mnt/cdrom.
Windows is just as complex under the hood … just take a look at the Registry.
——
Exactly. I sent you to a 41 page document to show you that there is a standard. No one follows it. You as a user shouldn’t give a rats ass where the programs are. They should work out of the box. If the distros followed the standard – your life would be easier. You care about your My Documents folder.
PS. Jim: I did admit there’s room for improvement.
Here’s Something For Everyone:
Set up a webspace somewhere and draw up a document how you think desktop Linux should be set up. Keep in mind these following facts:
1) It has to be backward compatible. Face it – no one, including MS, changes immediately.
2) It has to be inherently multi user capable
3) Certain parts of your file system should be expected to be mounted over a network. Config files should be separated into global config, user specific config and host config.. There is to appear to be no artificial distinction between systems.
4) It has to deal with libraries coming in from a large number of projects. Some of these could be host specific, some will be system wide.
5) It has to deal with a large variety of archs. Some 32bit, some 64bit. (these could have different libs as well)
6) It has to deal with older programs that may not know how to handle this (transition period)
7) It will have to have some structure for applications that users will install (where to place etc)
When you think you’re ready send it out to the standards groups etc etc. Till then improvements to the current situation will happen 1 step at a time. Its a lot more grunngy, less sexy than a total redesign and will get a lot of flack. Oh well.
Feel free to post your website here. I’ll be around.
I don’t find the Linux file system to be a mess. It is a different approach to organizing a system and like all design decisions it comes with advantages and disadvantages.
Personally, I much prefer text config files to something like the registry. They are for the most part human readable (yes I have seen sendmail.cf). They can be edited with simple tools i.e. any text editor. And most importantly, they are atomic. Which means there is little chance of me messing up more than one thing when I change a config file.
Has anyone read read JWZ’s critique of Linux? No one wants to mess with details that mean nothing to them. As a programmer, you learn that one of your most important jobs is to create abstractions. Why is it that Linux can’t abstract installation details away?
More importantly, why are there so many dependencies? No good engineer makes dependencies she can’t control. Installation is a solved problem. Programmers should RTFM for the solution.
The worst thing of all is RTFM. On one side, you have these tireless cheerleaders shouting that EVERYONE must topple the evil Microsoft, and one way to do this is running Linux. Linux is better! they argue. But wait — when people try this, they immediately are treated like idiots who should just use a consumer OS like Windows. If you’re a developer who hates these newbies, tell your cheerleaders to shut up. The users are not the enemy, it’s the cheerleaders.
I dislike Steve Jobs, but I hope he steals a lot of marketshare from the other Unixes.
Sure, OS X is great. And as soon as your average Compaq-buying, Dell-buying, Gateway-buying user can head over to CompUSA or Best Buy and pick up a copy of OS X to replace Windows with I’m sure it’ll be all the rage.
But when you show this user a copy of OS X, explain how wonderful it is, how it is just sooooooo much better than Windows and Linux and it’s based on a *BSD core so it’s rock stable … you’ll win him over, I bet. That is, until you then take him by the hand and lead him over to the Macintosh machine he’s going to have to buy in order to use it.
Then you get to explain why that pretty new Mac, with half the clockspeed of a P4 or AthlonXP, is worth paying MORE for than the x86 machine. “Clockspeed doesn’t MEAN what you think it means,” you’ll tell him. “Look at the AthlonXP! Clock for clock it trounces the P4. Hell, clock for clock, it it’d scale, the P3 does the same thing! It’s a whole different architecture!”
You can go on and on explaining it all to pieces and no matter how logical your comments are, no matter how persuasive, the fact remains that the user is most likely going to look at you with eyes that glazed over early into your pro-Mac dialogue, take another dubious glance at the Mac, and walk away.
Not everyone is willing to throw away thousands of dollars to switch. Why thousands? You’re not just asking them to buy a new Mac JUST to get to use OS X, you’re asking them to throw away all the time, money, and learning effort they’ve already put into their Windows box PLUS add more time and learning effort (not to even mention the money) to learn OS X. Everyone always says, “Mac is a HARDware company! THAT’s how they make their money.” Why, then, do you never hear someone saying, “Try a Mac! The hardware is incredible! Sure, the specs look painfully slow on PAPER but in the real world they’re magnificent!” You don’t hear that. You hear, “Try a Mac! OS X is so freakin great! And, um, oh yeah … you’ll, um, have to buy a Mac to use it on. But OS X is worth it!”
So Apple is a hardware company only because their software (OS X) is so great people are willing to buy the hardware to get the software?? That’s … a truly unique business model. But the fact remains that MOST people use AMD/Intel hardware, OS X will NOT run on it, and the vast majority of those people aren’t going to throw away that investment just for Aqua.
The prettiest, friendliest OS in the world isn’t going to win the day at a price that steep FOR CURRENT COMPUTER OWNERS. Mac should, IMO, be targeting NEW buyers, not entrenched Windows/Intel/AMD users.
Of course, if OS X DID run on Intel/AMD hardware (and all the myriad commodity hardware the platform supports) what would happen to OS X’s vaunted ease of use then? If Microsoft was able to hand-pick the hardware that would run in Microsoft-manufactured computers running a Microsoft-created operating system … well, I think you see my point.
Strath wrote:
If everyone here has such strong opinions, then shut your mouths and implement your ideas. The time you spend bitching about problems would be better spent coding. And if you don’t code, then don’t be a backseat driver to all the people out here who code for free.
Tell us that the Linux community doesn’t profit from increased mindshare. Then journalists can stop writing about it.
Forget installing as well as Mac OS X does. It just needs to get to the MS-DOS standard first.
Installing in MS-DOS? That’s hysterical. Installing in MS-DOS meant
md newgame
unzip -d newgame c:
ewgame
cd c:
ewgame
newgame
There was NO filesystem standard for MS-DOS. It was brain dead simple simply because it was barely functional. Remember what MS-DOS was originally named: QDOS. Quick and Dirty Operating System. Comparing Linux to MS-DOS is like comparing that drunk guy in jail with a harmonica on the Andy Griffith show to a professional orchestra.
Rob wrote:
Installing in MS-DOS? That’s hysterical.
I know what installing in MS-DOS entailed. Your example of using unzip may have been weird, but it was always in a readme, and had no dependencies. Plus, real programs had an installer, which were often simple scripts.
Comparing Linux to MS-DOS is like comparing that drunk guy in jail with a harmonica on the Andy Griffith show to a professional orchestra.
Don’t flatter yourself. Unix was once the Microsoft cancer of operating systems.
“I liken starting one’s computing career with Unix, say as an undergraduate, to being born in East Africa. It is intolerably hot, your body is covered with lice and flies, you are malnourished and you suffer from numerous curable diseases. But, as far as young East Africans can tell, this is simply the natural condition and they live within it. By the time they find out differently, it is too late. They already think that the writing of shell scripts is a natural act.”
— Ken Pier, Xerox PARC
Want more of where that came from? Read the following link, which has some choice quotes from Michael Tiemann (now of RHat) and Bill Joy.
http://www.molgen.mpg.de/~wwwutz/Unix_Haters/
Of course they had no dependencies. That’s why you had, for instance, WordPerfect shipping with its own proprietary drivers for every single printer in existence. And every OTHER program on your machine that printed functioned pretty much the same way. I thought it was pretty much accepted that duplicating functionality in EVERY SINGLE APP on a system was a BAD thing … THAT is what dependencies work to avoid. That way you don’t have 25 programs saying, “This is EXACTLY how to print this, and here are my drivers to get it done.” You have 25 programs saying, “Print this. I don’t care how you do it, just do it.” It’s about efficiency, and yeah, sometimes efficiency comes at a cost in the short-term, but the expense is justified in the long term.
And what does your Unix reference have to do with anything? So now we’ve gone from comparing Linux to MS-DOS to comparing Linux to antiquated versions of Unix? Linux is Linux, BASED on Unix but a far cry from BEING Unix. Following that logic, WindowsXP = QDOS, right? If you’re going to compare it to something, at least try to compare it to something relevant. I use Linux only. I use Gentoo Linux, not Red Hat, SuSE, or Mandrake. You know how many shell scripts I’ve written? A few, because I wanted to. I also used to write batch scripts in DOS. I wanted to. I didn’t then and I don’t now HAVE to.
I love how people throw out a quote someone, somewhere, once uttered as if, by virtue of having been uttered, it attains anything approaching unbiased truth or relevance.
“No one will ever need more than 640K!” — Bill G
(Paraphrased) “I can say with certainty that, going forward, OS/2 IS the operating system of the future!” — Also Bill G, pre-Win95
There’s a quote somewhere to fit every argument. Doesn’t make them valid or even particularly interesting. You could have just as easily trotted out a quote in support of cannibalism, peadophila, or IV drug use. Why not quote bin Laden to show why wanton destruction of civilian targets is a wholesome, worthwhile endeavour?
I see this disturbing trend in lots of computer science books and somem websites now. Every user is a “she”. Why not, “he/she” or better yet, “the user”.
Have we all become so politically correct, so brow beaten by the feminists, that we now have to learn to write and speak in a new way? “The user” by the way, would have been my preference.
Yes, computer science is a male dominated field, but not so much so. I see plenty of women in computer science, and personally — with the market saturated the way it is now, I’d prefer to see no more men OR women being encouraged to enter it.
There are fields like nursing, flight attendents, secretarial that were once female dominated… Do you think people from those fields are going out of their way to write “he” “he” “he” in every printed text book?
The world is filled with morons, the author of this article is a defenite example.
Dear Mr. (or is it Ms.?!)Name WithHeld,
“Every user” is not a she. I have always spent my life thinking of the generic person as a he. As I wrote this, I was trying NOT to think of the “classic” Linux user, so I tried picturing my sister. That’s why it’s “she.”
Your judgemental, sexist, ignorant, “politically-incorrent-because-I’m-sick-of-being-PC” attitude is weak and foolish. I feel sorry for you not only as a tech news reader, but as a person in general, that that is what upset you left with from this article.
Firstly I teach older people how to use computers. Though Windows is more intuative and easier to use than Linux it still falls a long way short of being easy/intuative. I don’t think it is the target Linux should be gunning for…. The problem is that most users don’t remember how tricky it was to learn your first operating system (users tend to be much more forgiving with the first system they learn – I suspect that unlearning is about twice as hard as learning.)
Installing programs on windows Is one process that a lot of users baulk at even on windows. I don’t think that emulating the windows method would be doing anybody any service.
I think that there neads to be a better installer for linux.
*It should be graphical.
* It needs to be able to handle source as easily as it does binaries (and warn the user that the install is going to take a while and that they can probably do something else while they are waiting).
* Installs should happen with one click and perhaps a password With an advanced button that lets you do things like choose different install directory… even compile flags.
* Allow the user to choose which libraries need to be installed on the system and which can be uninstalled when packages that depend on them are uninstalled. Hint these tend to be the libraries which the devel packages are not installedd for.
* Automatically search the web for dependencies in either binary or source form.
* Inform the user if other packages need to be upgraded in order to use package effectively. inform user of most recent version that will work without upgrading packages.
*set up private libraries for applications that would otherwise be broken by upgrading. add app to subdirectory of /usr/old create lib subdirectly move dependent libs to this directory replace binary with script that sets PATH and LD_LIBRARY_PATH variables then runs application. This option
uses more disk space and memory but keeps compatility going.
Any other ideas???
The reason why even smaller apps tend not to go in there own directories is…
With the current linux setup you can pretty much keep your system configuration in tact by backing up your /etc /home and /usr/local directories (for apps you install yourself). So even if you have a hard drive meltdown you can have your system back up and running as you had it within the hour.
Quicker than Imaging your system on to CD’s… Don’t forget to back up your databases as well.
I have been playing around with linux for the last 4 years and have seen many improvements but with the pace of change and the atitude of a large percentage of linux users and developers linux will always remain a bit player on the dest top.The best way to make things user friendly and productive is to take what works well and improve on it. At this point in time windows is the standard so copy and improve on it. If the geeks of this world continue to not understand the reality that the operating system is just part of a tool that must be usable by the vast majority of the population with minimal of training then Microsoft will always control the desk top. People are not coming to linux so linux must go to the people.
> I thought it was pretty much accepted that duplicating
> functionality in EVERY SINGLE APP on a system was a BAD
> thing …
Ok Rob, so we’ve both established that installing things on GNU/Linux is worse than on QDOS, Windows and OSX. I rest my point. Just please tell the Linux cheerleaders to stop, and the Linux businesses to quit selling. Because Linux only has excuses and RTFMs.
Linux and open source software is by and large written by unpaid volunteers who generously share the results of their labor with the rest of us. If you do not like it or find it does not suit your needs, you are not forced to use it.
Buy proprietary software if you do not like open source software. Only you can decide whether it is better for you to buy software with money (which for most of us takes time earn), or with your time and effort.
To me, the effort to learn to use Linux and open source software is the price of freedom.
Furthermore regarding Robert, I have never been told to RTFM. My questions have always been politely answered. But, I always try to find the answers myself first, and I ask my questions politely amd make it a point to thank people for taking the time to answer my questions.
Read my post again. I do not direct my criticism to volunteers, but to cheerleaders and companies.
I have always used a boxed, paid for Linux distro. And I’ve even completed contracts for “Linux companies.” I don’t say anything bad about Perl or other labors of love, tempting though it is. 😉 But Red Hat is a company, which wants my money.
Would you prefer if I criticized the author of this article instead, for useful observations? Apple zealots used to be terrible, but on OSNews, it’s the Linux zealots who just make excuses and flood inboxes.
The price of freedom doesn’t pays of to most people. Most people don’t care about having the source code, or free flow of information that Stallman dreams about.
Trust me, if you can convince them that they would be more productive on Linux (or any other OSS software) after an initial training period, they would move. A lot of my friends have moved to Mozilla from IE (Opera is not popular among them, I guess), because Mozilla gives them a lot of features that would help them while surfing.
>>If i want to drive a car, I have to take lessons and read a manual.
>I learned from my Dad. He told me what to do.
Somebody telling you or reading a book, it’s the same: it not intuitive.
>>If i want to use change the message on the voicemail of my mobile, i have to read the manual.
>I fool with it until I figure it out.
Yes, but did you understand it the first time you saw it? If not then it’s not intuitive.
>>If i want to use my microwave i have to read the manual
>I expect that it should make sense. I’ve never read a microwave manual, but somehow, my food is still cooked.
Did you really understand all the icons the first time you saw a microwave? (microwave, hotair, grill, defrost?) Do you know how to combine microwave + hotair when you want a pizza fast, So how long does it take? Half of the time compared to normal hotair? I don’t know from looking at the microwave.
>>If i want to program my video recorder, i have to read the manual
>Not me. I fool with the remote until I get it.
The fool around does not equal intuitive. Do you think your mother can program the video by fooling around? If not then it’s not intuitive.
>>If i want to use the alarm in the office i have to read the manual.
>My boss showed me how to use ours.
Somebody telling you equals reading the manual => it is not intuitive.
>>If i want to make music, i take lessons and read book, watch video’s.
>I play with my guitar, piano, drums, and bass. I taught myself drums and piano and I picked up guitar in college.
So do you know about all the different modes and scales, just by looking at the piano?
>I can and actively do all of those things. I’ve never read a manual. Most of these things have succeeded because people have been able to figure them out. Not because there was a long instructional process.
Yes, i figured apt-get out by fiddling around, but that does not mean that it is intuitive.
Are pretty dumb.
If you’d taken Mandrake, which is designed more as a desktop orientated operating system, then the experience in completely different.
I installed it for my younger brother recently and tried to take the GUI approach, instead of my normal one.
It was a case of download an rpm. Click on it and rpmdrake lauches automatically and takes you through a GUI that’s simpler than any windows installer I’ve ever used. It installs the RPM for you, fetches any dependencies in the process, and adds it to the menu.
The only other step I had to do for one program, not packaged as a mandrake RPM, was to launch menudrake and add the program to the menu.
It was very quick and easy, and by far the easiest RPM installations I’ve ever done. The only improvement that could really be made is for the program to detect a program that doesn’t have a menu entry and lauch the program to add it automatically.
I agree too, but let us not hold back. the bottom line is a matter of simplicity, which even the hardest hacker can appreciate. linux has made things hard on itself by holding on to a legacy file hiearchy. but no matter. joe user dosen’t want to know a damn thing about it. so forget installers all together! give me a menu for “new software” under each of my typical menu categories. each one should have a tooltip explaining it. when i click on it, since the app isn’t installed yet then the computer says: “hey! this isn’t installed do you want me to install it?” i say yes and boom (after however long). its done and running. now the app appears among my typical menus. then go the extra mile and put a timer on each of these. if i haven’t used in app in over a year then auto-uninstall it. don’t even ask me. i can always reinstall it if need be with no harm done.
finally let me say until we get the above, i’ll stick with debian.
over and out.
~transami
I realise that this article is written about a specific distribution, but it seems to me that so far, RedHat has been aimed more at server installations, while other distributions are more aimed at desktop systems. In fact recently RedHat issued some press releases to the effect that they are starting to address desktop systems with a new release of theirs.
I use Mandrake 8.2. When I need to install an RPM package I can use its graphical “Software manager” tool (rpmdrake). It works out all dependencies, warns me of conflicts, prompts me to insert any CDs it needs and/or downloads updated versions from the Mandrake site, etc. One can add new sources to this tool, such as FTP servers, HTTP servers, CDs, etc. I have never used Debian but this sounds a lot like apt-get. There is a command-line version of the same (urpmi).
There’s nothing stopping RedHat from adopting a similar system, or even forking off Mandrake’s software.
No-one mentions Synaptic which is a pity, Synaptic is a
great GUI to APT. Synaptic is available for Conectiva, RedHat and SuSE (other distro’s I don’t know).
With the tools provided by the apt4rpm project, it is
possible to create an APT repository from _every_ rpm
repository. There is not much involved to do this, once
set up it is run from a cron, that’s all.
Check out these pages to obtain all information about APT for RPM:
http://apt4rpm.sourceforge.net
(visit especially the repository and related matrial pages)
http://linux01.gwdg.de/apt4rpm shows a howto to setup
apt if the distributor (in this case SuSE) does not deliver it.
If you do not want to spent your live by solving dependencies. With DEBIAN & dselect (dpkg), the instalation of software is simple and COOL. All dependencies solved in a few seconds. I simply do not understand why people (from red hat) do use that rubish rpm system where you have to find dependencies
manually (I had RH for few month couple of years ago and it was really strugle with looking for dependencies on file xy, which I did not have any idea what it was). I have not seen better software installation management than dselect/dpkg.
To know how to fix a problem, obviously you must first properly identify your problem. I believe that most people using Windows at home are happy with how well it works.
So, is Joe Desktopuser even looking for something else? If so, why? Maybe it’s because M$ pisses people off. Not so much because of how well Windows works. I think people are having problems with the liberties that M$ takes. For example, changing their homepage to MSN after an IE upgrade. That’s just one example.
So, do I want Linux to be more like Windows? In ways, yes. But, I think I’d use OSX as a better role model. OSX’s ease of use and popularity proves that the underlying file heirarchy isn’t a problem. Can you imagine what would happen if OSX was suddenly open sourced and ran on the PC? Jeeze Louise. It’d be on my desktop faster than you can say “Reformat”.
So, for Linux to gain more market share on the home user’s desktop, Adam’s remarks make perfect sense. Emerge, apt and RPM’s are OK. But, there needs to be something better. A standardized installation method supported by a comittee made up of a few people from different distributions. Now that there is a LSB, the distros wouldn’t have to do anything different. They could continue using RPM’s for those who like them. You could even build from source if you choose. The new standard could just be built on the assumption that you will be installing the application on an LSB compliant distro.
Finally, for those who think that apt is good enough for end users, you are just wrong. For example, I had to teach my dad how to find programs (like hardware drivers), how to save it in a directory and, how to use Windows Explorer to navigate to that directory just so he could double click it. That’s typical.
100% on target! I am an irrational enthusiast when it comes to Linux. I manage practically all Windows-version at an advanced level, so I suppose you can say I am not a computer-inept. But I LOVE Linux and would love to use it for all my work. But it is EXACTLY what you say that keeps me from doing it.
Just look at this: I installed a SuSE 7.1 (and later 8.0) totally Standard-out-of-the-box (even with developer-packages, since I heard that you need them to install a program – which doesn’t make sense, either). However, when trying to install a simple little program (“kshowmail”, from source) it just didn’t. Libraries missing etc. pp. And when trying to install the missing library I run into the next problem and so on and so forth. That shouldn’t be that way.
And I could go on giving you examples for at least 10 pages. And if it keeps me, with my enthusiasm and certain knowledge, from using Linux, there are millions who would like to try or use it but don’t.
And I very much liked your proposal with the install file, a clean uninstall-option (it took me a week to get to know “make uninstall”), a clear definition where to install what type of software (until today I have not found out but I remember an article of Mosfet adressing this problem) etc.
Kpackage is fine and dandy but what if SuSE decides to compile kpackage without rpm-support (as has happened in one KDE-Version)? What if I want to update KDE itself (so I’d add that at least for those basic packages there should be a text-based installer-version too – or KDE updateable from within and the changes taking effect on the next login)?
Keep up the good work! If I can be of any help (not being a programmer but a multimedia-conceptionist), I’ll be more than glad to provide it.
– Stephan
The price of freedom doesn’t pays of to most people. Most people don’t care about having the source code, or free flow of information that Stallman dreams about.
—–
I was only speaking for my self. I choose to use Linux for my reasons. I also accept that having made that decision, some effort is required on my part to make it work.
To me (once again I am only speaking for myself) the small amount of effort it takes to learn to use Linux is well worth it for many reasons:
1. I will not support a monopoly with my money, especially one whose behavior has been as bad as Microsofts’.
2. information technology has become an essential part of participating in modern society. The Internet allows everyone to have a voice in a way that corporate controlled media never will. For this reason, I add my support to the community by using their products.
3. I like the openess and flexibility of Linux; everything is open and I can control/learn about any aspect of it I wish to.
4. It is _MY_ computer. I will decide what software runs on it and who has access to it _NOT_ Microsoft or some other corporation (see their new EULA). Furthermore, I will decide when/if I upgrade and what hardware I will purchase to run it on (if it was affordable, I would run a none PC based system).
5. The resulting system, while may be somewhat difficult to set up, is powerful , flexible, stable and secure.
I have always used a boxed, paid for Linux distro. And I’ve even completed contracts for “Linux companies.” I don’t say anything bad about Perl or other labors of love, tempting though it is. 😉 But Red Hat is a company, which wants my money.
Since you purchased a product from a company you have a right to expect to work as the manufacturer says it will. If it does not then you should rightfully expect the company to correct the problem.
But as you stated, Red Hat wants your money. Therefore they are motivated to keep you happy and I am sure would want input from their customers on how they can do a better job.
Would you prefer if I criticized the author of this article instead, for useful observations? Apple zealots used to be terrible, but on OSNews, it’s the Linux zealots who just make excuses and flood inboxes.
The only part of my comment that was directed at your post (not at you) was my comments about RTFM. And I was only relating my experiences. I have always asked for help in a respectful manner and have found people to be very helpful.
I also realize that people are taking their time to help me with no expectation of receiving anything in return and do not having any obligation to help me. Therefore I am grateful for any help that I receive and let people know that. How you approach getting community help is completely up to you.
As far as criticizing people, I would prefer you that you criticized no one. But, I respect your right to free speech whether I agree with you or not.
In general, I agree with the sentiments expressed by Adam. I do have a couple of comments.
1) “Is it really that common that we tell someone to “Read the F*cking Manual!” that we need to abbreviate it?”
Actually, yes. It is the /most/ common, appropriate response to a vast number of questions. There are a lot of really stupid users. Even the non-stupid ones don’t read the manual before trying to install; the non-stupid ones at least read the manual after they encounter problems, whereas the stupid ones immediately go looking for someone to hold their hand. Even if you put, on the download page, in BIG LETTERS: “You must do XXX to install this software, or it will do YYY”, I guarantee you that some idiot will email you saying “Your software is stupid. I tried to install it, and it did YYY”. “RTFM”, IMO, is an appropriate response.
2) Curiously, I was in another state when my wife, who has Mandrake running on her laptop, needed to install Yahoo Instant Messenger. She found the RPM, downloaded it to the desktop, double clicked on it, entered the root password when prompted, and it installed. I was fscking impressed, when I found out later. My wife is /not/ a “computer person”. That said, she was extremely lucky. I’d guess that 3 in 10 RPMS will install without some dependancy conflicts; that is because RPM is fundamentally stupid, and should be taken out and shot. urpmi, and others, doesn’t help much; they tend to be difficult to set up, and terribly inefficient.
3) The most impressive package management system I’ve yet seen is Gentoo’s, which has powerful dependancy resolution capabilities. It builds everything from source, which means that there are never any library incompatibilities. I haven’t yet run Gentoo on a desktop, so I don’t know if there’s a GUI for the installer — but one of these days, when I get tired of not being able to upgrade Mandrake because of that wasted space of a package management sysetm, RPM, I’ll replace it with Gentoo and write a front end if there isn’t. The downside to Gentoo is that you must have a working GCC on your system, which isn’t practical for a lot of computers.
While I agree, in principle, that Linux should be easier for newbies to use, I strongly feel that should not come at the cost of flexibility for the Linux community as a whole. That’s the job of the various distributions. Red Hat for newbies, Slackware for experts, Suse for those in the middle.
Or mandrake for everyone
The point is, Linux provides a base for distributions to build on, it is up to the user to choose the one right for his needs.
-Bruce
No home user gets Linux on their computer without intentionally putting it there, so she must have had some inkling that this was a geek oriented OS. It is unlikely that she bought a computer with Linux pre-installed because stores selling such things are as hard to find as a dodo. Corporate users don’t need to install their own software. That’s what sysadmins are for. Linux came into being because some people were dissatisfied with the available commercial OS’s. The important thing is that users should have choice in the marketplace. The geeks who want a hackable system should have the choice to have Linux. Those who want an easy system to get their work done have Windows and Mac available. Linux shouldn’t be forced on everyone, just as Windows shouldn’t be forced on everyone. It is great that Linux is becoming easier to use and more accessable to the less technically oriented user, but that doesn’t mean that it is appropriate for everyone. I don’t think it is a shame that many people will find
Windows or Mac more appropriate to their needs. There are different communities of users with different needs. The most important thing is that each user be able to find a system that meets his or her particular needs, not that Linux or any other system has “world domination.”
*Gosh*, so I have to start looking arround on the web to find a piece of software? Do I really have to download my software from some slow server?
Can’t I just start a program that gives me a categorized list of installable programs, which I then can easily install?
Oh wait, I can, never mind 🙂
I think there were a lot of good comments made. The main point to keep in mind from my perspective is the purpose of any tool whether it is a pen, screwdriver, pliers or computer system. The purpose of a tool is to help the user to accomplish some task in the most effortless natural manner. The computer is a tool. It should be intuitive to use the computer as discussed in human interface books.
I think your list of 5 reasons to use Linux makes an excellent point. Reading back through them, you’ll notice that none of them were based on how poorly Windows performs. The fact is that Windows, on the desktop, doesn’t perform all that badly. Still, they are excellent reasons to switch OS’s and they parallel my sentiments.
If you’ll notice, your only reference to usability was to the effect of Linux being difficult to set up. So, why is it such a bad thing to emulate some of the things that Windows gets right?
Of course, if double-clicking a file is a security risk, because the package might contain maliscious code, this would not be worsened by the double click system propsed by the author of the article. Honestly, do you think you are ANY less vulnerable when you download *.rpm from a random web site, drop into root and install it? Of course not…If anything, a double click installer file would only make it slightly easier to hose your system, but a novice user could easily do the same in the shell. In either case, I have used most of the major distributions and, in my opinion, application installation will almost always be difficults in linux unless these distrubution undergo major architectural changes. The morrass of libraries, in particular, those associated with Gnome, are, in my opinion, way too modularized. This is clear enough if you’ve ever tried to install Gnome without a handholding installer like Ximian. While there are advantages to having libraries seperated into many small parts, e.g. you don’t have to build a lot of redundant code into aplications, it makes installation a nightmare. Moreover, and this is very common in my experience, if you want to install a new package it is often the case that a different, crucial package relies on an older version of some library required by both, and there exists no updated version of the package. When this happens, you are stuck: use the new package and loose the other one, or vice-versa. This was the case for some time with GNUCash and Gnumeric, to name just one example.
Lastly, all of my linux wonderings have also taught me somthing else: if you need a powerful desktop environment with a real productivity suite yet demand a unix based system, you really only have one option at this point in time: OS X. Linux geeks who say linux is a powerful desktop probably don’t use it in the way that most desktop users do: I need, for example, to be able to paste a graph from one program into another without hassles (try pasting a gnumeric graph into abiword, for example); I need HIGH quality font rendering when I type a document or surf the web (there are loads of sites you could spend DAYS reading only to get a marginal improvement in XFS’s font rendering); I need to be able to configure the system without editing 13 configuration files and traversing the file sysyem looking for where the relevent configuration program is, e.g. Xconfigurator (granted, this is rapidly improving in the big distributions); lastly, and this concurs with the author, I don’t need hundreds of superfluous applications, e.g. 34 ftp programs, 19 web browsers, 7 text editors, all competing for the same crowded application menu space (remember the joke that was Redhat KDE menus in 7.x?).
Yes, I can do all of this myself, but really, like most people, I can’t be bothered…It’s an utter waste of time for what is really a marginal desktop OS (meaning Suse or Redhat, Mandrake, etc.)
If you’ll notice, your only reference to usability was to the effect of Linux being difficult to set up. So, why is it such a bad thing to emulate some of the things that Windows gets right?
I am all for Linux being easier to use and more widely used, if only for selfish reasons; larger userbase, more potential developers _AND_ a larger voice in the Internet, Government and media.
I don’t personally find Linux that difficult to use now, though it does take some effort on my part.
As far as emulating Windows installer, I think it would be great if there were widely available tools that would allow developers to package their software in a way that allowed easy installs. Both OpenOffice.org and Mozilla use very polished and easy to use installers. There is no reason that tools could not be developed to allow all software to be packaged in a similar way.
But as distros move to FHS and LSB, I think you will see something more like what Debian has. You go to a web site, or start some kind of package manager and you are presented with a categorized list of all available software. You pick out what you want, and the software installs itself and resolves all dependicies regardless of what distro you are using.
Another thing I don’t think people take into account about Linux and Open Source in general is that it is like watching a play, or a painting from the very beginning. You see all of the mistakes, false starts, and occasional arguments (OK, not so occasional) of its’ passionate creators.
With commercial software, all of this takes place behind closed doors and only the final, finished product is shown to the public.
I firmly believe that Linux and open source in general will rapidly over take commercial software. The free sharing of knowledge that is so important to open source is the very reason that science has been able to make such remarkable progress.
1) I agree that it should be an option (and probably the default option) for users to install applications strictly for themselves. )Might be nice if they registered themselves (via a standard command – into an ASCII database – not a binary registry!) into a system file, so that the system admin could see that 12 people have installed the newest P2P app, and decide whether or not to install system-wide, or block the ports it uses. However, not all linux users *have* root access on the machines they use. I’m sitting at a public library with 26 machines for general public use, and another 25 or so for staff use. All of these run linux, and we have some 4000 accounts on this system (don’t bother trying to hack us – the only thing of interest in all of this mess is most likely 50 copies of the latest Britney Spears MP3). We have several users who have found things like TuxRacer, Maelstrom, GNapster, and other apps they’re interested in, but they can’t get them to install into their space. You think make install is bad? Try make install –prefix=/home/luser/. I can install from RPMs, I’ve even built from tarball a few times – I ran into this trying to install Mutella, and I *still* can’t talk Mutella into finding the readline library that is already on this machine, *OR* the newest version, that I *did* follow the make install –prefix instructions for userland installation.
2) App developers – PLEASE, PLEASE, PLEASE!!! Quit using the latest bleeding edge libraries straight from the Gnome/KDE/MPlayer/whatever CVS repository! Is two major distro revisions of backwards compatability too much to ask? This same library has most of the machines running RedHat 6.x – Why is it getting almost impossible to find RPMs built for RH 6.x as it came out of the box? Even YMessenger, even though it has RH 6.2 RPMs, wants other libraries updated. And we don’t even actually have RH 8.0 out yet!
After reading the article, and all attached comments, I have to wonder why so many people insist on ignoring some of the solutions put forth. It just sounds like people like to talk, rather than listen – deflecting the conversation into unnecessary rat holes.
First, the basic premise that an unskilled, newbie user should be installing random software is silly. The idea was that a neophyte user would download a random piece of software to their system, then try to double-click to install. No. The newbie user ought to be installing software from a managed repository, such as the Debian stable distribution. Someone dismissed this idea (working from a managed software repo) out of hand, thereby dismissing an entire solution to the “problem”. The debian distribution, with its online repository and the dpkg/apt/dselect/+++ system simply does not have the dependency hell problem. End of story.
Second, it seems many people in this long thread simply ignore the fact that there are graphical installers available for managed linux distributions, that these installers do quite well at categorizing the software in ways that everyday users can understand (well, that’s really an underlying feature of the packaging system), that these installers show what software is already installed, or what’s available, and that the software install / removal process works flawlessly.
Yes, users could type apt-get install <package>, but they needn’t. When Joe or Sally User wants to install software, why not run dselect (okay, I suppose) or the graphical frontend mentioned in this thread?
I think part of the problem with this entire discussion is that nobody has highlighted the notion that leading linux distributions, such as Debian, are leading the way toward a new model of software distribution. It’s no longer “safe” to install random pieces of software on your system. You *should* be getting your software from an authentic, verifiable source – rather than from some random location on the internet.
If some linux developer wants to release software, he should be packaging it for distribution via one of the main channels, where the dependency issues are resolved. In the case of a Debian system, an independent SW developer can setup a repository of their own software (their “distribution channel”), which can then be merged via apt.sources with all the other software out there. This approach is already in practice.
Users should be going to known, trustworthy palces for their software – for their own protection, if for no other reason.
Sorry for the rant! Guess I just had to add my 2-cents to the discussion.
Great job on this article – I think you have it just about right. My thoughts on new user software installation are similar to yours in many respects, but to further increase the ease of installation and administration of software, you want to have a single repository of applications that going through some formal method to ensure/certify that they do in fact work with a distribution and install to a common location.
I think that FreeBSD, Gentoo linux, etc… have the best example of this so far, but in a too technical fashion for the newbie user – ports. Ports provides the single point of installation, using the same or similar commands to install software each time, installs to a default location that may be easily recognised by new and experienced administrators and can be monitored to ensure that each piece of software does as it claims (reduces the fear of new users installing software that they dont know – new windows converts always ask “but what if it has viruses? how do you know its going to work”). The question is how to take that great first step and create a user interface that people find easy enough to make the defacto method of installing.
I’m trying to get my household switched to Linux. It would be so much easier for me to administer, personally, because of the lack of Windows “magic”. I can’t reboot Linux and have things magically work that didn’t before, or reboot and have things stop working unless I did something specifically related to the boot sequence or the program that fails.
In any case, the only person in my household who is on Linux full time is my little sister. She does her email, web, etc. and has no problem with it. It works better and more reliably than Windows, it hasn’t just decided to quit reading DVD movies as was her previous OS experience and it burns CD’s just fine.
The rest of us are players of EQ and until WineX gets EQ support, I think we’re pinned into Windows. It’s a nasty habit and I keep meaning to break it. Further, my brother is into Multimedia Fusion, Games Factory, etc. put out by Clickteam ( http://www.clickteam.com ). There are no such tools in Linux and he’s afraid to even dual boot so he can learn Linux, despite knowing all of the benefits and wanting to switch. He’s seen how when I was dual booting, Linux would eat the Windows partition and Windows would kill the bootloader and in the end I wound up with a system that won’t boot from a boot disk and will only run one operating system at a time, despite all of the configuration hacks in the world.
Installation hasn’t been a problem at all for me or my sister. The problems we’ve had are in being able to do what we want to do (i.e. Games Factory, EQ). All other factors aside, these two pieces of software keep almost my entire household tied to Windows.
I can’t belive you people really prefer the windows way of installing software?
When I want a program I look in the software manager which appears in both my kde and gnome desktops. I search for what I’m looking for and I click install or uninstall if that is what I want and it works. If I download something then I download a mandrake rpm and i click on it a thing comes up telling me what it is and I click install and then try the program which appears in my menu. If I don’t like it then I click on software manager and remove it. I don’t know where it is installed and I don’t care. I don’t even know how it gets in the menu.
I also download tgz files and type uncompress them by you guessed it clicking on them and then I bring up the terminal emulator and type ‘su (enter my password)”./configure && make && make install’ and if it works I pat myself on the back and if it doesn’t I don’t care but what frustrates me is that I have no easy way to remove it if its rubish so I love rpm.
Rpm never presents a problem. On windows which I use for games and notation software loads of things don’t work when I install them and especially games. I often have to download patches and reinstall and all sort of other rubbish and then when I install something else I find it broke my software I already have and even windows itself and I almost pray for the day when I can have the notation software I need for linux and forget those windows wizard for good. I have a mac (os9) aswell and that gives me no good way to remove rubbish. But rpm does without the command prompt.
The easiest solution for all distribution is simple to use “make”. This provides
make install pkgname.[deb|tgz|rpm]
This is essentially what all BSD systems do … they provide a top level make system that keeps binary packages and source packages in order will work as a build and installation system and package/installation manager. It’s what Debian apt and Gentoo “emerge” are
Once that’s in place a simple web based front end to it all (for both local and remote pacakges) is the UI to tie it all together. Something like CUPS web based management interface but prettier (most XP applications and even directory browsing look more and more like web applcations). Browsing to localhost:124234 {port number for the package management service] would present the UI. Browsing to another machine on the netowrk https://someotherhost:124234/ you could package manage it too (with sufficient rights).
I dare say mine is the best idea presented here so I hope people will comment on it.
If you’re not contributing to whatever open source project you’re having problems with, nobody’s going to notice your problems.
I suppose that it is something that many of us are wishing for .. simple installation/removal of the same. In the same breath don’t take my CLI away as it is one of the most powerful features and the reason I stay with Linux. yes .. I am not a nerd but I feel really comfortable using the command line.
Don’t get me wrong, but I also resort to GUI for many other things and feel just as comfortable there as in the CLI.
The bottom line is that I do agree with the need to find a better way for the ordinary citizen out there to like Linux and still have the CLI for those who want it.
Your article is interesting to say the least.
Using MDK rpmdrake and installing everything is as easy as in
microsoft. Though through this way we cannot get the latest
version of the software by using tarball. But considering the
easy way to install the software just by click and run, I think it can be tolorated.
I think you’d get a LOT further solving installation problems if you paid attention to the parts that have already been solved well.
Frankly, installing new software in Debian Linux is already much easier, simpler, and far superior to the elaborate GUI system you would like to do from scratch.
I know people tend to ignore this type of comment because they presume you’re just a die-hard Debian fanatic, but there’s a reason why so many people are Debian fanatics, and it’s NOT just to be “more elite than thou.”
Once you find the software to install in Debian, it’s a very simple matter to install it:
“apt-get install package-name”
No mucking around with dependencies. Debian put all the dependency information into the package when it was built, and the APT software installs any necessary software to satisfy dependencies. Why step a user through a wizard to accomplish something when the packaging system already has all the pieces of information it needs to fix the problem?
Now of course this isn’t perfect, but the bulk of the problem you point out has already been solved. Debian’s system installs the software, searches and installs any programs the software depends on, and adds the software to the GUI system menu. Without asking the user to figure out dependencies or hunt them down himself.
So can we just pack up and go home? Of course not. Here are (most of) the drawbacks to Debian’s approach.
1) As someone mentioned, you have to know the name of the package you’re looking for. This is the only real problem with installing or uninstalling Debian software. Yes, Debian really needs a simple comprehensive guide to finding the package you need for the task you’re trying to do. However, do you really think this is just a problem for Debian? If it was, there wouldn’t be such a thing as RPMfind.
2) This only works for Debian. If you use another distribution that isn’t Debian-based, you can’t expedt the Debian-based solution to work. (Though Connectiva’s apt-rpm has potential)
3) This works for Debian only because Debian has strict policies about where and how software is installed. These policies make it possible for all Debian-managed software to work together.
4) Debian software is tested to see how well it works with other Debian software before it is officially released. This is necessary because policies can’t catch everything, but this is the primary reason why Debian releases tend to be behind. Of course, this doesn’t bother the Debian people because they’ve learned how to add sources from the development version of Debian and install any new software from there if they really want the new version now.
In spite of the drawbacks, Debian is a lot closer to simple package management than any other system out there. The reason is simple. Package managment is very hard, not easy. The only way to make it easy on the end-user is for the developers to do the hard work for them.
As for the objection that Debian doesn’t let you put the software where you want: Why should it? Does the average user really feel better knowing that Mozilla is installed in /Software/Internet/Browser/Mozilla instead of the default? What they care about is being able to find it on the menu/desktop when they want to run it. Would the average user want to have to remember where they put software so they could go delete it, or would they rather just find it in the package manager software and tell it to go away?
Just have to comment on this:
>Your judgemental, sexist, ignorant,
>”politically-incorrent-because-I’m-sick-of-being-PC” attitude
>is weak and foolish. I feel sorry for you not only as a tech
>news reader, but as a person in general, that that is what
>upset you left with from this article.
Seems like the pot calling the kettle black, or the “judgemental calling the judgemental judgemental”.
I wasn’t the person complaining about the “she” “s/he” usage, but I understand his point. I get absolutely sick of so-called “non-sexist” language. When you don’t know or don’t care about the gender of a hypothetical person, proper English is to use the generic masculine “he”, “his”, “him”. The reader knows full well that “he” implies “she” as well. It’s part of the English language.
On the other hand, flipping between “he” and “she” to try to be non-sexist, or just using “she” so you can’t be accused of being non-sexist, constitutes a drag on communication. When I start reading some How-to or commentary and start encountering “she”, I always think I missed something, because I see “she” as if the sex of the person has been identified. Then I realize it’s just someone who thinks this constitutes “non-sexist” language. Bad idea. If you want to establish yourself as non-sexist, make a little speech about being non-sexist and get back to communicating to the reader.
Of course, the “gratuitous she” is not nearly so bad an idea as those who use “he” in one place and “she” in another. It sounds like the writer is taling about two different people!
Am I being sexist to say “non-sexist” language is pointless and confusing? Hardly!
Am I “judgemental” to say that I find the whole practice to be gratingly bad English? I’ll accept the word in that narrow context.
Is this attitude ignorant? Not nearly as ignorant as the presumption that the attitude must be ignorant. That presumes that only ignorant people don’t believe in the virtue of “non-sexist” language.
want to untar anything, but want to do exactly what you’ve described. I’m been using Linux for 3 years now (former Mac/Windows user), and there are STILL problems with “normal” operations that MS and the new Mac OS X have that Linux devotees could learn something from. Linux people, if you don’t change with improvements, then there will be NO more world domination in the future. Each tool has it’s purpose, but tools do improve and other tools are created. Linux needs to mprove to reach the masses or die a slow death!
Someone complained about hardware support in linux. While i can understand the complaints, this is one thing that can’t be changed easily. The reason some hardware is unsupported is not because noone wants to add support for it, it is because the hardware vendors keep their specifications to themselves and don’t provide drivers themselves. If you don’t want to care about hardware support when you buy hardware, well, fine, but you probably wont have much success running linux them.
—–
For those people complaining that software is too hard to install, well, i can see how it can be a problem for you, but i dont think you should expect any general solution anytime soon. I doubt many of the developers see this as a too big problem, and those who do probably see it as hard and boring to solve. Thus not being said that i don’t care about the problem, i do. But i basically like the current way things works. You most likely need a package for your exact version of your distro because of binary compatibility issues anyway, so dependencies shouldnt be THAT big of a problem, at least it hasn’t been for me. In the future when things stabilize and binary compatibility wont be as big an issue then this could change, but if that is the case then dependencies shouldnt be a big problem either as you could just use a generic package found from a generic library repository.
In short: I dont think the solution is to create pretty wizards, and the like, but to fix the binary compatibility issues, and adhere to a strict file hierachy standard that would allow creation of packages that will work throughout all distributions on a given platform. Then automatic dependancy resolving shouldnt be a problem, no matter what you are using, but as long as binary compatibility is broken every 6 months, this aint gonna happen, install shield like wizards or not.
You could of course statically link all programs, im sure the ram makers would be very happy
Well, if the developers find trying to find an easy way for the average user to install Linux software hard and boring, they can kiss Linux on the desktop goodbye.
To tommy higbee – language is not static, it evolves. When evolving, difficulties arise. There are no set answers, but it will evolve, no matter what.