Why are Free-Desktop Developers Wedded to Linux? 528
An anonymous reader wonders: "We have been hearing promising predictions like 'This year will be the year of Linux on the desktop' for the last decade. However, the Linux of today seems to be as far away as ever from realizing the expectations of mass adoption we once had for it, without significant growth in home usage since the late 90s. Clearly, if Linux is unable to reproduce a third of Firefox's end user uptake over a much longer time-frame, there are deficiencies with the direction the GNU/Linux/X/Gnome/KDE system has taken. Of course, almost all free software and desktop efforts and development remain unquestioningly oriented around Linux.
Other free-desktop operating system projects which take different and innovative approaches like ReactOS, AROS, Mona and Syllable remain comparatively starved of developers and interest. An often cited reason for using a non-Microsoft OS is to avoid a monoculture, but free-desktop efforts have created a total monoculture around developing and promoting Linux, despite a decade of failure in supplanting Microsoft's proprietorial OSes with it. Why are free-desktop developers neglecting to consider an alternative to the penguin?"
Re:Why? (Score:2, Interesting)
One of the main problems with Linux is the vast number of distributions, all subtly - and often pointlessly - different. While I can understand the reason for why this happens and why it represents what a lot of people like about Linux, you also have to understand that it is one of the reasons that prospective Windows converts don't like it. So many choices, which one should they pick? Which is better between KDE, Gnome, XFCE, etc.? They don't know, and beyond a few fanboys for each nobody else does either because in all honesty none are truly better than the other. And while that's just wonderful for the guy who knows which is best for him, it's just another choice that the clueless average guy has to make blindly and another reason for him to go scampering back to that othe OS where there's no choices, just familiarity and stuff that works (more or less).
Unfortunately I think that if another non-Linux OS were to become popular developer-wise it might turn out the same way. More developers, more conflicting opinions, more forks. I think ultimately the succesful free OS might be one that's put together by a small core of developers who are able to make one solid desktop operating system. No "light version", no "enterprise version", no separate distributions, just one clearly branded and defined OS with all the requisite compatability and virtualisation to make other OS'es programs run on it.
Going slightly offtopic here but I think another issue with Linux is that it's not doing enough to bridge the gap for Windows users to jump ship. It's coming from the other side of the idealogical void from Windows and it seems too self-involved with it's "I'm Free and proud" baggage to really make something that's practical for Windows users to move onto. I'm not talking about applications here, I mean little things. The filesystem of course is one thing that instantly sets the two apart,
This isn't an anti-Linux rant. It's more of a Stop Thinking Outside The Box rant. Outside the box stuff scares people. Get back in the box and think how normal people think dammit.
Re:Not really (Score:3, Interesting)
Not an X-based desktop, a GTK (or Qt)-based desktop. You don't need X. There are GTK+ and Qt for windows.
On the other hand, it does seem that a majority of GNU developers do their work on Linux. And why not? It is arguably the "best" Unix out there. Yes, other Unices have features that Linux doesn't, but I don't think any of them have as many of them as Linux has that they don't. The BSDs are very close (some closer than others) but Linux is the "kitchen sink" kernel - of course you can modularize it or just not build things into it, so don't see that as a drawback. Besides, blowing a couple more megabytes of memory to have more functionality doesn't bother many of us any more.
a simple theory (Score:2, Interesting)
When you are creating something that is going to be offered to the general public as "free", the only significant investment you wish to commit is time. Oddly enough, time is the only resource we as human beings will always run out of, plus we do not know how much time is allotted to any of us, and therefore its value cannot be calculated (even though lawyers sure seem know how to put a price-tag on it).
Re:Why? (Score:2, Interesting)
Mayhaps they are thinking "Outside the box" because the box is a shape that is displeasing, and also is ON FIRE.
The windows way of handling filesystems and drives is more familiar to more people, true... but it's also kind of brain damaged (Example: No distinction between Hard Drives and Partitions in the naming schema). Also, people are either A)Technically illiterate, in which case they navigate the computer by set, static procedures, thus making ANY change of directory harder, but also meaning that keeping some similarities doesn't actually help. B)Technically competant, in which case, they can learn a new directory structure pretty quickly.
And, again. ON FIRE. A Windows box (XP Corporate) that is not running signifigant antivirus and antispyware can be locked up and owned in less than 10 seconds, remote execution through webpage. I SAW IT HAPPEN. Before I updated it, user miskeyed a search site, got a hostile webpage, and BOOM. Had to reformat.
Re:Not really (Score:2, Interesting)
But as it is GNOME (Sun makes considerable work on GNOME) and KDE works just as good at least on Solaris so I don't know what this article is all about.
Re:Not really (Score:4, Interesting)
It's slow mostly because it takes a noticeable time to start processes, and this bothers me, as it's something I do a lot. Also, the GUI takes up so much memory that there is less of it left to get work done with. Once this gets up to the point where it starts swapping a lot, obviously productivity is out of the window.
It's a hassle, because, although a lot of open source software technically works on it, not all of it is readily available. At least at the time I still used it (the situation may have improved since), there were fink, darwinports, and pkgsrc, each supporting some packages but not everything I wanted (pkgsrc worked best for me, but didn't provide binaries for OS X). Having to use different package managers and having to compile things from source are terrible time wasters. The software that Apple ships is either different from what I'm used to from other *nix systems, or it's the same software, but often an older version, which caused further problems.
Also keeping the software up to date is a nightmare when some of it is integrated with Apple's updater (which keeps pushing "updates" for software I don't have or want), some of it is integrated with some open-source package manager (fink and friends), some of it comes with custom updaters, and some of it doesn't have any update mechanism at all.
The final straw was that Tiger broke the ext2 driver, meaning the end of sharing files between OS X and Linux. Yes, Linux supports HFS+, but the interaction between the Linux HFS+ driver and Apple's fsck has given me...bad results in the past, so I'm not going there again.
Of course, none of this means that OS X doesn't look gorgeous and isn't a great OS if you just want to use the great software that Apple ships with it, and maybe a handful of third-party apps. However, for a command-line junkie like me, GNU/Linux beats OS X hands down.
Monopoly Lock On the Desktop (Score:3, Interesting)
I am no Mac fan but I actually think that apple currently has the ability to shake the market to it's core. They now have a intel version of the operating system, increase the driver support and put it on the shelves and I think it could really create a explosive impact on the home desktop industry.
Re:2 answers (Score:3, Interesting)
I know this is not a very detailed answer, and certainly this is not the only thing that i saw that wasn't quite right. Its the best example I can think of this far down the road.
Re:Not really (Score:3, Interesting)
I'll ask just the opposite. (Score:3, Interesting)
Re:Some do (Score:3, Interesting)
Former expert windows users sometimes prefer KDE, because it resembles mswindows better.
Gnome is, in my opinion, better for completely new users.
The magic bullet is that, Kubuntu for switchers, and Ubuntu for everybody else.
Re:Not really (Score:3, Interesting)
Re:Not really (Score:4, Interesting)
Yeah, that's probably about right. I imagine most of the GNU development was done under Solaris, NextStep, and AIX... in that order. The first time I got my grubby little 11-year-old hands on a Unix shell account, in 1993, it was on a NeXT box. Most of the utilities on that box were GNU utilities... GCC, binutils, tar, gzip, etc. I remember learning to unpack tarballs and running
Once I heard about Linux around 1996-ish, there was no going back. Here was a Unix-type operating system I could install on my own Cyrix 486SX PC, awesome
I haven't seen anything come along that's more versatile and all-around better than Linux. Sure, I think OpenBSD is great for ultra-secure servers, and they've been doing fabulous things with wireless driver support recently. Some Linux distros (cough, Mandrake, cough) have gotten way too far out on the bleeding-edge features curve and had stability and configuration problems.
But overall Linux has become everything I'd hoped it would be and more: free, good hardware support, well-documented, high performance, good community support, and UBIQUITOUS (my wireless router runs Linux, and I'm sorely tempted to put Linux on my girlfriend's iPod).
Re:Not really (Score:3, Interesting)
Yeah, that's probably about right. I imagine most of the GNU development was done under Solaris, NextStep, and AIX... in that order.
Actually, a lot of the machines at the AI Lab/FSF back in that era were DECStations running Ultrix. There was at least one AIX box (hal.gnu.ai.mit.edu, IIRC) and handful of HP/UX boxes, some BSD4.3 and a smattering of SGI boxes running IRIX. Most of the Sun boxes were still running SunOS 4.x and not Solaris. There was one NeXT cube that I remember.
Re:Not really (Score:3, Interesting)
All this assumption, and I KNOW of different OSes and hardware compatibility, etc. But, like it or not, Windows is a de facto standard, so a device can be honestly claimed to not work if it doesn't work with Windows.
All this aside, when I buy a device I also do my homework to see if it will work in Linux, and that can be a difficult and annoying process. Try buying a scanner sometime when you are looking for specific features that are only found on certain models of scanners. You have to browse the web looking up names of scanners, and their feature sets, until you find the scanners that will meet the functionality requirements. Then you have to cross-ref this with the Sane list of supported devices, which uses European model names for scanners and has some scanners listed as supported with reduced functionality. Then you gamble on a scanner's supported functionality matching your needs, and you also gamble on the scanner's functionality actually working as advertised by Sane. (Not to bash the Sane guys, they do good work but scanning is may latest hardware issue). Repeat this process for all hardware you wish to use, and consider that you can't upgrade some drivers in your Linux system without upgrading the kernel, which can be a problem.... ugh. Frankly the whole situation is annoying.
Re:BSD (Score:3, Interesting)
Maybe its not about being screwed. You obviously feel you or others have been "screwed" by people using you BSD licensed code in ways that the license clearly intended. Maybe sometimes people want to grant others the freedom to create closed source derivatives. Sometimes I want that people to have that freedom with my code. Sometimes I don't.
For me, it depends on the project. For example, I have written a frontend to Access and SQLite files. On a side not I released a new version today. See the link in my signature. It is GPLed. I wrote this originally to teach myself C#. I then began to use it to deal with access databases and someone pays me to deal with such things. It is a tool that makes my job easier. I give it away in the hopes that it will benefit others. This costs me nothing and if I can get other to try it, I can better improve it with there feedback. Perhaps I will get a patch from someone with improvements.
Now I have other things I would not mind under a closed source license. Most of these things are incomplete programs. For example, I have written some project and file templates for SharpDevelop [icsharpcode.net], an Open Source IDE for .NET. I would very much like to know that people have used those templates in closed source code. I'd like to see my templates used to make classes that were Open Source as well, but its nice to be useful in general. I have other snippets that I will soon release on SourceForge under a BSD license. They demonstrate patterns and are personally researched and developed best practices. I hope to avoid others the trouble I have gone through.
The BSD license was specifically developed to allow code to be used in a closed source application. Anyone that contributes to a BSD project should know this.