“Has it managed to completely escape the attention of the “open source” movement that Adobe, Macromedia, Corel, and so forth have blithely continued to remain virtually Windows-only while waiting for the dust to settle? Only now they have realized that it won’t settle and oh-so-quietly the rush of announcements of support for Linux has not translated into a rush of quality applications.” Read the editorial here. I’ve written a similar editorial a few years back.
Divide & Conquer: The Demise of Unix
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Most Linux people that I talk to want free software, me included. I don’t know if they would buy Adobe for Linux even if it was out. I would not be surprised if these companies have done market research that found that they would not see a return of investment, building Linux versions of their software.
I need Adobe & Flash, so I have my company buy it and I run it on the XP workstation they have provided for me.
Did anyone notice the date on it? Maybe he knows something we don’t?
Arguements over different distros is something I tire of personally, and I especially hate it when config files, or any other files are in a different place to others. Isn’t the LSB meant to be about trying to solve all these sorts of issues?
One example of a pet hate is distro specific config tools when, to me, it would better suit all our needs if they were made as contributions to (as a minimum) kde, gnome or whatever other desktop environment the distro supported.
I just do my best not to learn things that are distro specific anymore. I’d much rather have skills and knowledge that applied to more than one distro, ideally ALL of them, with minimal variation.
It’s the applications that matter, not the distro, surely?
If all distro’s would offer vanilla packages (slack does) then the differences between distro’s would be less dramatic. Many slack-based distro’s have a very “natural” look and feel (but lack bells and whistles).
Initiatives like x.org and LFS are great.
The war between KDE and Gnome is kinda worrysome though.
As a UNIX system administrator with 20+ years experience, and a Windows system administrator since Windows 1.0…
Was anyone really an admin for Windows 1.0 systems? It wasn’t even a useful system until 3.0.
customers didn’t choose Windows because it was better (or worse) than UNIX; they did it because Microsoft/Intel was careful to guarantee them a consistent software experience across a broad selection of hardware
Not at all. Microsoft/Intel was not careful to guarantee a consistent software experience across different hardware. Microsoft doesn’t run on that many architectures and Intel couldn’t care less about your experience on other hardware.
Equally important, application developers flocked to that consistent software experience because it meant their products were cheaper to develop without the headaches of version-specific differences.
Try telling that to the people still stuck on Windows 3.1, 95, 98, NT, etc, only because they cannot migrate their existing software to XP.
Commercial grade Windows software just works and usually keeps working
As I stated before, this is not the case.
I installed Linux on one of my systems the other day, so I could use it as a teaching vehicle for my class on system log analysis. But first I had to Email a bunch of my friends and ask them, “what version of Linux should I use? Red Hat? Debian? Gentoo? Mandrake? Slackware? Do you think I could get away with OpenBSD or FreeBSD?” The responses I got indicated that none of my friends use the same thing but that I could be sure that if I used Flavor X some adherent of Flavor Y was going to bust my chops about it, and that someone was sure to show up with flavor Z and have trouble making things work.
You’re being a little overdramatic, or your friends need to get out more. No one that matters really cares what distro you use and the differences really are not that great. What if you had people show up with different installations of Windows, or even just different .dll versions because of the software they have installed? Surely you would have just as many problems, if not more.
Do you hear the sound of distant laughter coming from Redmond? I do.
Only if it’s nervous laughter.
Has it managed to completely escape the attention of the “open source” movement that Adobe, Macromedia, Corel, and so forth have blithely continued to remain virtually Windows-only while waiting for the dust to settle?
That will be the death of them too. While they continue to do nothing the Open Source Movement will create products that rival or better current proprietary solutions and when those commercial offerings are finally available on Linux it will be too late. Who’s laughing now?
Is it too late to save the situation? Yes, I think it is. At this point there are too many adherents and investment in “not invented here syndrome” in the various free UNIX flavors that nobody is going to come to their senses until there is only one man left standing.
This is clearly not the case. Linux and the Linux Desktop are getting more standarized, not less. The LSB, freedesktop.org, posix compliance, and even things like autopackage are pulling these projects together.
They just sat back and watched free UNIX fail to become a credible threat because, well, frankly, it was in the hands of egotistical, detail-oriented amateurs.
Your harsh judgements are unjustified. You act like this same bickering does not happen in the commercial world. Do you know how many former employees of companies have quit/been fired over personal difficulties (can you say Oracle and PeopleSoft) and started their own rival company. I also find it amusing that you can refer to the people inolved with Linux as amatuers, considering most of the important ones work on their projects professionally for large corporations.
I think the Redmond cavalry deliberately stirs the pot. If them Native Americans ever unite, they’ll push the Europeans back into the Atlantic.[1]
Switching metaphors, the Redmond hare stays ahead of the superior penguin (in turtle drag) by keeping it fragmented.
Because 75 different X-desktops/editors/version systems is a good thing!
But we know who wins the tortise/hare race.
[1] Historical update: s/Atlantic/casino/
This is a poor rant that happens to agree with Eugenia’s opinions. Instant headline.
The man behind it writes lots of similarly ill-informed rants, including one in which he makes the classic audiophile mistake of assuming that the real world isn’t quantised or subject to sampling limitations. Surely, he argues, this 35mm negative must have unlimited data inside it waiting to be unlocked, and is thus superior to any digital image.
He writes that he’ll have a “nice 837727626 x 237362326 pixel 1024-bit color image” in 2020, made from this same negative — and so he might, although he’ll have to write custom software to create it and it still won’t look much better than today’s digital images. Why? Because sampling isn’t a linearly improving technology, below a certain threshold your signal is overwhelmed by quantum noise in your detector. Since we can reasonably assume that his 35mm camera doesn’t have a liquid nitrogen cooling tank, and his 35mm film isn’t made from some as-yet uninvented grainless substrate, the most he’ll actually get is ~20Mpixel at 16 bit per channel, and even in that image he’ll be able to see a lot of noise introduced by his choice of low quality analog media. The noise will not have improved in 15 years, in fact it will have got much worse.
He makes the same argument for lossy image compression and for color displays (Bzzt, quantum noise problems again, we can’t build better than 12-bit per channel displays and there’s no sign of that changing in 15 years). So he doesn’t understand the physics OR the technology, what hope is there for him?
Adobe, Macromedia and Corel provide poor support for Windows, and abysmal support for everything else. They’re not very good at what they do, but their customers (notice what kind of customers all three companies have…) are suckers and keep coming back for more. This has less to do with the UNIX wars than with his mis-directed anger at their shoddy proprietary software.
The guy either has a crystal ball or leaves in the future. In the latter case I don’t know how we are reading the article dated Dec 4 2005.
Why is “open source” in quotes?
I use Linux at work and OSX at home and if there is one thing that OSX beats linux at hands down its consistancy, sure its look and feel is nice, xcode is cool and all the other nice features are good but there are free versions that are as good if not better, but the shear inconsistancy of linux keeps me on OSX, where worrying about library versions, different look-n-feels (how many different widget sets does one OS need!), kernel versions and patch levels, and learning a thousand and one incompatable config file formats just to get your server up and running is very much the exception (you still have to do it because OSX uses so much free software!).
Maybe I’ve just grown out of the stage where spending hours fiddling around in the depths of some obscure software trying to get it to work was fun.
Yeah, I agree with you there – I’m feeling the same effect happening to me – I don’t wanna fiddle around anymore, I want anything that just works – and maybe is even pretty up-to-date and usable if I don’t update the base system (kernel) every 6 months or so.
I don’t have experience with Mac’s, but from all I’ve heard, it seems like the solution to me. And more than just working – it is cool.
But the troubles of linux were too much for me too. Tried it out 1 year or 2, from time to time, but switched back to my laptop’s XP some months ago, and I’m happy again. ONE Distribution *g*, Consistent look & feel, snappiness, a whole lot of proprietary AND free apps etc. Linux is a Unix Clone, but where is the Windows Clone? ReactOS? I guess I’ll just check back in 2 years or 3 (for Linux + ReactOS).
The comparison to the infamous Unix Wars are just plain wrong. First off there are only 2 major players on the linux market, with a lot of small factions also but going to them would be the customers choice and his responsability. This makes the linux market very much less fragmented than the Unix market was, coupled with a higher (near 100%) level of compatibility.
There is also no lock-in anywhere near the magnitude of what there was (and is still) in the unix world. Moving from Red Hat to Suse to Slackware is almost trivial (and low cost) compared to going from Solaris to AIX.
Oh and there’s always plenty of laughter coming from the asylum. Windows on the server is wedged between traditional unixes on the high end and growing linux market on the low end. I don’t see them dominating that market anytime soon.
Sure users still expect to see Windows on anything with a user interface (pc/pda) but open source is improving everyday and has so much momentum it will at least provide a challenge to Microsoft.
Also why was Cheapskate’s comment modded down ? It was funny and summarizes the article nicely IMHO.
I installed Linux on one of my systems the other day, so I could use it as a teaching vehicle for my class on system log analysis.
This has to be a joke. They have classes for that???
People love market leaders. It gives people a comforting feeling. In the networking world, it is Cisco. In the desktop world, it is Windows. In the database, it is Oracle. People like having a safe choice, having something where there will be lots of opportunity for job growth and availability. Unfortunately, in the Linux world, there is way too much choices and way too little majority. There are no safe choices. Should I learn bash, csh, or new super shell Y? Should I learn Perl, Python, or scripting language Z? Should I learn ext3, ReiserFS, XFS, or JFS? Lilo or GRUB? Should I learn rpm or deb? Apt or Yum? Gnome or KDE? Linux, FreeBSD, or Solaris? Sendmail, Postfix, or Exim? Fedora, SuSE, Debian, or Gentoo? A healthy IT market has a clear market leader (60-70%) with a lot of smaller companies each making 10-20% of the market each. In these markets, the little companies innovate, and market momentum allows the big fish to catch up. Everyone stays happy because whatever investment people have made don’t disappear.
Linux is just a kernel. RedHat, Gentoo, Debian are the OSs that are similiar but each do various things in a different fashion.
There’s no standard toolkit still. No wonder Adobe and Macromedia don’t want to write software for Linux.
I wonder what things would be like today if there had been a meta-distro that all other distros are based off. If we had a standard toolkit, standard app interop….
I believe things would be better if you had more standards at the lower end of the software stack, but NIVHS and unpaid contributors have no incentive to cooperate. Slashdot proclaims 2000 year of the linux desktop, but there hasn’t been this huge groundswell to switch over 5 years later.
I wouldn’t say it’s all FUD. The author has a very valid point, that one of the main reasons other (arguably technically inferior) platforms are preferred to Unix or Linux is consistency and compatibility. But I wouldn’t outright dismiss Linux or open source software either; Linux is growing in maturity, and many computer illiterate types use distros like Suse, Fedora or Ubuntu without problems.
Now, the “distant laughter” bit is funny, but of course not true: Redmond obviously takes Linux very seriously.
Yeah, choice and competition is killing UNIX. What we need is for everyone to run exactly the same OS and then software writing would be a doddle.
A good piece; very thoughtful. Highlights the reality that commercial software vendors will not write for Linux so long as no single commercial distribution dominates. Since just about any schoolkid can crank out a “free” distribution (and walk away from it when he gets bored) that will never happen: Linux will always be strewn with dozens and dozens of essentially identical and purposelss distributions.
If Linux people really want commercial vendors to write for Linux, they’d drop the free distros and all run out and buy RedHat or SUSE or Novell.
If Linux people really want commercial vendors to write for Linux, they’d drop the free distros and all run out and buy RedHat or SUSE or Novell.
A quick web search :
From the Microfocus site : “Micro Focus Application Server™ is a set of powerful run-time facilities that enable COBOL applications to run on all leading Windows, UNIX and Linux platforms”
From the Oracle site : “Oracle’s Linux commitment began in 1998 with the first commercial database on Linux. Today, Oracle is the only major software vendor to collaborate with and provide first-line support for Red Hat and Novell/SUSE, all Oracle products are available on Linux, and Gartner Dataquest says Oracle is #1 on Linux with 360% growth.”
From the IBM site : “The combination of WebSphere and Linux gives medium and large-size companies a highly reliable, open-standards-based platform for extending Linux-based applications across an IT infrastructure — to run on Intel® processor-based servers; more robust POWER4 servers such as pSeries and iSeries; all the way up to the zSeries™ mainframe.”
Should I go on ?
is when IBM, Oracle, or an ISV provides commerical products for Linux, they only support RHEL and SLES. Does this stop you from running, oh say Oracle 10g, on Gentoo, Debian, etc? No, you just dont get support. When Linux becomes more predominant on the Desktop, the same will occur. ISVs only targeting *major* commerical Linux distros. In the future, if you want to run Photoshop on Linux, your distro will have to be compataible with the major commerical distros. Your distro is not standards* compliant, no commerical software for you. Simple as that.
Linux fragmenation is overstated.
*Modular LSB, FDO Platform, etc.
OK,
I have run FreeBSD since 2.2.2 as an owner of a small ISP and played with linus since Turbolinux3x. Here is my question to you all who want’s a system that just works. I can download and install Mandrake 10.1 and it will find most of my hardware, and give me a pretty good desktop environment. I just upgraded the harddrive on my Dell Smartstep laptop that is less then two years old. It took me all day to reinstall XP the service packs, drivers, office, adobe etc. I bet that same copy of Mandrake woulden’t take that long?
Now I will tell you that I use Gentoo and FreeBSD because I am addicted to setting up(screwing with)the OS. This is a problem for me and one of the reasons why I can’t use Linux as my work computers now days.( I sold my ISP and am now an Insurance Adjuster)
Buy a mainstream copy of linux, as in SUSE, Mandrake or JDS and update every six months or as necessarry due to security issues. Just say no to the newest version of the kernal!!
> Linux fragmenation is overstated.
Obviously the first thing you will notice is that startup scripts in Linux (debian, slack, RH, SuSE) are all very very different. Atleast with BSD/Solaris/AIX it’s well defined.
Secondly there’s no standards for creating menu items – SuSE puts KDE in /opt/kde/share/foo and others put in /usr/share/foo. Still there are issues with Libraries and locations – Linux/Unix people flamed Microsoft for DLL hell, well it’s no better on UNIX either.
Finally devices – SUSE uses devfs – so your root becomes /dev/disk/controller0/disk0/part1 and Redhat uses /dev/hda or whatever and so now where’s device compatibility?. Finally package management – Redhat and SuSE use RPMS, but when you install XMMS – it’s stuck in a different menu than Redhat.
Sure when you have source and then configure; make; make install will work but what about integration into the system? Like when you install a software using c;m;mi your system has no record of what version is installed.
Now everybody on this board will be quick to retort “hey it’s open source and choice is good”. I don’t know how to answer that!.
That was absolutely, unequivocally, the dumbest thing I’ve read this year.
It is nothing but an emotional rant based on no facts or deep thought whatsoever. Yeah… gnu/linux obviously has failed since it isn’t making billions. It’s only making millions for a lot of companies each, instead of full concentration in one company.
What will you do when your entire theory here goes up in smoke? Apologize? I doubt it. I guess we’ll have to settle for the fact that you’ll probably end up using linux, or the Microsoft version of it when the day comes.
“The sting of a reproach, is the truth of it.” — Benjamin Franklin
With respect to servers, the current model of multiple distros, multiple OSs, works out pretty well. Linux, FreeBSD, and others are highly competitive with Windows in the server space.
With respect to desktop software, it is a serious limitation. KDE or GNOME, or WindowMaker of Xfce, there’s no consensus on which is better and most have a strong dislike of one in favor of the other.
You know what I’d like to see? Apple to release the old NeXTSTEP windowing system as open source. That was a slick, streamlined, and beautiful interface. It worked quite well on 68040 systems running 25 MHz, on 12-bit graphics, with 16 MB of system RAM.
And I don’t see why so many are looking to mimic the functionality of the Windows environment. It’s Apple’s environment that everyone should be gunning for (that’s what Microsoft does, anyway).
This guy was spot on. I think he’s accurately described the hill that the open source community has to climb to get anywhere – as high as its climbed so far. The worst part is that he’s accurately described what has already happened many years ago.
He has an interesting point, but there’s some holes in it. Free software isn’t going to go away even if the vendors do. Commercial *nix systems go away if the business model doesn’t work. Linux and BSD distros are a hobby projects.
In other words, Free software can go through periods of ebb and flow. There could be setbacks and then years from now a resurgence. Techies will always pick their favorite distros that may or may not appeal to the masses. But all it takes to grab a chunk of the desktop share is for one distro to fill the gap at the right time. A lindows or xandros kind of deal. Personally, I could see a $free distro doing the job. Simple, with the right gui config tools built in. Sent out in the mail on cds for free. Users that need support or want to subscribe to a package database could pay a fee. That’s just speculation though.
The point is, Free software could die, but not the way unix vendors died.
-b
…then you probably should investigate that further before you start bashing him for his opinions and embarassing yourselves further. Marcus has well established credibility, something that goes a lot farther than witty comments on an online discussion board.
And yes Rayiner, they have classes for that. There’s a good chance you could benefit from one. Maybe when you get out of school, you’ll come to understand that log analysis in the real world involves more than typing ‘less /var/log/messages’ in a terminal.
One of the limitations of the open source community is they don’t like to hear criticism, no matter how true it is.
Open source is here to say, but I don’t think the author is talking about open source in general. He seems to be talking about desktop environments.
Open source operating systems work great on servers. Highly competive and incredibly compelling, as well as highly successful in terms of usage and competing with Windows.
But what about the desktop? Open source software like OpenOffice and Mozilla/Firefox have seen great success, but much of that (especially Firefox) is sucess because it’s running on Windows.
There just aren’t that many desktop systems running Linux or other open source operating systems. When you remove sysadmin users from the equation and only consider non-IT business and consumer ues, the usage percentage goes down even further.
If you think he’s so wrong, totally off base, and dumb, then why do you think that open source operating systems/GUIs haven’t found the same success running on desktops that they’ve found in the server realm?
for ISVs looking at desktop software for Linux/UNIX there is really only one thing that matters: profit.
the reasons to target or not to target a platform revolve around the profit potential. so, for instance, if there isn’t enough volume represented by a certain OS, very few ISVs will write closed source, pay-per-license software for it.
ISVs need a fairly limited number of things to make a venture profitable. they include a binary stable ABI for the system libraries (in the case of Linux things like gcc, g++, glibc), some basic file system consistencies (the layout of /etc hardly matters for desktop software, though =) and some installation standards. we have only really gotten there with the larger Linux distros in that last year or so.
gcc has finally put together some back-to-back ABI coompatible releases; distros generally agree on the usage of /opt and /usr via the FHS; there are standards for where to put your .desktop files so your apps appear in menus….
now we just need the requisite volume and a few more releases to prove to ISVs that the stability of the necessary components can be counted upon.
Adobe already does make software for Linux/UNIX (Acrobat Reader, for instance) so it’s not like these companies are incapable or unaware. they are just waiting for profitability to be a reality.
this is why things like Oracle’s database runs on Linux, despite it being an even thornier world for server software that it is for desktop software.
we don’t need to wait for everyone to agree on every detail, we just need a stable enough platform with enough volume behind it and it will happen. it has nothing to do with packaging formats or ISVs demanding they be able to support every UNIX-like OS out there or how long it takes to run ./configure.
and no, i don’t think anybody would turn their nose up at Photoshop if Adobe used ToolkitX for their app just because they used The ToolkitY Desktop.
p.s. UNIX fragmentation wasn’t the only reason it was hobbled on the desktop. not only did it not have a compelling desktop experience (which we do have that today), they were also horribly expensive and lacked a commodity hardware market (Linx/BSD/Solaris/etc on Intel provide that)
>> From the Oracle site : “Oracle’s Linux commitment began in 1998 with the first commercial database on Linux. Today, Oracle is the only major software vendor to collaborate with and provide first-line support for Red Hat and Novell/SUSE, all Oracle products are available on Linux, and Gartner Dataquest says Oracle is #1 on Linux with 360% growth.” <<
I find it amazing that our campus, a closed M$ shop for nearly five years, is now scrambling for people within the IT shop to support a *nix-based campus administrative program. Only two people, myself being one, know anything about it and three volunteers have wanted to learn it.
Can you hear the sound of sobbing by those who should have been paying closer attention to possible changes?
Linux distributions are analagous to a democracy with 3,000 parties. Because noone can compromise on anything, and people are so die hard about every little detail, there is very little agreement. The parties all do what they want but there is no unification or majority that can agree to something and then move forward to other issues.
For example, as a software developer wanting a rapid development environment, do I use Gnome or KDE? If I choose one, how do I reach all the die hards who refuse to run the other or either? How do I even know the market is not going to completely change and waste all my development time? And even if I am able to choose a target, where do I put the default install? Do I let the distro choose? If I let the distro choose, what kind of standards are there to pass parameters to the distro and how do I know this distro has implemented it?
Just look at the real world and you can see examples of how important standards are versus technilogical superiority. It will be difficult for anyone to argue that English is superior to all languages, yet it has taken off as the worlds unofficial official language because the world needs something that just works. Look at TCP/IP in the networking world. TCP/IP was supposed to die in the 90’s for a true OSI Layer Network Protocol. Today, all major infrastructure is being built on it. Was it the superior protocol? No, it worked, people agreed on it, and people were free to use it.
Competition is good, but certain things just need to be agreed on and taken for granted.
SGI went with Linux/Itanium, and Intel will pull the rug from under their feet. SGI will die, but not because of Linux, because of Intel trying to save their own skin. Shame.
Obviously the first thing you will notice is that startup scripts in Linux (debian, slack, RH, SuSE) are all very very different. Atleast with BSD/Solaris/AIX it’s well defined.
Granted, but it doesn’t seem to be much better in Windows – IIRC there are something like four different ways of making things start at boot, and only one or two are visible without regedit. Ever wondered just why Nortons starts at boot without being in Startup?
Secondly there’s no standards for creating menu items – SuSE puts KDE in /opt/kde/share/foo and others put in /usr/share/foo. Still there are issues with Libraries and locations – Linux/Unix people flamed Microsoft for DLL hell, well it’s no better on UNIX either.
Okay, but that’s really not a biggie eh. So there are four potential locations of KDE – it wouldn’t take much to code something in to choose one! To use Windows as an example again, they’ve gotten away with changing from WindowsStart Menu to Documents and SettingsUserStart Menu, and everything associated – surely Linux can manage a few options too?
Personally I have no issues with libraries or locations, but I guess that comes from a decent package manager…
Finally devices – SUSE uses devfs – so your root becomes /dev/disk/controller0/disk0/part1 and Redhat uses /dev/hda or whatever and so now where’s device compatibility?
I’ve used devfs and udev, and both had /dev/hda1 entries. Maybe SuSE are just being dumb with it? Who knows.
Sure when you have source and then configure; make; make install will work but what about integration into the system?
See previous comments about package manager – this approach isn’t acceptable for most people though. I’ll wait twenty minutes for something to compile, but I don’t think everyone will.
Back a little more on topic, I don’t think these closed products will pick up huge popularity on Linux, because the Linux devs have spent years developing open alternatives.
Why would I want Acrobat Reader when I have KPDF? I can print to pdf’s using OpenOffice which will probably cover many people’s requirements for Acrobat.
Who needs MSN, AOL or ICQ when there are Gaim and Kopete? And very few people see any need to pay for Photoshop when there’s the GIMP.
Yes, I know some people do – but most don’t want to spend that kind of money. Those that do are a fairly specialised subset of users.
Or the ultimate example – who would want to pay for MS Office (which is bloody expensive if you’re not a student) when OpenOffice is free? Sure, it’s not really as good in many ways, but that’s not going to justify the cost differential in most cases.
Hey,
Just a thought, from my perspective (an admitted moron) a lot of the discord here is the clash over what software should accomplish. I think the difference is those that view software as an income as opposed to those that see software as something their grandson told them about. Most non computer-savvy people I’ve ever met (90 percent of all users) could care less what they’re working on or with. They don’t care about the Unix wars or MS monopolies they want to get email from the kids or get free porn. To these people who don’t care about interoperability and standards whether they run win98, fedora, freebsd is immaterial. In my opinion the quality of an OS means nothing to the average consumer. They want a computer. They get a computer. It has whatever OS on it they learn and become accustomed to it. The fear isn’t to me that OSS won’t win and get the whiz-bang photoshop port… the fear is that it will cease to be and everyone will pay 100 bucks extra for a computer with no choice as to what is running on it. Oh wait nevermind. At least we have the 14 year old punks to fall back on.
It’s kind of funny to compare professional application, like Photoshop to GIMP, or MS Office to OpenOffice.
I’d definately agree that most people don’t use all the features of those apps, but I’d like to know if any professional designers are using GIMP or financial analysts use anything other than Excel for their spreadsheet works.
Finally devices – SUSE uses devfs – so your root becomes /dev/disk/controller0/disk0/part1 and Redhat uses /dev/hda or whatever and so now where’s device compatibility?
I am not doubting this is not a real situation, but device files are standardized by the kernel people; specifically the LANANA http://www.lanana.org (Linux Assigned Names and Numbers Authority) or devices.txt in your kernels Documentation directory. SUSE is incorrect for just using /dev/disk/etc…, if that is true.
Even though we will probably never agree on a standard package format, or atleast one transparent to the user and “just works.” I think the LSB is doing ok, and should atleast propose a standard package format, and format for init scripts.
Unlike in the proprietary software world, non-compliance dosn’t nessacarly mean stuff won’t work, it will just mean you are non-compliant.
In other words, customers didn’t choose Windows because it was better (or worse) than UNIX; they did it because Microsoft/Intel was careful to guarantee them a consistent software experience across a broad selection of hardware. Equally important, application developers flocked to that consistent software experience because it meant their products were cheaper to develop without the headaches of version-specific differences.
uh huh. and the price point of the wintel TCO vs virtually everything else had nothing to do with it?
his arguement may stand up if you completely ignore the mac, which offered a significantly higher level of consistancy then microsoft and unix put together. the reason that microsoft won the os war is that it is damn near impossible to dominate a market by selling quality when there is a substancial price difference. this doesnt just apply to operating systems either, there are a great deal more honda civics on the road then mercades benz. i dont think anyone has any doubt to which is the better car, and i dont think anyone would really question why there are so many more civics. and anyone using marketshare to justify quality in this case would be laughed at.
If Linux people really want commercial vendors to write for Linux, they’d drop the free distros and all run out and buy RedHat or SUSE or Novell.
A quick web search :
From the Microfocus site : “Micro Focus Application Server™ is a set of powerful run-time facilities that enable COBOL applications to run on all leading Windows, UNIX and Linux platforms”
According to: http://supportline.microfocus.com/productreleaselevels/unix.asp I see they support SUSE and RedHat.
From the Oracle site : “Oracle’s Linux commitment began in 1998 with the first commercial database on Linux. Today, Oracle is the only major software vendor to collaborate with and provide first-line support for Red Hat and Novell/SUSE, all Oracle products are available on Linux, and Gartner Dataquest says Oracle is #1 on Linux with 360% growth.”
http://www.oracle.com/technology/tech/linux/htdocs/oracleonlinux_fa…
SUSE and RedHat (and Asianux). I quote:
“Oracle does not want to support fragmentation in the Linux operating system market. There are an indeterminate number of Linux distributions in the market. Customers have consistently asked for stability, better performance, and reliability of the Linux platform as well as enterprise-class support. Oracle cannot be effective in servicing customers if we attempt to support a large number of different Linux distributions.”
From the IBM site : “The combination of WebSphere and Linux gives medium and large-size companies a highly reliable, open-standards-based platform for extending Linux-based applications across an IT infrastructure — to run on Intel® processor-based servers; more robust POWER4 servers such as pSeries and iSeries; all the way up to the zSeries™ mainframe.”
http://www-306.ibm.com/software/webservers/appserv/was/requirements…
Linux for Intel: RedHat, SUSE.
Linux for zSeries: RedHat, SUSE.
Linux for pSeries: RedHat, SUSE.
Linux for iSeries: RedHat, SUSE.
Should I go on ?
Yes, please…I’m on the edge of my seat.
As many here are fond of saying “RedHat != Linux”, and while in the community that may well be true, in the commercial sector RedHat == Linux.
When companies say they “run on Linux” it’s just like saying they “run on UNIX”, but what they mean is that they support and run on specific implementations of “UNIX” and “Linux”. They run on Solaris, but not FreeBSD, say. They run on RedHat but not Slackware.
“Linux” is more buzzwordy than “RedHat” (much to RedHats chagrin, but they’ll take it either way), but when you look at the man behind the curtain and ask them what they mean by “Linux”, they mean RedHat and SUSE (Novell).
[i]Today, if you want a consistent software experience, you have little choice but to go with Windows[i]
Is this a joke? This guy must be on drugs.
This guy doesn’t realize that all apps are turning into Web Services. Linux makes the *perfect* platform for serving this *new* application paradigm.
Linux’s strength is in the server and database support *example. Oracle server farms using Linux clusters*. I don’t see the linux desktop becoming more than a hacked up version of Linspire or a really slimmed down embedded Thin Client with a web browser.
The Perfect example is Google and what they are doing with their custom in-house Linux environment.
Good post Slash
It’s good to see the majority of the comments contradicting the usual groupthink drivel as if there is never, ever anything wrong with Linux.
I’ve said it a thousand times, and will probably say it a thousand more. There’s got to be standard toolkit. Freedesktop is doing good stuff, and dbus will eventually handle the IPC stuff, but you still need a standard toolkit. X11 is already a standard, but Xlib is too lowlevel, but once you mention a standard toolkit, you get the groupthinkers screaming “don’t take away my choice”.
The author or that article, or rather, rant, is completeley wrong on all counts.
And I wouldn’t brag about being a Windows admin, specially for the pre NT versions.
Linux cannot be compared to former commercial Unixes because unlike the Unixes, Linux is licensed under GPL.
The beauty of Linux is exactly that of choice. Choice of distro, desktop environment, apps etc.
The microsofties would like precisely that, for Linux to become homogenous, monolithic monster like Windows.
That’s exactly what would destroy Linux, not its beautiful veriaty.
There is nothing technical preventing Adobe or Micromedia to write apps for Linux, other than market demand.
If Ooo.org and Mozilla can write their apps to run on any distro with little or no problems (even multi platform actually) then I can’t see why it would be technicaly difficult for Adobe or Macromedia.
Other than the fact that those companies like to charge exhorbitant fees and have product activation for their products, which wouldn’t fly with Linux users.
That’s exactly why we have Linux, Ooo.org, Mozilla, Gimp and so on so we don’t need Adobe and Macromedia.
It’s all about freedom.
The article is nothing more than poorly thought out rant.
And the author is probably a former Unix admin who lost his job and is now a stressed out MCSE Windows admin.
>>I can’t see why it would be technicaly difficult for Adobe or Macromedia
Software development costs money, Adobe or Macromedia or any other business is by definition “for profit” organization, their one and only purpose is to make money, so if they invest in software development they should be shure to get good ROI and sell it in feasible quantities.
Now, ask yourself, how many Linux enthusiasts will buy such expensive professional software?
First of all, we don’t even have usable 3D user interfaces today, talk little of during the hay days of motif.
Secondly, the free UNIX-like operating systems today, are nothing like their over-priced proprietary UNIX ancestors. Linux, the most popular free UNIX-like system, is the most customizable operating system under the SUN. The proprietary implementations of UNIX are barely versatile. That’s why they are dying and will continue to. In contrast, Linux runs on almost anything and can be used for almost any purpose, only limited by the imagination of its users.
Thirdly, proprietary UNIX died, or is dying, because they could only run on expen$ive proprietary hardware. Did anyone know of any Solaris, HP-UX, Irix or AIX for the the home user who wants to play Tetris and Solitaire, learn how to hack, or surf pr0n in the 90s? Windows survived because it ran on cheap hardware, period. But it has met its match, Linux. Sir Bill Gates is even loosing sleep over Linux and for valid reasons too.
Finally, yes, Linux is killing proprietary Unix, but proprietary Unix has outlived its usefulness anyway. I mean, do they do anything special anymore? To me they are more like a cog in the wheel. But you can’t deny the fact that Linux is coming after Microsoft and will reach it the moment Microsoft’s marketing camp begins to loose credibility and fumble. I’d argue it’s already happening. I smell rot in the air. How is it said? “You can fool some of the people some of the time, but you can’t fool all of them all the time.”
The public is still overwhelmed by the usefulness, openness and versatility of Linux. Give it time, it will blossom. IBM said it best, “The future is open.”
Obviously the first thing you will notice is that startup scripts in Linux (debian, slack, RH, SuSE) are all very very different.
That doesn’t really have an impact on app compatibility or consistency (which I guess is the crux of the criticism), unless your app is started at boot.
Secondly there’s no standards for creating menu items – SuSE puts KDE in /opt/kde/share/foo and others put in /usr/share/foo.
Actually, location of binary files and such is irrelevant to menu item creation. I’m not sure where you got the idea that they were somehow connected… The important thing is that KDEDIR is set, and that all executables are in PATH.
Finally devices – SUSE uses devfs – so your root becomes /dev/disk/controller0/disk0/part1 and Redhat uses /dev/hda or whatever and so now where’s device compatibility?
devfs and udev use symlinks to maintain compatibility.
Sure when you have source and then configure; make; make install will work but what about integration into the system? Like when you install a software using c;m;mi your system has no record of what version is installed.
Use checkinstall to make a deb/rpm package, which is then registered in the corresponding db. It’s as simple as “./configure && make && checkinstall make install”. If you don’t like the app you can then use the package manager to remove it.
That said, there are a whole slew of distro-neutral installers for Linux (the Loki installer, more recently used by Codeweavers for Crossover Office; the OpenOffice installer; the source-based XFCe installer, a derivative of Arkollon; autopackage; etc.).
Crossover Office is a good example: they only have a single installer, and yet I had no problem using it on my Mandrake install. It correctly added menu entries to my KDE desktop, and comes with an uninstall feature.
I don’t think “fragmentation” is a good reason for why Adobe and Macromedia aren’t porting their flagship applications to Linux. It think it has more to do with market share and (call me paranoid) agreements with Microsoft. That said, they have made some apps available for Linux, i.e. Acrobat Reader. Too bad it sucks bad (the new KPDF blows it out of the water).
The problem that Adobe will face is that if a proprietary app isn’t available for Linux, then an open-source alternative will come out that may eat at its market share on other platforms. While Gimp 2.0 doesn’t have every feature that Photoshop does, it is coming very, very close (to a point where a lot of people could in fact switch over).
I’m pretty sure the Gimp wouldn’t have known such a rapid (and successful) development if Photoshop had been available for Linux…
Why do people view the success of Linux and free software as being irrevocably tied to the fortunes of Microsoft?
Why is the assumption made that the ultimate goal of the first two are absolute world domination?
The world is a much bigger place. The author and article both suffer from major tunnelvision.
The guy has a point. Choice is not always a good thing(TM).
You know what I’d like to see? Apple to release the old NeXTSTEP windowing system as open source. That was a slick, streamlined, and beautiful interface.
Well, personally it’s not my cup of tea, but you can always check out AfterStep or OpenStep.
And I don’t see why so many are looking to mimic the functionality of the Windows environment. It’s Apple’s environment that everyone should be gunning for (that’s what Microsoft does, anyway).
This is matter to debate, but personally I’ve always felt that Gnome was going more towards Apple, while KDE was closer to Windows. Of course, both are themeable and configurable to look like either.
I installed Linux on one of my systems the other day, so I could use it as a teaching vehicle for my class on system log analysis. But first I had to Email a bunch of my friends and ask them, “what version of Linux should I use? Red Hat? Debian? Gentoo? Mandrake? Slackware? Do you think I could get away with OpenBSD or FreeBSD?” The responses I got indicated that none of my friends use the same thing but that I could be sure that if I used Flavor X some adherent of Flavor Y was going to bust my chops about it, and that someone was sure to show up with flavor Z and have trouble making things work.
Seems like all this guy’s friends have already switched from Windows to some flavour of Unix. There’s a saying about not seeing forest for the trees…
If you think he’s so wrong, totally off base, and dumb, then why do you think that open source operating systems/GUIs haven’t found the same success running on desktops that they’ve found in the server realm?
a) customer intertia
b) lack of marketing
c) Microsoft blackmailing OEMs (and other anti-competitive tricks)
Those 3 programs are enterprise-level programs, aimed precisely at the same market Sun targets. A thousand such programs could be on the market and they would have no impact on the status of Linux as a mainstream consumer OS that attracts new development and ports of existing products. In other words, every Oracle user in the world could switch to Linux and Adobe would still have no reason to port PhotoShop.
Aaron J. Seigo wrote: for ISVs looking at desktop software for Linux/UNIX there is really only one thing that matters: profit.
Before an ISV can make profit, the company needs to invest. Investments are risky: They might not result in a loss, not a profit.
The existance of several look’n’feels on the Linux desktop makes an investment riskier than your average single look’n’feel desktop (Windows, Apple). As long as users can see that an application uses a different look’n’feel than their preferred desktop, they might simply reject it — or even worse: developers will try harder to code a competitive application for ‘their’ desktop.
Another problem results from the missing direction of the Linux desktop towards a few target markets: I’m quite sure that Apple’s market share for publishers, designers, and musicians is _not_ 3% but much higher in these specific markets.
For these reasons, my personal guess is that Linux needs at least four times more market share than Apple got right now: We won’t see much ISV participation before Linux got a 12% – 15% percent market share on desktops.
In addition, I think the fragmentation claim based on the number of distributions is quite overstated. A lot of the various distributions are simply branches of a more popular distribution tweaked for a certain market (low-powered desktops, education applications, games, high-security, a specific language.) Once it’s installed, for all practical purposes you have Mandrake, Debian, SuSE or Red Hat.
In regards to the whole “choice” thing in regards to desktops. What about the recent article posted here on OSNews in regards to Gnome trying to slim down memory requirements? One of the reasons for multiple GUIs under development is the large diversity of software out there. One of the ways in which Microsoft both builds its market and reduces its support costs is by cutting off the lower end of the desktop spectrum every few years. Personally, I’m quite glad to see continued XFce, Blackbox and Fluxbox development. It certainly is better than taking a risk by running older versions of windows that are no longer supported.
I am amazed that these mainstream companies still havn’t released any major applications to Linux/BSD in a cut down form for free. If some of those cut down versions were available to the community I bet it would sting our Open Source alternatives in the a**e big time. Their big names and reputation should pull at least 50% of novice users from appllications (like The Gimp) over to their apps. How many people use Adobe Acrobat Reader for Linux? (Because KPdf/KGhost?Xpdf suck i guess). They are probably making enough money as it is with the Windows world.
Software development costs money, Adobe or Macromedia or any other business is by definition “for profit” organization, their one and only purpose is to make money, so if they invest in software development they should be shure to get good ROI and sell it in feasible quantities.
Now, ask yourself, how many Linux enthusiasts will buy such expensive professional software?
how many “windows enthusiasts” will buy it? i would venture to say that a significant portion of the photoshop install base is illegal on both windows and mac. if anything, that percentage would be lower on linux as there is a significant amount of users who use it BECAUSE of a strong sense of ethics.
that being said, there wont be that many as the linux install base is a fraction of that of windows. however, adobe hasnt dropped ps for mac, if there is a market for it on linux they would be smart to start work on it, there is nothing keeping linux users for paying for it that dont apply to windows or mac.
Well, they don’t drop mac version, because mac is popular among graphic designers.
> We won’t see much ISV participation before Linux got a 12% – 15% percent market share on desktops.
Marketshare doesn’t matter unless the users are prepared to spend money and have a reputation of spending money. Right now the OSS desktop userbase isn’t known to be really big spenders – case in point – Loki closed down because people weren’t buying games. Apple’s userbase known to be well heeled and will drop $$ on software. The OSS desktop userbase is only interested in free stuff – or you’d be seeing everybody buying Star Office and ignoring Open Office. Same goes for Opera Browsers and games.
This is not a generalization – we know because we have created a “for pay ” plugin for XMMS and WMP to enhance audio. We’ve sold more licenses for WMP because users are willing to pay for the Windows version while we’ve not sold more than a handful of XMMS licenses and on top of it someone already cracked the Linux version
(if interested see: http://www.oss3d.com)
Dev wrote: Marketshare doesn’t matter unless the users are prepared to spend money and have a reputation of spending money.
Yes, that’s true. Linux on the desktop attracts users that are a) free software enthusiasts, b) willing to spend time instead of money, and c) have lower or different software needs in general.
I guess, Star Office vs. Open Office may be due to c), and Opera vs. Firefox may be due to a). Your XMMS vs. WMP example isn’t really a prove unless you don’t take the base distribution of each app/operating system into account.
However, with 12 to 15% market share, there will be less people from classes a), b), or c). Just look at linuxquestions.org: The longest discussion is about “What software do you like to see ported”. Most answers cover Photoshop, Dreamweaver, and similar apps – and several people expressed their willingness to pay for these apps.
Well, personally it’s not my cup of tea, but you can always check out AfterStep or OpenStep.
Those are add-ons to current environements, or emulations of some of the features of NeXSTEP that run along side X11/MS Windows.
NeXTSTEP was it’s own windowing server, completely different from X11, in the same way that the Mac OS X GUI isn’t X11. Mac OS X’s interface is based on NeXTSTEP’s windowing environment.
X11 suffers from bloat. Well, that’s not fair. X11 doesn’t, but in order to get X11 to the level of a modern windowing system requires GNOME/KDE/TrueType/etc/etc. Layer upon layer upon layer.
NeXTSTEP did almost all of that (it didn’t have smoothed fonts) and did it on much more modest hardware. There’s a bounty right now I believe on making GNOME run on 128 MB of RAM right now; NeXTSTEP was fast, full-featured (again, missing only smooth fonts, not bad for 1993) and responsive on 16 MB of RAM.
Not to mention the KDE/GNOME diehards, who’ll boycott the other (and the small band of growing users sick of both system’s bloat and are going for lighter-weight environments such as Englightment and Xcfe).
Look at yourselves, this is exactly the ego problem raised in the article. Refusal to believe there’s any problem with yours, immediately brushing everything off as FUD, your way is the right way, blah blah blah. If you don’t see the problem, it’s because your ego has conveniently tuned it out.
EGO.
Mac OS X’s interface is based on NeXTSTEP’s windowing environment.
Err…not really. It is true that NeXT became OS X, but the Mac OS X interface is an evolution of the Classic Mac interface (with added unixy goodness). Let me suggest that it’s been a while since you took a look at the Next interface:
http://www120.pair.com/mccarthy/nextstep/intro.htmld/desktop1.gif
I see very little of Mac OS X in there.
X11 suffers from bloat. Well, that’s not fair. X11 doesn’t, but in order to get X11 to the level of a modern windowing system requires GNOME/KDE/TrueType/etc/etc. Layer upon layer upon layer.
It’s a modular approach, not a layered one, and that doesn’t make it any more “bloated” than other modern desktop environmnents. Of course, GNOME/KDE + X won’t run on 16M of memory, but OS X won’t either. They’re modern DEs, with a lot of features, including a lot that are mostly eye candy. However, that’s irrelevant, because 512 megs is as expensive today as 16 megs was in the heydays of NeXT.
The NeXT was a great machine. So was the Amiga. So was the BeBox. Let it go.
NeXTSTEP did almost all of that (it didn’t have smoothed fonts)
Come on, that’s not the only thing it lacked compared to modern Desktop Environments…
Not to mention the KDE/GNOME diehards, who’ll boycott the other (and the small band of growing users sick of both system’s bloat and are going for lighter-weight environments such as Englightment and Xcfe).
I like Xfce (not Xcfe). I often use it when upgrading the system (since I’m supersititious about upgrading KDE while running it…). It does load very fast on my slightly-obsolete Athlon 900, and feels a little bit snappier when moving Windows around.
Enlightenment, however, isn’t exactly a lightweight environment (and e17 is probalby going to require recent hardware if you want to use all the eye candy…).
Then again, on current hardware KDE and Gnome are in fact very snappy.
…the bounty is not to make Gnome run on 128M, but rather to make it more usable with that amount of memory. I’ve run Gnome on a Pentium 166 with 96M of Ram. Was it crappy? Yes, it was. But it ran, and it didn’t choke.
This guy’s article is so wrong it’s scary. He claims to be a computer expert and then writes things that are meaningless.
Modern computer desktop software isn’t written to platforms, it’s written to APIs. The fact that he dismisses Mac OS X as a Unix that run Adobe, Macromedia, and Corel software shows his stunning ignorance. He says they don’t count because they were written to the “Macintosh interface”. Huh? If he means the “Macintosh API” — in other words, Carbon — then he’s right. But Carbon, as an API, could run on any OS if it were implemented for that OS. Obviously, since Carbon works on both OS 9 and OS X — mind-boggingly different platforms.
The reason Adobe, Macromedia, and Corel don’t have major software releases for common Linux distributions simply comes down to one fact and one fact only: the amount of time and cost it would take to re-write their apps using “standard” Linux APIs (QT, GTK+, etc.) couldn’t be offset by sales. If Adobe thought it could sell X amount of copies of, say, Photoshop on Linux, and those sales would bring in considerably more money than it would cost to port Photoshop to Linux, then they’d do it. Period.
There’s no technical reason why commercial desktop software can’t run on Linux. The legacy Carbon and Win32 APIs certainly aren’t that fun in a modern .NET/Java/Python/etc. world, so QT or GTK (from what I’ve seen) would actually be a major improvement in terms of clean and productive API design.
Again, the person who wrote this article is without a clue. If Microsoft created its own Linux distribution and “ported” the Win32 API to Linux, I wonder what he’d say then?
Jared
I think he’s quite allowed to discard OSX – basically he wrote an article about Linux/BSD. Maybe he didn’t want to include OSX because it is a beast of a completely different colour and he didn’t want to scatter “except for OSX” everywhere.
Yes, Carbon *could* run on any OS if it were implemented for that OS. But AFAIK it hasn’t been, so don’t go talking about it like it has.
The question of MS “porting” their Win32 API was covered in another article recently, and it was basically a dream. Why on earth would they want to give people another excuse to migrate to Linux?
Err…not really. It is true that NeXT became OS X, but the Mac OS X interface is an evolution of the Classic Mac interface (with added unixy goodness). Let me suggest that it’s been a while since you took a look at the Next interface:
http://www120.pair.com/mccarthy/nextstep/intro.htmld/desktop1.gif
I see very little of Mac OS X in there.
Yes, really. You’re just looking at the superficial. The “plumbing” of Mac OS X is all based on NeXTSTEP and its windowing system, although there is quite a bit that’s been added (a lot of it eye-candy). I don’t think they even changed the name of the process that runs the windowing server.
There’s a direct evolutionary step between NeXTSTEP and Mac OS X called Rhapsody, released shortly after Apple bought NeXT and the reverse-takover began. It was basically the NeXT operating system (including the windowing system) ported to PowerPC, with some of the window UI converted to the “Copeland” look/feel. There was still a lot of NeXT “blacktile” menus in there as well, such as going into the nimaster controls.
Apple then took that and worked in new bells and whistles, and that’s what became the Aqua interface of Mac OS X.
If you’ve ever used NeXTSTEP, and you’ve used Mac OS X, you’ll immediately see how closely they’re related.
It’s a modular approach, not a layered one, and that doesn’t make it any more “bloated” than other modern desktop environmnents. Of course, GNOME/KDE + X won’t run on 16M of memory, but OS X won’t either. They’re modern DEs, with a lot of features, including a lot that are mostly eye candy. However, that’s irrelevant, because 512 megs is as expensive today as 16 megs was in the heydays of NeXT.
When comparing GNOME/KDE to Windows XP/2000/98, a system will run Windows faster than GNOME/KDE on 128 MB of RAM. With Mac OS X it’s a bit different, since it’s doing more (a lot of it is eye candy), so 128 MB of RAM for 10.3 is going to be a bit slugish (although not so much for 10.0).
Come on, that’s not the only thing it lacked compared to modern Desktop Environments…
I’m wondering if you’ve ever used NeXTSTEP. In 1993, before Windows 95 was ever released, you had a complete desktop environment, you had fully interopable objects in drag-and-drop (you could pull a sound file, drop it into a sound editing app and it appeared, or you could drop it into an email, and it would automatically encode it for you), you had an object-oriented development, it even did filled-in window moving. With NeXTSTEP you had a mail app that *still* rivals today’s Outlook or Evolution. All before 1995.
Unfortunately, the hardware was very expensive ($12,000 for a workstation).
I like Xfce (not Xcfe). I often use it when upgrading the system (since I’m supersititious about upgrading KDE while running it…). It does load very fast on my slightly-obsolete Athlon 900, and feels a little bit snappier when moving Windows around.
Which is the problem, you like Xfce, someone else is a GNOME supporter, and someone else won’t use anything but KDE. Such a variety of incompatible (in that you write an app for one it can’t easily be written for the other) environments, that you either have to write your own widet set (adding bloat to your app, not to mention time in re-inventing the wheel), but also making sure that they’re all installed with the appropriate versions on your system. And then you’ve got to contend with multiple distros that have different verions of all those environments.
This was never a problem in the server world. Libraries are much simpler, and a lot smaller, and anyone running a server was able enough to resolve it rather easily. The average spreadsheet user is lost in the realm of QT/GTK+ libraries and apt-get and rpm -i.
Yes, really. You’re just looking at the superficial.
I’m sorry, but that’s what the interface is. You’re talking about the “plumbing”, i.e. the software engines beneath. It may be semantics, but by “interface” we talk about the actual widgets, menus, controls, i.e. whatever the user uses to interact with the system.
So the interface is definitely Mac-like (it wouldn’t have made sense to get away from it, as Mac users might not have followed), even if the underlying layers are from NextStep.
When comparing GNOME/KDE to Windows XP/2000/98, a system will run Windows faster than GNOME/KDE on 128 MB of RAM.
I’m sorry, but you can’t put WinXP and Win98 in the same category. And with 128M the WinXP system will quickly become sluggish if you have a couple of apps open at the same time.
Incidentally, KDE 3 isn’t too bad with 128M – it’s actually faster than KDE 2 (and Gnome).
Which is the problem, you like Xfce, someone else is a GNOME supporter, and someone else won’t use anything but KDE.
That’s not a problem. To each his own. BTW, KDE is my main desktop. I use XFce when logging in as root because I usually do some quick tasks and it takes less time to load.
This is like saying: “the fact that some people like Windows and some people like Macs is a problem” or “the fact that some people like the PS2 while others have Xboxes and others will only swear by the Gamecube.” It doesn’t matter, because in the case of Linux DEs they are but a shell and all Linux programs can still be used.
Such a variety of incompatible (in that you write an app for one it can’t easily be written for the other) environments, that you either have to write your own widet set (adding bloat to your app, not to mention time in re-inventing the wheel),
What? Why would you re-write an app? GTK apps work fine in KDE, Qt apps work fine in Gnome, and both GTK and Qt apps work fine in XFce. I use Gimp (a GTK app) all the time in KDE. My GTK and KDE themes were chosen to look good together, and everyone’s happy.
but also making sure that they’re all installed with the appropriate versions on your system. And then you’ve got to contend with multiple distros that have different verions of all those environments.
No, you only have to contend with the distro you’re using. Install from the distro’s repository. I suggest Mandrake, as they have a huge software repository.
I’m sorry, but all nostalgy for dead OSes aside, you’re basically lamenting about false problems. There’s no problem about using Qt and GTK apps in either KDE or Gnome. It is completely fallacious to claim that they are “incompatible”. Sure, you’ll get two different “Save” dialogs. Big deal!
It often seems to me that the proponents of “consistency” (an utterly minor issue which has nonetheless become the new favorite of Linux critics) seem to think that users are so dumb that they will be completely confused by having two different file dialogs and widgets that are not 100% identical – never mind that Windows apps use quite a few different widget sets themselves (how many Windows toolkits are there, in all? 4? 5?).
The fact is that if an interface is intuitive, people will know how to use it. Look at video games: they all have distinct UIs, themed to go with the game, however most of them will follow the same design guidelines (in addition to requirements from the manufacturers, such as Sony’s TRCs). The result is that players rarely think twice about navigating them.
Give computer users some credits. Minor inconsistencies doesn’t really matter as long as interfaces are intuitive.
Which is the problem, you like Xfce, someone else is a GNOME supporter, and someone else won’t use anything but KDE.
As an Xfce user, it’s not really a problem. Xfce runs the same toolkit as Gnome. I’m using Opera (built using static Qt), and just for fun I popped up epiphany (from Gnome) and if I wanted to I could use Konqueor but I’m rather nuts in that I don’t mind Xffm. I can probably run most of these applications under Fluxbox as well.
Secondly, I think the state of the Gnome vs. KDE debate is that any of the big players could settle the debate de facto by picking one or the other. Hasn’t anyone considered that if Macromedia or Adobe jumped onto a Gnome or KDE bandwagon, that everything else would follow?
Third, the argument that we need less diversity in the Free Unix world is seriously baffling to me. Almost all of the distributions have standardized on Gnome or KDE, leaving the rest for people who need or want a different user environment.
I mean this literally.. my eyes hurt..
I’m sorry, but that’s what the interface is. You’re talking about the “plumbing”, i.e. the software engines beneath. It may be semantics, but by “interface” we talk about the actual widgets, menus, controls, i.e. whatever the user uses to interact with the system.
So the interface is definitely Mac-like (it wouldn’t have made sense to get away from it, as Mac users might not have followed), even if the underlying layers are from NextStep.
You’re talking about interface, and I’m talking about technology. I don’t really care about the particulars of the interface, so long as it’s easy to use, intuitive, and lighter than what we’ve got now. Heck, make it look exactly like GNOME/KDE for that matter. Even have Sun, Novell, RedHat, and groups like Gentoo, Fedora, Slackware, and others create their own customized version, just like they do with GNOME/KDE.
The NeXTSTEP windowing environment is a lightweight, incredibly powerful, increadily featured system running on top of BSD. So powerful, that it’s barely behind even now, when it hasn’t seen any development since 1997 when NeXT got bought by Apple (and even then developement had stalled).
I’m talking about a windowing system alternative to X11. There have been, and there are currently, better environments than what X11/GNOME/KDE/Xfce. Heck, there are open source projects reviving BeOS, another windowing system that was ahead of its time.
Third, the argument that we need less diversity in the Free Unix world is seriously baffling to me. Almost all of the distributions have standardized on Gnome or KDE, leaving the rest for people who need or want a different user environment.
Why is it that those who champion choice freak out at the notion that perhaps there’s a need to move to something other than X11?
It often seems to me that the proponents of “consistency” (an utterly minor issue which has nonetheless become the new favorite of Linux critics) seem to think that users are so dumb that they will be completely confused by having two different file dialogs and widgets that are not 100% identical – never mind that Windows apps use quite a few different widget sets themselves (how many Windows toolkits are there, in all? 4? 5?)
That’s a completely different consistency than what I’m talking about. I’m talking about two or more different ways to write a program that shows a button, each way requiring a different set of libraries.
What? Why would you re-write an app? GTK apps work fine in KDE, Qt apps work fine in Gnome, and both GTK and Qt apps work fine in XFce. I use Gimp (a GTK app) all the time in KDE. My GTK and KDE themes were chosen to look good together, and everyone’s happy.
This isn’t about two different save dialogs.
A program with “gtk_widget_show (image);” isn’t going to work on a system with only KDE and libs installed, and a program with “QPushButton hello( “Hello world!” );” isn’t going to work on a system with only GNOME/GTK+ installed. So what would Photoshop for Linux be developed in? GNOME? KDE? The ugly X11 libraries? The slightly less-ugly motif? Or write their own?
There’s a lack of simplicity in all of these solutions. A common way of doing software installs (but different install apps available, you know, for choice, and doing it graphically like you can for every other environment, rather than every distro having their own method of package management that they’d rather die than use another.
One of the things I like best about NeXSTEP/Mac OS X is simplicity. From software installs, to security updates. There’s so much duplicated effort, most Linux desktop environments now do some things very well, and other things terribly. Great KDE but poor GNOME; great GNOME but poor KDE. Great install methods but poor layout/user experience, etc.
Secondly, I think the state of the Gnome vs. KDE debate is that any of the big players could settle the debate de facto by picking one or the other. Hasn’t anyone considered that if Macromedia or Adobe jumped onto a Gnome or KDE bandwagon, that everything else would follow?
There’s no killer app that I could think of, other than Firefox or OpenOffice perhaps, that has a product going in that could do it, and I don’t think GNOME or KDE would like having Macromedia or Adobe control their fate.
And besides, it’s such a risky move. Perhaps Adobe were to port Photoshop with built with GTK+. But then everyone moves to KDE, and now they’re in a bit of a pickle. Like others, they’d likely just end up writing their own widget set, and reinvent the wheel.
It’s a real weakness of the open source community, not wanting to acknowledge weaknesses in areas where we’re weak. Rather, it’s easier to asail the messanger, and lament all the other reasons that we’ve convinced ourselves on why something isn’t flourishing like we think it should. After all, clearly any criticism is FUD.
There’s a reason why open source operating environments haven’t flourished in the desktop realm like it has in the server realm. And maybe neither you or I have hit on the exact reason, but it’s clear that it doesn’t present the same compelling benefit to users as it does in the server realm.
Jared, you missed the point. It doesn’t matter whether, technically, the commercial software could be ported to APIs which could be potentially made availabe on alternate platforms. No commercial OS vendor (Microsoft, Apple, etc) in his right mind is going to port his bread and butter APIs (ie. Win32, Carbon) to another platform. And most “write once, run anywhere” APIs (Java, .NET) are lowest common denominator — which means you wouldn’t want to use them for crafting innovative GUIs (doubt it? name any *widespread* commercial app that uses Java or .NET — and, no, I’m not talking about freeware Limewire or anything like that). So, what emerges are partial solutions such as WINE. Or VMs (Vmware, Virtual PC, etc). None of which offer any particular advantage in contrast to the inconvenience of using such products.
Here’s a question for you all. How many Linux users still use Windows? I think it would be safe to say that a majority still do. Even if it was only for games, it still means that Windows fulfills a need that the alternative OS can’t.
Let me put it this way, if I was suddenly required to use Windows only on my desktop PC at home, it would realy suck but it wouldn’t really matter either, life goes on as usual. If I suddenly had to use Linux only, however, it would be very, very difficult.
There are some things that I just won’t be able to do (although for there are alternatives for the majority of what I do). I would also have to spend considerable time and effort retraining myself (although that’s a given and I think I’ll enjoy that). But the real pain lies in the world around me, which requires the use of Windows based file formats. Now that’s one royal pain. Simply put, they won’t bend over backwards for me. Sucks, but that’s the way it is.
The truth is, Linux on my desktop PC right now is a “nice-to-have” but there is no real need for it. Windows on the other hand. . .
Although it’s right that the multiplicity of distributions give the impression of the fragmentation that occured in the Unix times, It remains true that any linux apps (take Opera or Firefox) will run on ALL current linux distributions.
It is specially true of commercial software that will typically come with all required libraries in its install path (/opt for example) if not statically linked to them…
And if that wasn’t the case, then there is still Java that guarantees consistent software experience on all supported platforms. Take Limewire, Poseidon, Eclipse.
The lack of quality applications is due to the lack of perceived market (Only corporation buy things like the Acrobat suite) for these apps.
And lastly, the apparent proliferation of distros is due to the fact that it doesn’t cost anything to appear on distrowatch and other tech directories. But I am sure that less than 5 distros make up 60% of linux users and if you take into account that Debian offshoots are for some reason pretty faithful to the Debian structures and ways, then it’s a lot less fragmented suddenly.
But it would be good if every year, the main distros and maintainers agreed on the versions of the kernel and essential libraries that should be considered live and agreed to call it something like “Linux 2005”, that would clear up the mist and would stop this kind of misguided articles to get us all excited.
> Secondly there’s no standards for creating menu items
Actually there is
Thought these links would be relevant.
KDE Usability Project
http://usability.kde.org/
Gnome Usability Project
http://developer.gnome.org/projects/gup/
Real Programmers DO Care.
Your article goes to show that you’re absolutely CLUELESS.
we have only really gotten there with the larger Linux distros in that last year or so.
I don’t think we’ve got there at all to be honest. What ISVs want is the ability to write a piece of software in a standard way and create a graphical installation package. They don’t want to target a particular version of glibc or any other particular component they may find in someone’s system. With servers this doesn’t matter so much as slightly different packages can be written for different distributions without too much trouble, but at the general purpose desktop end this just isn’t acceptable.
This become rather more pointed when it comes to the desktop. There’s a long way to go.
Here’s a question for you all. How many Linux users still use Windows? I think it would be safe to say that a majority still do. Even if it was only for games, it still means that Windows fulfills a need that the alternative OS can’t.
Let me put it this way, if I was suddenly required to use Windows only on my desktop PC at home, it would realy suck but it wouldn’t really matter either, life goes on as usual. If I suddenly had to use Linux only, however, it would be very, very difficult.
And people like you are useless to the free software community. You are useless, your advice is useless. On the other hand, most in the community perhaps use Windows sometimes for some purpose, but they do all they can so that it is no longer the case later.
While you sit on your a** telling everyone to emprison themselves, others do sth. I respect them, they have value, you have none, and I do not respect your opinion.
Because you do not value your freedom does not mean others should do the same.
I don’t think we’ve got there at all to be honest. What ISVs want is the ability to write a piece of software in a standard way and create a graphical installation package. They don’t want to target a particular version of glibc or any other particular component they may find in someone’s system. With servers this doesn’t matter so much as slightly different packages can be written for different distributions without too much trouble, but at the general purpose desktop end this just isn’t acceptable.
This become rather more pointed when it comes to the desktop. There’s a long way to go.
I know pretty well most of the people trying desperately to bash Linux here or elsewhere do not know what they are talking about, because what they are saying is almost always contradicted by facts that someone using Linux as its primary system daily (like me) should instantly see.
I use a custom Linux (only I have one like this, I made it), and guess what ? Adobe Reader installs just fine (but I erased it, free software tools are way better), Flash too, Real Player too (erased also).
Guess what ? I installed the privateer remake (a game, as you would not believe it) in 5 minutes on my custom Linux system, packaged with the several years old Loki installer, without any problem (it is compatible with every distro), and the game appears in my Gnome and KDE menus !!!!!!!!!!! And this was done by what you would call amateurs (not even ISV). So when I read clueless people like you, I have the impression that ISV have less clue than amateurs.
But actually, I know that that is your arguments that are clueless.
Obvioously, , when bashing sth, you should know what you are talking about. The author of the rant is clueless, so he is harmless, just making a fool of himself, using the syndrome of false authority (being a security expert does not qualify you to talk about desktops).
But there are other clueless ones in the posts, such as this one :
Obviously the first thing you will notice is that startup scripts in Linux (debian, slack, RH, SuSE) are all very very different. Atleast with BSD/Solaris/AIX it’s well defined.
Actually, they are all as well defined in different brands of Linux than on different brands of Unix. Actually, they are basically the same. Actually, it depends only on your version of init, so on Linux, I think there are 3 versions (the third one not present in Unix environments is the one I use, simpleinit-msb). Anyway this is irrelevant, I can use all 3 without any problem, as they all use the same principles.
Secondly there’s no standards for creating menu items – SuSE puts KDE in /opt/kde/share/foo and others put in /usr/share/foo. Still there are issues with Libraries and locations – Linux/Unix people flamed Microsoft for DLL hell, well it’s no better on UNIX either.
A pretty clueless one again. There IS a standard for creating menu items, and at least Gnome and KDE adhere to it. It is so amazingly flexible, that you can not even understand that the place you put these files in is irrelevant, it will work (the privateer remake game I installed today on my custom system is a good proof of it, the game appeared in the menus, without me telling it where they are, and they are not in default places).
And there are NO issues with libraries and locations, but you could not know this, you think it is flawed like on Windows (at least you admitted that).
Finally devices – SUSE uses devfs – so your root becomes /dev/disk/controller0/disk0/part1 and Redhat uses /dev/hda or whatever and so now where’s device compatibility?. Finally package management – Redhat and SuSE use RPMS, but when you install XMMS – it’s stuck in a different menu than Redhat.
I used devfs for years, and devfs HAS a compatibility layer, but it is too clever for people like you. Actually, if the compatibility layer is activated, /dev/hda WILL appear ONCE you try to access it. But you could not know that, you do not know what you are talking about. There is NO compatibility problem at this level on Linux. Heck, I switched to udev recently because devfs is deprecated, and even there, there is a compatibility layer for devfs space names.
Sure when you have source and then configure; make; make install will work but what about integration into the system? Like when you install a software using c;m;mi your system has no record of what version is installed.
You are not lucky really, all your arguments are BS. I use exactly the environment you describe. Of course I do not do all this by hand, I use automated install that do the c;m;mi (it is called nALFS). Now I’m bleeding edge and it is a breeze to install anything (I took several hours creating the XML files though). The auto install used installwatch and a custom perl script (nuke) to manage the packages. Now I use paco (it took a 5 minutes scripts to migrate my 1000+ XML files to use paco), and it is even better, with a nice GTK+ interface to manage packages and files if I want (I don’t). My system has full record of what is installed, but you would not know, you are clueless.
Now everybody on this board will be quick to retort “hey it’s open source and choice is good”. I don’t know how to answer that!.
No, I retort “Hey it is free software and way better than anything I tried till now, and all your problems are non existant in Linux” (and I tried a lot, from AmigaOS, which was my favorite, to BeOS, which I liked a lot, to MacOS X). But you could not come up with anything but your cliché, as you are clueless about the Linux desktops.
linux lusers spend all their time configuring the system, after making the video card work, the audio fails .., they have forgotten women exists, get a life! use windows!
To end my contribution against clueless posts, the situation you describe is EXACTLY why the last lot of migration to Linux happened.
When nearly every evening you want to watch some movie with your wife (like mine) because she asks for it (like mine), and it does not work 80% of the time (because Windows screwed himself on some update, or just plainly erases codecs, or they are incompatible, …), and I spend 2 hours to fix it, one more evening lost and wife unhappy.
Since I switched the Windows client to geexbox, no more problem, no more windows puker, it works 99 % of the time (exception being some old real files).
“Trolltech has provided us with an intuitive, powerful tool. Qt simplified our task of developing Photoshop Album by providing high-level tools that we could customize to meet our needs,” said Mike DePaoli, Photoshop Album Engineering Manager. “The product is excellent, the support was outstanding and we are extremely pleased with our decision to go with Qt.”
http://www.embeddedstar.com/press/content/2003/2/embedded7410.html
http://www.trolltech.com/newsroom/announcements/00000120.html
Maybe to release Adobe Photoshop would mean building different binaries/packages for Red Hat, SuSE, Mandrake or whatever. I would have thought this, along with lack of demand from their customers, are greater problems than the lack of a ‘standard toolkit’.
Maybe there will never really be a strong demand for closed source software on Linux, because of its inherent problems such as a lack of flexibility/customisation without the source, along with reduced longevity compared with Free Software equivalents.
Mac OS X has both Cocoa and Carbon apis, which have gradually been integrated, and you can’t tell if an app is written using one or the other. The Gnome and KDE toolkits are heading in the same direction via cooperation such as freedesktop.org, and I’m sure in a year or two it will be hard to spot if a Linux app was written with one or the other. That doesn’t mean there might be reasons for prefering one over the other, but having a choice of toolkits has no bearing on the viability of releasing a commercial app.
What this guy is circling around but does not get is: Standards are important. What else should the wish to have a consistent interface/programming environment/devfs across distributions be. Now, the only company torpedoing standards wherever possible is Microsoft, simply because they do not want to loose their stranglehold on their customers.
Java, XML, OpenDocumentFormat, … all torpedoed with embrace and extend, patents and ignorance.
So in effect in every “healthy IT market” (according to that guy) where one market leader gains 70%+ marketshare that market leader can squeeze ALL other competitors out of the market by simply BEING the institution for de-facto standards (see MS Office).
This situation can be broken up by highly cost-effective, innovative competitors who do not let themselves to be bought – like the FOSS culture. Customers do not care about consistency, they care about getting “good enough” for as little money as possible (what MS delivered in the 80ies and 90ies). FOSS is doing VERY well in that respect, as soon as you can see 400$ laptops sold with either Linux or Windows pre-installed where the price reflects the purchase cost of the operating system you will see LOTS of people buy the cheaper Linux variant. You will also see Microsoft’s prices go down.
From the article:
“I installed Linux on one of my systems the other day, so I could use it as a teaching vehicle for my class on system log analysis.”
Rayiner Hashem wrote:
“This has to be a joke. They have classes for that???”
Why a joke with a three ???
Obviously, analysis of the behaviour and effects of a strange process based on log records can get very complicated. If you can have firewall rule writing classes (I’ve been to one) why not have system log analysis clases?
I mostly agree with the article. The petty politics of so-alled FOSS communities weakens them a lot.
The communities lack leadership (I’m not referring to the Linux or BSD kernels, but to other software) because there are a lot of ego-driven people in those communities, and because there are a lot of ego-driven people in the world as a whole, too. But at this point you may ask “Isn’t Microsoft ego-driven, too?” It definitely is, but the paycheck that it generates for its individual employees kind of forces those individual employees to unite as one coalesced ego, because they do not want to lose their paycheck. Whereas in the case of the FOSS communities, there’s (generally) no paycheck coming from a single source, so there is no stimulus for them to act as a single ego, and thus we get the fragmented situation that we have now.
If you disagree with my above 2 paragraphs, I’d love to know why.
I’m running my own software company and I can tell you about the following in the article:
“So the battle in the free UNIX space is entirely over command line options, system administration paradigms, installation packaging, and 3D GUI features. I’ve got news for you: Real Programmers Don’t Care about that garbage.”
I agree that “Real Programmers DON’T Care about that garbage!” because those things are not crucial to the functionality of a product, unless, of course, it’s an administration product, an installation packaging add-on module or a GUI addon. But those products are only a very tiny fraction of all products. For all other types of software they are not essential because they do not in any way contribute to the features of the software or its usefulness. So I agree, but I would have to add that they DO care if they develop one of the very few products that have a significant relation to the administration paradigm, for example (and the others).
What happened to UnitedLinux?
“linux lusers spend all their time configuring the system, after making the video card work, the audio fails .., they have forgotten women exists, get a life! use windows!”
You know, there are other things apart from women
— such as mental and emotional health that you have to build solely by yourself. I am a Linux luser, I am also a Windows user and a FreeBSD user, and an OS/2 user, and one notable thing I have found out is that Linux and BSD can waste a great deal of time to keep the systems up-to-date as opposed to the little time it requires in Windows and the largely irrelevant, but good old OS/2. For example, I’ve been having multiple problems over the period of several months with Fedora Core 3, because occassionally there are clashes between packages-to-be-updated with installed packages, or missing dependencies. Resolving those required my manual intervention, and sometimes I had to invest time on the order of a couple of hours to track down the clash and wait for the updates to download, so that I can be sure things are really resolved. Whereas modularity through various packages is definitely good in the FOSS world, it requires more care than the smaller extent of modularity in the Windows world, all the more that in the Windows world most of the modules are controlled by one vendor. With greater modularity and more sources to supply the modules, Linux and BSD distributions have to be extra extra careful if they want to ensure a smooth up-to-date and safe (as far as it is possible to be safe through technology alone) computing experience.
It often seems to me that the proponents of “consistency” (an utterly minor issue which has nonetheless become the new favorite of Linux critics) seem to think that users are so dumb that they will be completely confused by having two different file dialogs and widgets that are not 100% identical – never mind that Windows apps use quite a few different widget sets themselves (how many Windows toolkits are there, in all? 4? 5?).
Making this kind of ridiculous straw-man argument doesn’t do you any favours. You may consider Linux inconsistency to be a minor issue, but you’re not in a position to speak for all users.
The inconsistency between some Linux apps goes far beyond the kind of minor cosmetic inconsistency found in Windows. It may not make Linux totally unusable, but it does make some tasks significantly more difficult and time consuming.
I wouldn’t switch to an OS that wasted my time by crashing regularly. I wouldn’t switch to an OS that wasted my time by running significantly slower than the alternatives. So I’m not going to switch to one that wastes my time by having a UI that I find vastly inferior.
For me Linux would need to have major advantages over Windows to make up for it’s UI problems, but the only advantage I can see is the money I’d save by not having to buy Windows. That’s totally inconsequential compared with the time I’d waste fighting with Linux problems like it’s crippled copy and paste between apps.
The point I started the previous post for but forgot to make, is that Linux / BSD are NOT more difficult to use than Windows if you are able to counter the temptation to update your system all the time (and have to deal with problems I mentioned in my previous post). That, and given that your hardware is supported, makes Linux and BSD totally usable on the desktop today.
To those people earlier who said I missed the point, I think they missed the point themselves. This strange and meandering article seems to make the argument that: because it’s easy to develop for Windows and hard to develop for all of the *NIXs out there, Windows will win. But, again, you don’t develop for an OS, you develop your app in its own little cocoon which talks to the rest of the world via APIs. If you have the same API on multiple OSes, porting your app becomes relatively trivial.
Now I never said you’d see Carbon or Win32 appearing anywhere else. My argument was that big Windows-world developers don’t port their apps to one of the major Linux OS APIs because the sales can’t justify the costs. There’s NOTHING stopping Adobe, Macromedia, or Corel from creating software that runs on the major distributions or Linux Standards Base-compliant distributions. If Windows had 1% desktop marketshare and the combined marketshare of the Linux distributions were 95% or so, which OS do you think Adobe, Macromedia, and Corel apps would be running on? Duh.
This all about marketshare and APIs. Nothing to do with the number of Linux distributions or the Linux vs. BSD controversy whatsoever, as the article’s author suggestions.
Jared
i would go so far to say that there are a handful of dms that are significantly MORE usable then windows. using linux is quite easy, its administering it that requires the obscene amount of knowledge and work.
You’re talking about interface, and I’m talking about technology.
Yes, I understand that, but you originally said: “Mac OS X’s interface is based on NeXTSTEP’s windowing environment.” There was some ambiguity to this statement, which probably explains why I misunderstood what you were saying.
I’m talking about a windowing system alternative to X11. There have been, and there are currently, better environments than what X11/GNOME/KDE/Xfce.
Better in what sense? What matters is the end result, really, and on current hardware Xorg+KDE or Gnome runs just find. It seems to me what you have is a solution in search of a problem.
Why is it that those who champion choice freak out at the notion that perhaps there’s a need to move to something other than X11?
Because there’s no real need. The current framework works, and works very well – especially since the switch away from stagnant XFree86 to the dynamic team behind Xorg.
If there really was a problem with X, then projects such as DirectFB and Berlin would be developing much faster than they are. As it happens, they are progressing very slowly, because most people involve don’t feel the need to reinvent the wheel.
I’m talking about two or more different ways to write a program that shows a button, each way requiring a different set of libraries.
And what’s wrong with having two sets of libraries? Hard drives are under 1$ a GB these days. Why does a couple of MB of extra libraries matter that much?
So what would Photoshop for Linux be developed in? GNOME? KDE?
Seeing as they used Qt for PhotoAlbum, I’m pretty sure they’d go with Qt for Photoshop as well (if they were to port it). That’s not a problem, the installer simply verifies if the libraries are there, and if not it installs them (possibly in its own specific folder, so that they don’t compete with other installed versions of Qt). After all, this is what most programs do in Mac OS X and Windows, right?
And besides, it’s such a risky move. Perhaps Adobe were to port Photoshop with built with GTK+. But then everyone moves to KDE, and now they’re in a bit of a pickle.
Why? KDE users still use Gimp, even though it’s GTK2. You don’t hear them complaining about it (I certainly don’t).
It’s not a big deal, really. And it’s certainly not the reason why PS hasn’t been ported to Linux, or why Linux still has less than 3% of Desktop market share, IMO.
It’s a real weakness of the open source community, not wanting to acknowledge weaknesses in areas where we’re weak.
Well, what you perceive as weakness I percieve as strenght. Competition is good, as it pushes each competitor to come out with a better product. Would the pace of KDE/Gnome development have been so rapid if there hadn’t been a competing DE to beat? I don’t think so. Eventually, the better technology might “defeat” the other, but that’s not important. The important thing is that competition breeds innovation, and monopoly breeds stagnation. I’d rather have the former than the latter.
There’s a reason why open source operating environments haven’t flourished in the desktop realm like it has in the server realm.
There are multiple reasons, as I’ve indicated in a post above. This IMO isn’t one of them.
The inconsistency between some Linux apps goes far beyond the kind of minor cosmetic inconsistency found in Windows. It may not make Linux totally unusable, but it does make some tasks significantly more difficult and time consuming.
The single thing you were able to bring up, as far as UI inconsistencies go, is the fact that copy/pasting images doesn’t work for all apps (I did ask you what those apps were, but you never came back to me). I personally have not experienced those problems as there are workarounds that are as efficient (and even more intuitive, as in the case of drag and drop).
So I’m not going to switch to one that wastes my time by having a UI that I find vastly inferior.
Yes, we all know you’re not going to switch. But, to paraphrase you, you’re not in a position to speak for all users.
I have argued that the inconsistencies are minor and not worth the fuss. Obviously, they disturb you to a great degree. I have irritants in Windows as well, yet I don’t go around claiming that it greatly reduces my productivity, because I share a quality what most computers users have: adaptability.
Why is it that those who champion choice freak out at the notion that perhaps there’s a need to move to something other than X11?
I’d be happy to support something other than X11 as long as the must-have applications I need get supported on it as well.
There’s no killer app that I could think of, other than Firefox or OpenOffice perhaps, that has a product going in that could do it, and I don’t think GNOME or KDE would like having Macromedia or Adobe control their fate.
And besides, it’s such a risky move. Perhaps Adobe were to port Photoshop with built with GTK+. But then everyone moves to KDE, and now they’re in a bit of a pickle. Like others, they’d likely just end up writing their own widget set, and reinvent the wheel.
Now I would argue that if Adobe were to pick gnome, that it would pretty much establish GTK+ as the desktop standard. Would the toolkit makers like it? I don’t think they would have much of a choice. For a historical comparison, the electric current wars were pretty much settled by the adoption of AC for the Chicago World’s Fair. After that companies based on Edison’s DC technology were forced to adapt to the new standard, or drop out of the energy business.
Sun needs no help from the opensource people to go out of business, they will manage that perfectly well by themselves.
Look at Java one day they think it should be standardized, the next day they don’t, one day they tink it should be opensource the next day they don’t. One day they want to opensorce Solaris the next day they don’t and now they do again. One day they believe in Solaris x86 the next day they don’t and now they do again.
By not making their minds up they loose customers in all camps. Who in their right mind would do business with Sun. Who knows where they stand tomorrow.
Most Linux people that I talk to want free software, me included. I don’t know if they would buy Adobe for Linux even if it was out. I would not be surprised if these companies have done market research that found that they would not see a return of investment, building Linux versions of their software.
They do, this was the reason there is no FrameMaker for Linux. They even made beta tests for it, but most Linux customers was not prepared to pay even half the price Microsoft and Mac people did. This was a couple of years ago so it might have changed by now as Linux becomes more and more a matter for business users.
I’m agree with what he say’s and I think the only linux suvival is to be united!
but about photoshop and falsh and etc I’m not agree with him because first why they should do it! they have a good market now! and second I think Microsoft can’t live without this softwares so their must be a lot of secrets that we don’t know!
I did ask you what those apps were, but you never came back to me
Some examples: OpenOffice, GIMP and Scribus
I personally have not experienced those problems as there are workarounds that are as efficient (and even more intuitive, as in the case of drag and drop).
Having to save data to disk then import it into another app is not as efficient as copy and paste. If it was then there’d be no point in having copy and paste at all.
Drag and drop often isn’t as quick and efficient as copy and paste. Both documents have to be visible on screen before you can drag and drop anything between them. That can require window resizing and shuffling that’s more time consuming than switching between windows with alt+tab or the taskbar. Drag and drop is often totally impractical when working on a small screen such as a laptop’s LCD.
Anyway, IME drag and drop doesn’t work much more consistently in Linux than copy and paste does. For example, try dragging and dropping a selection of an image from GIMP to OpenOffice or Scribus.
Yes, we all know you’re not going to switch. But, to paraphrase you, you’re not in a position to speak for all users.
I’m not speaking for all users, I’ve made it very clear that I’m purely speaking for myself. It’s you who seems to be arrogantly claiming that your opinion determines what is or isn’t a problem. Just because something doesn’t bother you, it doesn’t mean that it isn’t a significant problem for other people.
I have argued that the inconsistencies are minor and not worth the fuss.
For you that’s obviously true, but that isn’t necessarily true for other people. Are you really so conceited that you can’t understand that not everyone has the same needs as you?
Obviously, they disturb you to a great degree. I have irritants in Windows as well, yet I don’t go around claiming that it greatly reduces my productivity, because I share a quality what most computers users have: adaptability.
More smugness and arrogance.
Does your ‘adaptability’ mean that you could happily adapt to using a single-tasking computer that crashes every hour and runs at 10% of the speed of your current system? Or would you find a downgrade like that frustrating and damaging to your productivity? I know people who would be perfectly happy with those limitations, does that mean that they’re more adaptable than you or I?
I could get work done on a Linux system, it would just take me more time than it would on a Windows system. I could get work done using Windows 3.1 on a 486 if I had to, but that would probably slow me down even more than having to use Linux. Instead I use the best tool for the job, one that doesn’t suffer from problems that make me less productive. Why is that so hard for you to accept?
The war between KDE and Gnome is kinda worrysome though.
I would say that most this is not much of a problem anymore. At least KDE developers often try to be Gnome compatible (I’m not implying that Gnome people doesn’t try to do the same with regards to KDE, I simply don’t know as my background is mostly KDE). Free.desktop.org makes a wonderful job of keeping both sides in line. But I doubt this is a very hard task. Today most people know the importance of interoperability.
The problems will be even less in the future as Gnome most likely will win the battle for the Linux desktop of the average Joe. It will do so mostly for usability reasons. If you read KDE usability mailing lists, you get the impression that users spend 99% of their time configuring their systems and 1% for work. Sure when the KDE control panel is rearanged for the I don’t know what time, it will be much better, but by that time Gnome users will enjoy a system that not only is resonably simple to configure, but also simple to work with. Guess what most people will choose.
I know pretty well most of the people trying desperately to bash Linux here or elsewhere do not know what they are talking about, because what they are saying is almost always contradicted by facts that someone using Linux as its primary system daily (like me) should instantly see.
You have no facts to bring.
Guess what ? I installed the privateer remake (a game, as you would not believe it) in 5 minutes on my custom Linux system, packaged with the several years old Loki installer…
ISVs are simply not going to write their own installers like Loki did, nor are they going to engineer their own installation package because the native packaging system of the distribution wants to pull in umpteen packages when you install anything. They want to target one snapshot system that is stable and backwards compatible whatever happens to it, and they do not want to field calls from customers who are asking why their system falls apart and want to update practically their whole system. Your argument (whatever it is) falls apart right there. You simply don’t understand what is required.
Wiping out half a dozen toolkits and only having one isn’t going to solve the problem either – sorry fanboys. That tookit, whatever it is, will have a dozen versions in any given year and a major version where backwards compatibility is broken. That toolkit will always have dependencies on other libraries and those libraries will continue to have multiple, iterative versions. That’s the crux of the problem – being able to update vital components in a user’s system, adding new features while maintaining backwards compatibility and not giving ISVs a constantly moving target. That happens regardless of how many toolkits there are.
Windows has many toolkits (many used and created by Microsoft), and there are many third-party companies who have their own tookits on Windows. Some of them look vastly different and just don’t fit in. Yet, they do seem to fit in because they all have one thing in common – they use, or link in at some point, core Windows technology and the IPC system that binds them together. You can program differently, but have everything link in.
That’s not to say something like Freedesktop can solve all of these problems. Many vendors will want to create control panel applets, system tray applets and they will need a consistent and reliable API to do so. Even if desktops adhere to an API there are always differences – there always are. And because you’ve got an open source system it’s certainly possible that many unscrupulous vendors will stupidly replace your packages with their own patched versions breaking a heck of a lot of things, even worse than they do at the moment. The list is endless.
I’m an open source advocate as much as the next person, but the above are nevertheless problems – some of them we still have to face if we’re even lucky enough to get there!
At least KDE developers often try to be Gnome compatible (I’m not implying that Gnome people doesn’t try to do the same with regards to KDE, I simply don’t know as my background is mostly KDE).
You won’t hear any (or many at all) Gnome developers talking about interoperability of any kind, especially with KDE. It seems that KDE is simply expected to adopt what Gnome is using, and if they don’t they’ll probably be accused of splintering.
The problems will be even less in the future as Gnome most likely will win the battle for the Linux desktop of the average Joe.
This has been said many, many, many times by many, many, many people about over the past five or six years. It hasn’t happened, nor is it ever going to happen. If it were it would have happened by now. There’s a pattern
.
It will do so mostly for usability reasons.
Regardless of how usable Gnome is (however you choose to define that), technically it simply isn’t up to the job. There’s been a few articles and threads on here that have confirmed just that. That’s a far more difficult nut to crack I’m afraid, but KDE has quite a bit of work to do and some rivers to cross.
Guess what most people will choose.
I’ve never seen a poll or survey that has suggested otherwise – ever.
…but by that time Gnome users will enjoy a system that not only is resonably simple to configure
I thought they had that now
?
That’s the problem. People will rather make pointless and wishful comments, even after years of nothing happening (you’d have thought they would have learned by now), rather than look long and hard at what is actually required. If you point it out, it’s FUD – which isn’t actually what FUD means at all.