After AKAImBatman’s wildly popular (and heavily criticized) look at what the future of the Linux Desktop might look like, Mr. Batman has decided to write a followup article where he attempts to clear up many of the misconceptions readers have about the original article. In addition, Mr. Batman uses the second part of the series to give technical details about how some of the concepts might be implemented.
As long as you don’t take individual statements out of context, I think Mr. Batman gets it.
Is his list the only thing that would be required to make Linux an ideal desktop? Nope.
Is his list or any list of improvements necessary to get a novice to use Linux? Nope. I have installed Linux for total novices, and after some tweaking and very minor training they are quite happy with it.
No, I have not tried Ubuntu yet. I’m still waiting for my CDs to arrive due to this blog keeping me too busy to download and burn CDs.
You know what: Linux is multi-tasking.You can do all at the same time.
“No, I have not tried Ubuntu yet. I’m still waiting for my CDs to arrive due to this blog keeping me too busy to download and burn CDs.”
Jesus…
And please stop calling him Mr.Batman, that’s just too childish.
Not sure if I agree with all of his ideas, but one thing is for sure that the current status quo won’t bring Linux into the mainstream desktop. And as long fanboys accept what their told to accept we’ll probably have the status quo for a long time.
I think I’ll repeat this until it sinks down people’s head. When Linux, the Desktop, has a billion dollar per year marketing team it will begin to smell more market share.
Linux ain’t difficult to use or broken, geeks just have a completely retarded view about the intelligence humans possess. Call it hubris.
“Linux ain’t difficult to use or broken”
Well, yeah, it is difficult to use, for like 90% of the people using a desktop, specifically at home.
That’s the problem. And yes, Linux fanboys have a very skewed view as to the intelligence level of your average desktop user.
Your average accountant/HR person/lawyer/salesman isn’t going to want to fight with getting their wireless card to work, or their USB pen drive to be recognized, or their files to be shared with another computer.
When Linux obtains this level of ease (which I personally believe to be help up largely due to software and hardware vendors, not the Linux community itself) it will indeed grow by leaps and bounds.
I have news for you: Everything is hard for the home user. I’ve spent hundreds of hours of my life answering stupid Windows questions alongside stupid Linux questions. The difference is usually that the Windows user hasn’t even tried to consider the problem; or they have and they’ve come to the conclusion that their printer is broken (which is often right it seems).
Even Mac fans have lots of questions: But I think they’re more willing to read, and docs/people are more available to them as mac fans often tend to congregate.
Using *nix with KDE ,Gnome, XFCE or some other modern desktop env is no different to Window$. You just Point and click! (or double click).
My family inc kids, employees and friends have never had a problem with day to day usage such as email, wordprocessing or using a spreadsheet. The only thing they said when first confronted with Linux is “Gee this look a bit different but kinda cool”. It literally took seconds without prompting for them to discover where the apps were and could start being productive. They did have to ask questions though, but then they were used to saving to a c: drive.
I’ll admit that it does take some expertise to set up a good Linux/BSD based system with all the apps and drivers installed and configured but once that is done it just works. It seems to be that way we deploy these systems which seems to be at fault. Systems should be sold with a full compliment of applications/services/plugins which are configured and ready for use. Just add user accounts. We should forget trying to convince people to install Linux etc. from a CD and sell/rent Linux appliances.
Some distros offer such systems. The problem is, distro X doesn’t support your modem, distro Y yes but doesn’t support your graphics card, distro Z supports both but lacks support for…
Nowadays one of the major issues in choosing a distro is knowing whether or not does it support every piece of hardware in your computer. Sometimes it’s hard to tell.
“When Linux obtains this level of ease (which I personally believe to be help up largely due to software and hardware vendors, not the Linux community itself) it will indeed grow by leaps and bounds.”
That’s “held up”, not “help up”
=)
5 years ago we were hearing the year of the linux desktop, but nothing has really changed all that much. The numbers really haven’t increased that much. There isn’t much in the way of desktop standards (even with freedesktop.org), every little distro does their own thing.
If there’s not going to be standards on the desktop then you might as well try something different like Gobolinux coupled with autopackage.
Autopackage
Gnome 2.0
KDE 3.0
GUI Installers
Easier GUI Installers
Windows Mockups (XandrOS)
Crossover Office
WineX/Cedega
Yep, nothing’s really happened in 5 years.
Think the guy needs to show a little more modesty – what’s with “I this…..” “I that….” “I the other….” as if he’s discovered America – many of us have been pointing out/arguing about 90 % of the points he raises on a plethora of Linux forums over the last 3-4 years…… I must have seen on around 250 or more separate occasions complaints about “app install difficulty….” “FSH complaints in relation to the familiar Windows folder layout….” and so on and so on……. trivial points really that we’re all well aware of !!!
To be fair to him, he’s fleshed out a number of points reasonably well – however, he could show a little more modesty and recognise that most of his points have been brought up on hundreds of occasions over the last few years…….
I agree, he seems to know a great deal about what he’s talking about. However, he seems to think he’s the worlds greatest developer (who obviously doesn’t need to develop but just tell the rest where to go).
I think people forget how much has improved in a couple years. Of course, it’s still as convoluted to figure out what distro you want to start with as it ever was; heh.
“Users don’t read the manual before they drive a car, microwave their food, use a refrigerator, or even use a Desktop computer.”
You can’t drive a car without supervision or a license; therefore you’ve been read a manual (allbeit in a slightly untraditional way).
Anyone who sets up a refrigerator better know what they’re doing. There’s no such thing as “using” it. Opening a door simply does not count as use; besides you were taught that in your childhood as well.
Most people don’t read Microwave manuals; until you get one that wants an odd combo of buttons to enter a time (two of mine want you to press cook before you do anything, a manual was read and not by myself).
Ever met someone who doesn’t understand anything about their dvd player they just got and doesn’t consult the manual to plug those component cables into their tv? They probably spent a long time and were eventually told how to do it; or just got lucky and noticed the manualesque colorization
.
People read manuals. The trouble isn’t that they don’t want to read docs; it’s that on every operating system those docs aren’t both easy to find and complete (Windows help is pathetic, there’s no semi-advanced topics, if it doesn’t work normally you’d better call support). On Linux you have to know about man, and likely man -k; or google. Google of course helps every OS, especially linux and Windows (I’ve found most Mac help sites to be very wanting in the same way Windows Help is).
Appdisks won’t work. Well, actually they will; in fact it’s a great idea and I love it (seriously I do!). However, it won’t do what he wants. Because packages are interdependent (if you did appdisks for applications and not libs it’d work, but you’d still have the lib problem which is the real problem anyway). It’s a cool package idea, but it doesn’t really solve any problems that stable package managers haven’t fixed: Making sure no package ever steps on anothers toes by tracking every single file. Then only other package managers (Make) can step on toes: And since fBSD works so well without making rediculous file modifications (usually just compatibility modifications, not changing filenames) I don’t think this is a high occurance problem either.
The cool part of appdisk would be that it’d really make the package manager sort of dissappear from user site: You could hide it very well. Of course, to some level, autopackage does this. All you have to do in it is make the package executable and it installs (even installs autopackage if it needs it). You’d probably end up doing something easier for appdisk.
Of course, appdisk would provide incredible developer freedom to not be careful about filenames. But I think everytime someone’s found a way to “save developer time” we’ve discovered a new way that not being careful messes up your plan (or usually, lack there of cause you thought you didn’t need one anymore).
I think the way we should look at this is AppDisks not replace but augment package managers. By including the necesary libraries in the AppDisks, it would make the program distribution neutral.
So a developer for a program that is not included in many distros could just do one easy packaging to be cross-distro.
You could even use this to cut down individual programs each distribution maintain. They could limit maintained packaged programs to:
1. System programs/libraries
2. Packages that are affected by Distro tailoring (or maybe make an alternative AppDisks)
You then have a cross-distro AppDisks repository on the internet and Distros include those they want for their install CDs
You already can do that with autopackage, and you don’t give up two things that appdisks lose:
1.) Load speed: Moving into a tar and running things isn’t as fast as just opening; not to mention you can’t prelink (I suppose you could, with a nice patch to prelink but that’d kill the KISS idea).
2.) Shared libraries: If you put the libraries with the app, unless it has some say: Try to use shared libs, then use local shared libs; then you’re going to have the same annoyance that techies have with Windows: They don’t want the png decode in memory 18 times.
Of course, if your appdisks don’t include their libs then it’s a cool package idea. It reall is. It’s little more than appfolders in my opinion, but I think it’s a cool way to build on the appfolders idea.
He should write it.
It’d be even cooler if it was built so that it could be user-level and source or binary. If it was source, but automatic building, it could really be multi-platform with shared libraries: That’d be pretty neat.
Don’t fucking use Linux!
Go away.
Just don’t write idiotic reviews without any experience – go and tinker with your Mac or Intel.
Do you contribute with code?
No? So shut da fuck up.
Why dumb down a tried and tested operating system just to please the stupid housewive?
Again *buy* a Mac, *use* Windows.
Just fuck off away from here – unless you have solid experience with Linux; then only you have a voice.
When I join a Karate club, I don’t go about teaching the Master how to fight – do I? Linux is here to teach you about computing – the user shouldn’t dictate the other way round.
Again much like art-critics that can’t paint fuck all only criticize. We are getting these mediocres writting the most inane articles.
Save your time for something more constructive – learn Python and contribute code, idiot!
Yes Linux is a server operating system,
Yes Linux is a kiosk desktop system – it is meant to be stable and robust, and administered by someone with experience (who will gladly install your USB/Modem/ADSL/Printer/etc) …
it is *not* and should *not* be for the Joe Average to administer.
Joe Average is still as thick as paint
Joe Average and Housewives still click on attachments.
How far can you dumb down an operating system (at the expense of many security issues) just to please the thick masses?
The distinction between administrator and user should be set.
Easy to use yes. I’ve installed 3 Linux Desktops on absolutely clueless users. They are not finding anything hard, much on the contrary – very easy to use, stable and fast.
If they need anything extra I am a phone call away.
That is how it should be.
The “personal” and “easy” computing created all the security havoc you see today: worm, viruses, trojans, spyware, etc.
Do you fix your own car or do you employ the help of a mechanic?
Don’t dumb down a solid operating system – let it evolve naturally. If you feel strongly about an issue – contribute with code, documentation, even artwork with do – do something yourself as well.
Don’t try to make the Desktop Environment more Windows/Mac like than it already is – the price to pay will be too high.
Try converting a frame of mind thats locked down into the windows world you’re not gonna win. It’s like the freaking Intel vs. Amd frame of mind where A. The stock motherboard has no advance settings thus easy to use. Thats why schools love Intel the savings on the money the ease for the techs. Last school board IT dept I’ve worked at flat out refused to listen to my proposal because the techs where set in their ways.
Note even with the backing of the Head Administrator and resident CCNP i couldn’t win my arguement.
The excuse was Amd is to hard to set up too many options in ones bios. Same as to why they’ll never switch to Linux with all the Intel and Microsoft endorsements. The only linux box they have is for the SAM which is obvious I was doing the up keep because the 2 admins fergot their Linux.
Sad to say this but if all 3 where right out there with the same marketing cash flow at the very beginning, likely the os world would of been quite different today.
Sad to say this but if all 3 where right out there with the same marketing cash flow at the very beginning, likely the os world would of been quite different today.
Actually, if there would have been more efforts put into the desktop after XFree got ported to Linux (early ’92) then things might be different. Remember, this was when people were still running win 3.1 and DOS.
But your comment infers that there could have been some big company behind linux from the beginning. But there has to be an incentive for businesses to do marketing. Even now you have RedHat and Novell executives not sure of the viability of the linux desktop. Just last year some RedHat VP was saying that it doesn’t make business sense for many people to switch to a linux desktop and RedHat has been wishy-washy about desktop efforts until recently. And obviously the vast majority of revenues for RedHat and Novell are for server installations.
As long as there is the status quo mentality, things aren’t going to change.
I think he has some good ideas for the linux desktop. Is it ready today? Thats for you to decide. But some of the features he’s talking about do seem like a pretty good thing to implement in a future distro.
I agree with you totally just that if you look at the hardware market their is the need to thrive on developping technologies that are needed asap. We look onto the Software side required to use this new hardware. A. We are locked into Windows world. B. Vendor locked into buying the whole package and C. We go the route where a third option is available. D. Someone writes up the OS to use the certain features of hardware.
A clear cut plain example i took my mother to hospital for a CT they we’re running on Windows 2k network between all the machines and they’d show you results on 19 inch LCD’s tilted 90 degrees. Seen how that day the whole infrastructure was down the whole apmostphere felt like we went back to the dark ages of machine code.
Case in point Vendors are forced to side now with the dominating market. This shouldnt be at all. In the end its all to earn a quick buck.
There are so many articles and/or blogs on the internet about “Linux not ready for the desktop”, or “what Linux needs for desktop acceptence”, or “The Linux Desktop of the future”.
All of these keep saying that Linux is not ready, and lacks this or lacks that.
But real life evidence suggests otherwise.
First, Linux, according to IDC, is now ahead of MacOSX (as of last year, actually), and that’s not counting all the downloads and disc sharing that goes on, only units shipped.
Second, there is tons of anecdotal evidence that Linux works great for average users. There are so many testaments to someone installing Linux for their Mom, neighbor, friend, Dad, cousin, wife, or whatever, and that person using Linux quite easily and happily.
Third, cheap PCs with Linux pre-installed, like the ones with Linspire and Xandros sold at Walmart.com or Frys, are selling quite briskly.
The major problem is perception. People still think that Linux is only for geeks, in spite of the existence of desktop oriented distros like Linspire, Xandros, Mandriva, and Mepis. Also, people keep saying that this will be “the year of the Linux desktop”. This is flawed because there will be no “year of the Linux desktop”. The progress of Linux on the desktop has always been, and always will continue to be, evolutionary rather than revolutionary. In short, it will always be steady and gradual.
I agree, he seems to know a great deal about what he’s talking about. However, he seems to think he’s the worlds greatest developer (who obviously doesn’t need to develop but just tell the rest where to go).
While I didn’t notice him being modest, I didn’t see him being arrogant either.
He does lay out the technical issues in detail, and unlike many articles that talk about ‘what Linux fails at’ he does list specific ways to address the problems he sees.
If he didn’t provide any details on how to solve these problems, he would likely be blaimed for not knowing WTF he was talking about.
Yet, by showing he does now he’s labeled as not being modest.
I think people forget how much has improved in a couple years. Of course, it’s still as convoluted to figure out what distro you want to start with as it ever was; heh.
Yep. Fantastic improvements. While I am mostly in the ‘Linux is as easy as any other OS’ camp, it doesn’t mean that it is ideal yet. My biggest complaint is the seperation of the file system and shell from what is shown in the UI — under any X window manager or desktop environment. So, you could say I both agree and disagree with him as my emphasis is slightly different.
Batman has provided detiled and helpful comments — especially in section 2. He is attempting to show how to make things better, not just that things currently suck. That’s a good thing, even though personally I like Linux quite a bit right now.
Do you — and others — have constructive comments?
If you do, it would be good to hear them.
It is easy to criticise, though Batman has stepped forward after being criticised before with what seems to me to be a good and thoughtful response.
(He does reply to a few people in the first section, but he is not very hostile and his replies are well reasoned. He has been called worse — such as arrogant — in this thread alone so even if he were hostile, it would be somewhat understadible if not entirely acceptable.)
I made several constructive comments, thanks for ignoring them. I also complimented him repeatedly.
And Klik (automated building of customized cmg files with serverside-apt) is even cooler, but I’m not at all convinced that is really offers many advantages over existing practices which have a much lower overhead and therefore far more chance of being widely used.
The status quo, centralized distro specific apt or yum or whatever repositories, works fantastically well for FOSS applications. Where centralized repositories fall down is for distributing binary-only commercial applications. Mounting disk images a la cmg is a nice solution, but what advantage does it offer over simply providing a tar.gzip’d folder which contains all or most of an applications dependencies?
A simple folder with “inline” dependencies can also be easily integrated into a distro existing package management solution. For instance, I would add an entry to my sources.list, update, and install foobar, a meta package consisting of individual packages for the application and each of its supplied dependencies, which would then all install under /opt/foobar (dependencies, config files, documentation, etc.. going under /opt/foobar/bin, opt/foobar/lib etc…). Done this way, when I run “apt-get update && apt-get upgrade” the dependencies for foobar would get upgraded as well.
Those running a distro with out a packagemanager could simply download foobar.tar.gzip and unpack it to /opt or /usr/local or wherever, and then add the approprate path entry if desired.
Of course, all of the above could easily be wrapped in a user friendly delivery/installation mechanism, such as a file on a website like foobar.install–containing the repository info for apt, yum, whatever–and a script could be set as a mimetype handler for text/install which would launch a little gui dialog that prompted for the root password (or alternately allowed the package to be installed in user’s home directory), pop up a simple progress dialog, and finish off by adding a path entry or creating a symlink to /usr/bin.
An example of what I’m talking about would be XAMPP (http://www.apachefriends.org/en/xampp.html) which rediculously easy to install on both Linux and Windows.
Funny, I was just thinking about a similar idea today, how to push package managment upstream with the option of forking downstream.
For namespace I think the URI standard is enough, coupled with HTTP we get a nice way of keeping a “HEAD” URI of any branch we whish to keep.
I was than thinking that we could have some RDF based language to create files that would be retrieved when GETing those URI via HTTP.
The most flexible way of expressing relationships between version will probably be a simple href to the previous version.
I think we have here a real world application, that is actually implementable, for the semantic web.
F.ex. relationships with VCS systems and how to retrive the source could be included in the same file.
Oh well, someone has to do some serious research to get this going though. Might be me, but don’t count on it ;-P
20 years from now, people will be using different operating system so Linux doesn’t have all the time in the world to make an impact on the desktop. Maybe linux is fine for the server and other operating system are more suited to the desktop. there is nothing wrong with that.
20 years ago people were using Unix. Unix is still in use I remind you, and mostly died off because of its proprietary multiple personalities.
His followup consists of pointing out faults in posters critiques of his origional paper, this is no hard thing todo as stupid posters are innevitable. I really not think the entire topic is that revolutionary, we all know Linux on the Desktop has a ways to go, we also know about GoboLinux and the fact most people don’t understand the unix history.
His followup consists of pointing out faults in posters critiques of his origional paper, this is no hard thing todo as stupid posters are innevitable.
You do know that there’s a second part, right? Click on the link that says “Go to Part 2 >>” for the rest of the article.
Do you guys all really think gobolinux is a good idea?
Linux wont work on my ’emachine’. Ubuntu didn’t recognize my integrated graphics or network card.
Linux isn’t ready because there are too many distributions and all have slightly different kernels.
This is the achilles of linux. NO HARDWARE SUPPORT. No winmodems.
Pull your hair out is what it’s good for.
I will stick with retard Bill’s operating system. Just don’t download any P2P apps. Use a FIREWALL !
This is the achilles of linux. NO HARDWARE SUPPORT. No winmodems.
There is some. But you ware correct, that is the biggest problem.
I agree. If you are a geek with lots of spare time, you could get most hardware working under GNU/Linux. But most people aren’t, so Linux isn’t an option for them.
There are a lot of efforts being made in order to improve installation, application installation, GUI… I’d like to know what kind of efforts could be done to improve hardware support. Maybe the only way is to spread the software widely enought to attract manufacturers develop drivers for Linux?
Depends what distro you show. Doesn’t it make sens to show desktop distro like Linspire instead of Gentoo which is really hardcore?
How do I observe this? Well, I installed Linux at my mother’s PC 3 years ago. She only had experience with windows previously, but guess what – she had NO trouble with Linux. I had to nurse her Windows system once a week on average, smoothing over mostly system wierdnesses. Now I have to nurse her system once a month on average, smoothing over user wierdnesses.
A system with all the gadgets discussed by Mr. Batman would not change much for my mother, nor would it change much for me. I simply would have to learn the new things, my mother simply would not get bothered with any feature she cannot see.
I don’t know exactly, but I guess for my mother the DBFS idea would not be that good an idea, because this could result in ONE spreadsheet document (e.g. holiday accounting) beeing displayed in TWO folders (the spreadsheet folder and the holiday folder), which can be quite confusing for her.
She is more used to the physical way of seeing things, the real paper scissor is in ONE place only, and does not magically appear everywhere a sheet of paper is placed in her flat.
For total newbies it is totally irrelevant if they learn Windows, Apple or Linux. They have to understand the basic concepts of COMPUTERS, and then it is more a question of learning by trial and error which is more confusing in Windows than in Linux (repeatability of error messages in Linux is higher).
Linux TODAY already is equally or better usable than Windows, the two things slowing widespread adoption of Linux desktops are pre-installed Windows boxes and inertia of computer users.
Looking at most of the the positively scored comments here, I see quite some astroturfing going on.
These comments are of the sort of “yeah Linux is OK, but <fill in percieved flaw>”, which are then promptly followed by a reply of the sort “you are not entirely correct, but in principle you are right”.
These comments are designed to build up a picture of Linux in your mind which makes you believe that Linux will likely not run on YOUR machine although it undisputably runs on the machine of XYZ.
Fact is, I have not yet found a machine on which Linux does not run, and the usual peripherals are supported quite well. MOST machines run Linux perfectly.
I have a 1400 MHz Athlon, 512 MB RAM, 110 GB harddisk, a DVD burner, a DVD drive, a HP Deskjet 640C printer, a MICROSOFT
explorer mouse, an ATI XPert2000 grafic card, a WINTV BT787 TV tuner card, a camera, a mp3 recorder and a headset for skype.
All this zoo of peripherals works perfectly and is correctly auto-detected in Linux (Mandrake). Everything but the tuner part of the TV card is even auto-configured.
In Windows which I keep for playing games the grafic card driver and the TV card driver are at war, so I have to leave the tuner card uninstalled.
“Except for ReiserFS. I considered the possibility of Reiser solving the DBFS need, but I found its meta-data support to be incomplete.)”
to my understanding… isnt reiser modular? and just because its incomplete now doesnt mean it can be amended? I think the future is reiser 4 and autopackage.
He said there would be performance issues but they would be “Negligible” with modern hardware. If reiser is the fastest, and can fix its issues, why not get the best of both worlds ?
Here is what i would like to see: http://autopackage.org/ui-vision.html
Windows wont work on my ’emachine’. Windows didn’t recognize my integrated graphics or network card.
Windows isn’t ready because there are too many versions and all have slightly different kernels.
This is the achilles of Windows. NO HARDWARE SUPPORT
Shut da fuck up:
http://www.linuxant.com