“Forget SCO & Microsoft. The single biggest thing keeping Linux from the desktop is Linux itself. It’s time to stop adding new features, and finish what’s there.” Read the 3-page editorial at ExtremeTech. Our Take: While there is no such thing as “finished software”, as software by definition can always be improved, it is true that Linux projects (especially desktop apps) need to outline and release full-featured, solid, well tested versions of their software and offer less smaller ones that require the user to constantly upgrade and be in a state of flactuation. What is needed in smaller projects are real release engineers that can outline goals and releases wisely (the Gnome project, PostgreSQL and Apache are good on that for example, smaller projects could learn a thing or two from these bigger projects).
Linux: We Have Met the Enemy, and He is Us
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
That is SO true
Where are the GUI tools? All over the place. Apparantly he’s never heard of Google or Freshmeat.
And he acknowledges he’s using enterprise tools (that would cost him thousands on Windows), yet he still complains that its “hard”.
Most projects are thought out well. Sure some are big messes, but those are usually one-man volunteer projects.
>>and be in a state of flactuation
They make creams for that.
A lot of articles have tip toed around this. I really, really hope these ideas start caching on with developers. We need a shift in mentality. Take the GNU mindset, for example; not only should you intimately be acquainted with your computer to get it working, but you should also see that user-friendliness is just another way of saying dumb. Or is that the FSF? In any event, I know there’s a lot of closed-mindedness about this stuff in these circles. Talk about a lack of creativity, an overabundance of excuses coupled with narcissism, and an inflated aire of superiority. On beginning with Linux, I tried to see things the GNU way, but these days it just sounds like excuses. If people are going to be touting free software, why not also make it decent quality without having to have it installed by default by a top distro?
Not that I don’t like what GNU and the FSF are doing. I do, and very much so. I’m just saying the general attitude rings of a time when a lot of developers were made fun of in school, and their only defense was trying to convince people they were dumb.
I agree for the full 100 % (surprise, surprise from the author of the Hell & Bliss articles
).
There’s not much more to say, he totally got it right.
ABSOLUTELY. I have been saying this everywhere I go online (Slashdot etc.) for YEARS. 90% of the users don’t need 10% of the advanced features most apps offer, but we need the *basics* to be ROCK SOLID, well-designed, with a half-decent useability and GUI and a simple install. If we had that, we’d rule the world. Bill Gates is quaking in his boots that the Linux community wakes up and actually finishes what they start instead of getting bored of their 80% complete SourceForge project and let it die as they start another. We leave all the hard work (boring stuff like bug fixes and GUI work) to someone else, but often there IS noone else. So the projects suffer.
From the article:
“Windows may be the devil, but at least you can easily figure out where an application installs itself.”
I really like the author to tell me how I can know the list of files installed by an application on Windows. Kind of “rpm -ql …”
Or which file will be installed by a given application (“rpm -qpl …”).
MythTV works fine once you have installed all of the required software and drivers. It says right up front what you will need. It isn’t the MythTV developers job to make sure your kernel has an ALSA driver, and if it were, there’s no way they’d ever get anything done on their actual application.
There are graphical interfaces for MySQL, but none of them are going to save you any time when all you are doing is catting a schema into your database.
The application developers could include the dependencies with their software, but what if you already have the dependencies installed and they are the version required for some other software? If you use a package-based distribution, perhaps the distribution maintainers will think all of this through for you. If they haven’t, and you can’t or won’t go hunting for the dependencies yourself, you need to alert them to the problem or switch distributions.
A lot of software out there thankfully makes no assumption about which distribution you have. The README prescribes the versions of the software you need, and before you type ./configure you verify that they are there. It is somewhat time consuming, yes, but it means I can run MythTV on anything from Red Hat to Arch Linux.
So, what is this about us being our own enemy? What does that mean? It sounds like maybe Linux might be your enemy, or perhaps your distribution in particular, but I don’t see how that hurts Linux.
These kind of articles strike me in the heart and do this also to Linux. People like Dave Salvator must understand the power of Linux and how it is developed. Dave wants the holy grail for him alone. That is his right, but he should work on that himself (and hopefully give some back to the Linux community). Linux is there for us all and we are blessed by having it around just as it is today. And tomorrow it will even be better thanks to the contributions of a variety of hard working people.
Actually, I’ve spoken with RMS’ right hand man (Brad I think), and he acknowledges that there’s a lot of user-unfriendliness in free software, and that’s an issue they are trying to deal with… GNU Hurd was always intended to be Mac-like in its ease-of-use. Until that happens, encourage coders to read the Gnome Human Interface Guidelines, etc. and submit changes to projects to help them be more user-friendly. I’ve done it, and the programmers seem very friendly and want to fix their app – until they just give up from lack of community interest and input and let it die unfinished (see my previous post for more bitterness and details)…
“These kind of articles strike me in the heart and do this also to Linux. People like Dave Salvator must understand the power of Linux and how it is developed. Dave wants the holy grail for him alone. That is his right, but he should work on that himself (and hopefully give some back to the Linux community). Linux is there for us all and we are blessed by having it around just as it is today. And tomorrow it will even be better thanks to the contributions of a variety of hard working people.”
I almost wanna call this a troll. It’s just too near-sighted and developer centric. How do we win the desktop like that? Much less how do we keep the media’s attention if we stay a huge, popular hobby OS?
Our Take: While there is no such thing as “finished software”, as software by definition can always be improved,…
Here’s an essential problem. By maintaining and conveying this belief to others in a public forum, you encourage users and developers to maintain false truths of what completeness is.
You argue that software can always be improved but this isn’t true. You can only improve something until there is nothing more to change that provides reasonable additional benefit. Much of today’s professional software tools have already reached the point of completeness and are now headed backwards down the hill to destruction in the form of bloat and overcomplexity (go ahead and name some for me, it’s easy).
A software design requires a goal and a goal is a finish line. For example: “to provide a photographic retouching, editing and painting tool that is fast and easy to use, with all the features that professionals expect to have.” Once you get to that goal, you should stop. However, most companies rarely stop there. Due to the structure of their business and their true goal of increasing market share and increasing share value, they will continue to add new features, often unneccesary and counter productive features (such as Photoshop’s file management features and about 45% of Microsoft Office) just to be able to release a “new” version so people will buy again. This isn’t just built-in obsolecense. This is “the never finished” syndrome. This is the wrong way, the wrong reason, to go about providing new functions to users and this is why software appears to never be complete; it is forced into being that way, quite against the natural flow of things (most things have a beginning and an end).
When the goal is reached, the product should be complete. Sure, it’s good to release fixes to problems and improve things as needed to maintain the software on new OSes and to provide as efficient a workflow as possible for users. This is not what most software developers do, though. Once they reach “completion” they continue to poke and prod the software, adding marginal value and often ruining the product through bloat and rearrangement for the sake of appearances.
The Open Source world and Linux has this problem, but in a different way. There does not seem to be clear goals for Linux and its components to follow in the first place. The clearest goal we can find is “we can do that too.” I think the original goal of creating Linux was for a school assignment and/or to prove that “it could be done.” (please be lenient on me, my Linux historical knowledge isn’t exactly the greatest, but I’m sure if you don’t try to pick at little details, you can agree about the general idea)
The biggest problem with software is the very belief that it is never finished. The unwillingness of developers to get to a point of completeness and then stop fiddling with it (in the commercial sector, the product just keeps changing for the sake of new releases, after a certain point, while in the open source sector many projects just stop altogether for lack of interest). Why is this? Laziness? If a company has a product and they can resell it every year, they usually try to do so, especially if that means that they need not create anything new (or they buy someone else’s product, which is the ultimate in laziness, such as Adobe buying Cool Edit instead of making their own audio software – how well do you think the first version of their “new” product will integrate with their other products?).
Sorry, folks, but it’s the general attitudes about computers and software overall that are the real problem and few people are willing to admit or are even able to see that this is so. Commercially and non-commercially. Computers and software are not created for specific purposes and no one ever allows them to be completed because they don’t believe there is such a state. Without a clear goal and a clear finishing point, you can’t ever make anything truely good. There are a lot of almost great products being ruined by “continuing progress.” That’s about at good as it gets with the current attitudes. I bet that the majority of people here will think I’m just 100% full of crap and miss the point entirely because they are just so used to thinking about things the way they do, without really spending any time thinking about why. Look around you and see how other industries and products have stabilized and matured. Computing needs to do the same and it will not happen until people sit still for a moment and think, really think, about what the purposes of these tools are and stop running towards an infinite goal.
Your kind of attitude is the problem.
//
I almost wanna call this a troll. It’s just too near-sighted and developer centric. How do we win the desktop like that? Much less how do we keep the media’s attention if we stay a huge, popular hobby OS?
//
How do we win the desktop? Maybe we dont want to. The community in general doesnt want to sacrafice power and freedom just to get grandma to run linux. The ideal situation is for people to know more about basic computer administration.
How do we get the media’s attention? If the media will only report on whats popular and flashy, then we don’t need that kind of publicity either.
A manager I had once said to me “Sometimes hard things are hard.” Making your Linux system into a PVR can be hard. You need to have infrared drivers configured for your remote. You need to have a database set up. You need to make sure that your tuner card is supported and that you have a driver compiled for your kernel. If you decide you want ALSA, you need to make sure it is installed properly for your sound card.
This is a specialized use of your hardware. Maybe you can get a PVR up in 2 minutes on Windows, I have no idea, but half of your problem there was solved when you installed your hardware and had supported drivers because of deals Microsoft made with those companies, deals that Linux developers usually aren’t in a position to make.
There are several separate issues here, and for now, they must be dealt with separately. If your hardware falls into a certain, well-defined set, and so do your needs, perhaps a distribution can be made for this specialized purpose.
There is no harm in trying to come up with a solution for these problems, but it doesn’t help to lump them all together into one big “someone else’s fault”. We are lucky to have come to a point where we have all of these components working well enough that we can use them together for something like a PVR. A lot of work goes into each of them, and I don’t think even the developers want them to be hard to use.
For desktop software, user-friendliness is absolutely critical. And even on the server end, the idea that a network admin in hurrying to bring back a server online has to go hunting for stuff all over the place is inefficient.
He wants Linux apps to be finished because not to far from now SCO will win its case and Linux will be dead, and if the SCO suit doesnt kill it MS will sue Linus Torvalds for cloning and violating their IP, so either way in at least 2 years Linux is dead. You want apps that just work use Windows or a Mac. Leave Linux and its ragtag band of zealots alone.
I am sorry, but I do not agree. You over-philosophize here. Yes, there are over-complicated, bloated software around, e.g. MS Office and OpenOffice.org, but I am talking about apps like Gedit or Galeon or Konqueror. These apps (and hundreds more) are not “finished” in my mind, I find errors in their design, implementation and final output, all the time. I am not happy with them, they have a sour feeling of being “not well designed, not well implemented”, simply because they are released too often, without a greater plan behind them.
You can argue this all day, you have some good points, but we are not necessarily are talking about the same thing.
It’s awesome, but what about the legal problems
I agree with the writer in general terms. However, after witnessing how fast and stable open source apps (Mozilla, OpenOffice) are on Windows I wonder if X11 is the real culprit here?
Oh, and btw, regarding “software improvement”, you Jace seem to think only about “feature enhancements”, while “improvement” can just be code optimizing, architectural changes, code cleanup etc. I think your problem is that you see the whole thing as “feature bloatware”, while this is not necessarily what we are talking about here!
The author suggests people should look at what they do not as a project, but as a product. This is just misguided. If you want it offered as a product, look to third parties (like Ximian with gnome) who will actually sell it as a product. (BTW this is one reason why it IS best to wait until something is offered in a distro)
Have these people never heard of the expression “never look a gift horse in the mouth”? Talk about ungrateful.
Exactly.
Listen there are fine established “products” out there that are “finished” in the sense that the author speaks of.
The problem is three-fold. Projects feel the need to release products in an unfinished state. They make no bones about the fact that these products are unstable or in a state of flux but people still insist on using them as if they were released finished products (Rhythmbox anyone?).
Not only that, many distros insist on shipping the very latest release of certain software products regardless of whether or not they are truly finished by any sense of the word. Even when gnome 2.x had just been released as a developer’s platform certain distros were releasing products including these packages.
Still, there is the problem of the hordes of users that insist that distros include these unfinished products. They demand that this or that latest release of such – and – such product goes in. If they don’t then you hear people complain the way some people do about many desktop distros not including the very leading edge KDE setups.
If more distros stuck to their guns about only releasing stable versions of the packages they include and the users did not raise such a high fuss everytime the very bleeding edge of version of mister random app was not inlcuded with their distro then this might end up being a null issue.
Until then, I can always go back to listening to people complain about lame Debian is because stable is so far behind the usual insane curve of FSF development packages.
I liked the first page. He says:
I took my time. I read all the documentation. I diagnosed and fixed multitudinous glitches. But instead of success, I’m sleep-deprived, frustrated, and ready to chuck the box out the window.
There’s the key: documentation. He read it all and still had problems. Good succinct accurate docs are what it’s all about.
On the 2nd page:
I wasn’t so lucky with MySQL. The yum-based installation was incomplete – and the program is so Byzantine I still haven’t gotten it working right. Yes, I know it’s a heavy-duty enterprise-class database server, but where are the graphical configuration tools?
Sounds like a combination of a yum error, along with issues with MySQL docs.
Page 3:
I’m just asking that Linux application developers think their projects through from A to Z, with Z representing a polished product that installs successfully with minimal fuss.
That is to say: better docs.
{/rant}
What legal problems? Does SCO own it too?
Urgh. What a waste of time. 3 pages of bitching because he couldn’t get alpha qulity software running on his system. It’s freakin version 0.11. It doesn’t matter if it’s on Linux, or wrapped up in a nice, “friendly” InstallShield on Window, you’re lucky to get anything v0.11 running at all, let alone have flawless installation. Sheesh.
We all know you’re a big Gnomie fan, but can you bother, for the sake of editorial sanity, to say Gnome/KDE. Unless of course you think KDE has “childish and immature” release cycles and I’m sure we’d be all interested in learning how and why.
As far as the article goes, he doesn’t seem familiar with Softwane Engineering paradigm (for those who took it in college). Most of OO projects are released on a feature-based skeleton. So instead of developing an app (like what game companies do) and then releasing it after 3 years of development span with all featuers (and bugs!), the open source model adopts a more rapid release philosophy because OO is competitive and you need to constantly add features and improvment to your software to stay afloat.
This is not to discourage source and binary compatibility (a problem many users face), but we should balance compatiblity vs. usuablity so that end users can still enjoy great software without suffering.
In fact, MythTV is a cake-walk to set up if you use apt-get and the repositories provided by Axel Thimm (http://atrpms.physik.fu-berlin.de/).
The difficult parts come from getting all the underlying bits to play nice. (Even that doesn’t have to be all that difficult.) MySQL isn’t hard to set up, lirc can be depending on which hardware you use, if there is a config file for your remote, etc. The TV capture (v4l) drivers can require a bit of twinking to get running right. I never had any problem getting xmltv to work, basically you just tell it where you leave and who your provider is and away it goes.
Basically, MythTV depends on a lot of other projects to provide a complete PVR setup, which is as it should be. The MythTV developers couldn’t hope to reproduce lirc, or capture drivers, or ALSA and nor should they try.
Furthermore, this Dave guy never once posted to the mythtv-users list… where all but the most obscure problems get diagnosed very quickly. (I know, because I just looked at my archive of the list.)
What else annoys me about this article (even more than giving MythTV a bad rep as being too hard) is that it repeats a very common thread amonsgt individuals who like to write about free software. Basically that idea is that whenever the person in questions finds a bug, missing feature, or “design” problem, they claim that the developers need to “finish” or “polish” or “add the features users really want.” The reality is these complainers probably haven’t tried looking into the software… have they asked for that feature? Maybe that bug was a lower priority to fix than some other bug.
The bottom line is just because you can get yourself published on extremetech.com doesn’t mean your needs or perspective jives with those of the majority of the users of the software you’re writing about. And if you aren’t going to avail yourself of the resources the developers put into place (in this case, the mailing list) then where do you get off complaining?
Shows how much you understand or know, you can’t even spell my name properly.
>We all know you’re a big Gnomie fan,
I am not. I also help out XFce, what does that mean? If I was a “gnomie fan”, I would actually use it. I am not.
> but can you bother, for the sake of editorial sanity, to say Gnome/KDE.
Nope. Gnome has better release cycles IMHO (not that KDE’s are bad), and at the end of the day, there are a dozen other projects that have good release cycles, I can not mention them all. I just mentioned three representatives and thats all. There is nothing much more to it, KDE is not much better or worse on this matter.
> Unless of course you think KDE has “childish and immature” release cycles and I’m sure we’d be all interested in learning how and why.
I am sure that you understand that if something is childish, immature and “oh, why daddy?”, this is your comment. Sometimes you should come to my position instead of trolling from the third party chair you are sitting on.
<rant>
I don’t understand why these articles are popular; both for the readers to read or the writers to write. They are pure drivel. They serve no purpose in my opinion but to generate a bunch of Seth-ish whinging and other equally banal comments, which often times are neither insightful nor helpful.
When it comes right down to it, who the hell really cares? I mean, Eugenia complains that GEdit “isn’t finished”, which amazes me since it far surpasses its Window’s counterpart Notepad. So, use something else! It isn’t that hard. Another thing, perhaps the GEdit developer’s scope for the project and how the project turned out are entirely in sync. Just because somebody doesn’t like the color of the icons doesn’t mean it is a bad editor. Does it edit text? Yes. Then it must be a text editor.
Let me give a smaller scope example of the situation as I see it. I hope this makes sense.
I like Vim and don’t care for KDE’s editor, but I’m not going to write up a big, stupid article about how KDE’s editor should never be taken seriously because it doesn’t work like Vim. I’m not going to declare KDE’s editor “unfinished” because it doesn’t have all the functionality of Vim, I’m just going to use Vim. Duh! That concept scales pretty well to the OS level too.
</rant>
I feel much better. Thanks.
This article addresses a very important point about Linux application development. It seems that there are a number of people posting here who don’t seem to want any flavor of Linux to be user-friendly. The author merely suggested that developers of Linux applications take a look at things from a user’s perspective. They just want things to work.
Worldwide there is a potentially huge corporate market for Linux desktops, and great potential for commercial applications once the users of those desktops realize that Linux is a viable alternative OS for their home computers. I suspect that some of these pencil-necked geeks didn’t read what the author was saying, or else they’re afraid that if user-friendly Linux applications are developed, they’ll lose their hobby OS’s.
I liked what he had to say up until he called MySQL an enterprise class database that’s too hard to install…
while I still agree with many of his points, he discredits himself with 1 sentence…. a shame
MySQL has awesome documentation… in fact it’s extremely complete and it should be held up as a shining model of documentation in the free software world.
Mdd: you said it yourself “it’s version 0.11 so it’s unlikely to work”… but that’s the problem he tried to mention. Many apps never reach version 1 !!!
May be the presentation of the author is biased by frustration ( I tried MythTV too with same result) but I like the idea of “finisher”. Until now companies like Ximian are doing some part of it by wrapping GUI around common tools. But I agree that more polishing would help many apps to have the recognition they deserve.
AMEN
AMEN
AMEN
[quote]These apps (and hundreds more) are not “finished” in my mind, I find errors in their design, implementation and final output, all the time. I am not happy with them, they have a sour feeling of being “not well designed, not well implemented”, simply because they are released too often, without a greater plan behind them. [/quote]
Can you be more specific as to what exactly you consider to be poor software design and implementation. The architecture is public, so we can easily see what the design goals and principles are. What issues do you have with them? No one goes about writing thousands of lines of code without a plan. What do you mean by a “greater plan”?
I’m not convinced that the bug/feature ratio is any higher for open source/free software than it is for any other significant software project. Indeed, since the former is generally put together by loose confederations of coders, which is at odds with “generally accepted” software engineering practices, it’s rather remarkable that there is a decent level of cohesion and usability.
The author merely suggested that developers of Linux applications take a look at things from a user’s perspective.
Of course, the author made the assumption that this isn’t already the case. The problem is that anytime someone with a byline can’t get something to work, they assume it’s because free software developers aren’t paying attention to end users.
That’s absurd. Free software developers take feedback in any number of ways, from bug reports to feature requests, user surveys, and so on.
It kills me to hear people complain that free software doesn’t consider the end-user. As if the commercial software industry does? Maybe big corporate customers have pull, but actual users like you and me? No way. Which is not to say that commercial software companies don’t listen to their customers, but go ahead and call Remedy and tell them you don’t like the way their software works. Call M$ and say “I think the menus are confusingly laid out”, see how far you get, even if you are a paying customer.
The bottom line is this guy had a direct line to all the key developers and a supportive user group… which he chose not to even attempt to use before spouting this drivel.
> > On the 2nd page:
> > I wasn’t so lucky with MySQL. The yum-based installation
> > was incomplete – and the program is so Byzantine I still
> > haven’t gotten it working right. Yes, I know it’s a
> > heavy-duty enterprise-class database server, but where
> > are the graphical configuration tools?
>
> Sounds like a combination of a yum error, along with
> issues with MySQL docs.
MySQL has awesome documentation… in fact it’s extremely complete and it should be held up as a shining model of documentation in the free software world.
Sorry. I said it sounded like MySQL doc issues because the author of the article said he’d read the docs but still had trouble. (I’ve never used or tried to install MySQL.)
Hmm…
> The author merely suggested that developers of Linux
> applications take a look at things from a user’s perspective.
> They just want things to work.
But which developers? How many applications does it take to make your PVR work? xmltv, mythtv, lircd, mplayer, mysql, etc. Most of these do what they are supposed to do already. Someone previously mentioned product vs. project. What you want is an easily installable “product” that just works.
Nothing wrong with wanting that. I got MythTV going myself without using anything but tarballs, and could do it again, but sure, I’d use a prepackaged version if I could. I’d be amazed if anyone could make a one-size-fits-all package for it, however. Note that Freevo tries to do all of this anyway; they supply a distribution with its own versions of the programs and libraries used. And it still isn’t going to work for everyone.
I don’t expect average users to understand the difficulties of development; if they did, they’d be developers. They seem to think they have the hang of management, though.
You’re saying that you prefer Gnome release schedules over the KDE ones, in response to an article talking about how products need to be more finished?
KDE release cycles are feature and stability-based. A feature is created and then stabilised and the thing is not released until the developers are satisfied.
Gnome has time-based release schedules, taking snapshots at intervals and releasing them as full releases.
In terms of a more finished, stabilised product, KDE offers it, while Gnome releases are more bleeding-edge and true to their CVS. Both have merits, but venerating such a dubious article and then backing an opposing release mentality is a little poor.
I want it to “just work.”
Common people. It’s nice when everything “just works,” but this is a free OS. You are going to have to expect to put some time and thought into getting it to run the way you want it to run.
Quit thinking of this as a bad thing! It forces you to get inside the system and learn how everything works.
This whole article is a rant by a guy who’s too incompetent to install a PVR and now he’s bitching out the whole linux community cos he’s not willing to spend the time it takes to learn. boohoohoo. He should goto tellmommy.crymeariver.com and suckle the breast that emasculated him.
I’m a hobby developer and I don’t know how to create rpms. According to the author, I shouldn’t create a project on my free time, and create a some program that I think may be usefull unless I’m prepared to figure out how it should be installed on his computer. I admit I’ve been frustrated trying to install some half completed project in the past, but then I waited a few months until my distribution came up with an rpm for it. Linux will never be windows. They are different systems. One is closed and controlled the other is open.
Maybe we should listen to all the microsoft wannabees out there and kill the x-server, have only one widgit style, only one shell, one installer format, etc. But wait then we would have windows (or mac osX).
My point is that Linux isn’t a corporate project but a hobby for lots of home developers, that the corporate types feel they can exploit to make money. It doen’t fit the development model windows uses. And it shouldn’t.
The “extreme” part of extremetech is that they know absolutely nothing yet still pass themselves off as experts.
it’s for professionals and geeks to play around with. Trying to endlessly use Linux as a desktop OS is like trying to endlessly use Windows95 as a server OS. It just isn’t going to work because there are fundamentals of the software dead set against it. That’s why Syllable.sourceforge.net is around.
How do we win the desktop? Maybe we dont want to. The community in general doesnt want to sacrafice power and freedom just to get grandma to run linux. The ideal situation is for people to know more about basic computer administration.
Why do you assume that just because something becomes easier/more intuitive to use, right to the point that “grandma can use Linux” that it would result in Linux being less powerful?
Just because GUI based admin tools and default installers with all the package dependencies resolved, etc. doesn’t mean that you have to ditch the command line tools, manual configurations, etc.
Look at OS X. It is easy, intuitive, and hasn’t any capabilities as it has become more easy. It still has a commandline. It has the ability to compile software. Everything on OS X just works…out of the box…which is the way things SHOULD WORK! If anything, it is even more powerful. The more intuitive something is, from a user standpoint, the more powerful it is.
I agree completely with the author. The Linux community is holding itself back. Maybe not all of it, but a significant portion. Look at Ximian. When you install XD 2, it is pretty easy to install. The Red Carpet installer takes care of the dependency issues for you, and guess what. When it is done…everything just works. Has any capability been lost? No. Mozilla is the same way. Download, compile. Bingo…you are surfing. The same could happen with more open source software projects if the Linux community as a whole would give up their elitist attitude that Linux should be kept for people who like to write all their config files by hand. If you want to do that, fine…that ability doesn’t need to be taken away just because more of projects and Linux as a whole becomes easier to use.
Ease of use does not negate power.
The fact of the matter is the hard-core center of the Linux world doesn’t WANT to win the desktop. They don’t care about beating out MS, they just want a zillion different options for their own personal use. The rest of the world looks at linux as a potential savior from MS while the linux elite want no part of that and would rather navel-gaze and produce a million different applications that all do the same thing. The problem with linux is the old guard and their inability to step up to the plate and do what needs to be done. There is no great personality to stand up play the reluctant leader. Linus? He made it clear long ago he’s just the father of Linux, not it’s parent. And like so many fatherless or motherless children, Linux will continue to evolve and grow old but never really be good at any one thing because it lacks the focus a parent brings. In a nutshell, the Linux community is amazingly talented but just as gutless.
“” I mean, Eugenia complains that GEdit “isn’t finished”, which amazes me since it far surpasses its Window’s counterpart Notepad.””
Ok, this is just unfair. Notepad’s development cycle is, as far as I can see, finished and has been for some time. It does what it says on the tin extremely well, ie It provides a pad on which to type…*drumroll*…notes. If it was trying to pass itself off as a fully featured word processor/text editor then maybe you’d have a point, but it isn’t (AFAIK the last feature to be added was the ability to insert a date/time stamp). It may well be that you’ve picked the single piece of software on the Windows OS that actually feels totally complete and fulfils all requirements you’d expect from it (I have zero sympathy for anyone using an application called Notepad to try and view large files, unfortunately MS have Windows setup so that Notepad is the default viewer independent of file size, that’s a Windows problem, not an application problem).
Back on topic :>
I think most of us that use Linux, or the other hobby OS know that when everything is going well it goes really well, but when things go bad they go terrible. Eg I had no problems building my system from scratch, then lost quite a few hours trying to compile ChromiumGL (Neat game, pain in the ass to get working).
As with anything else, just because xxx app worked out of the box, or installed easily, for you doesn’t mean it’s going to be a cakewalk for someone else. It’s a bit arrogant to start insinuating people are dumb when all that’s happened is they’ve zigged when they should have zagged. Sometimes Linux itself can be a bit like writing code, in that it’s usually much simpler to see errors in someone else’s code than in your own (Something psychological going on I believe, you see what you expect to see instead of what’s actually there), likewise it’s usually easier to see what someone else is doing wrong with their setup than figure out what’s wrong with your own after staring at it for a while.
I want it to “just work.”
Common people. It’s nice when everything “just works,” but this is a free OS. You are going to have to expect to put some time and thought into getting it to run the way you want it to run.
Quit thinking of this as a bad thing! It forces you to get inside the system and learn how everything works.
Most people don’t want to “get inside the system and learn how everything works.”
Further, there is other free software that [i]doesn’t> require the user to “get inside the system.” Why is that people such as yourself seem resistant (don’t mean to be insulting, but there seems to be resistance to idea of making opensource software more polished and robust) to the idea of making application installs polished/intuitive/easy? It seems very backward. Its almost like wanting to run DOS again at times.
>”I mean, Eugenia complains that GEdit “isn’t finished”, which amazes me since it far surpasses its Window’s counterpart Notepad.””
Gedit is not Notepad, it has more goals and features than Notepad. You can not compare applications, each one is different, regarding “completeness”. But the point is, I find Gedit slow and unoptimized compared to other applications in general. I don’t have any problems with its feature-set.
Again you people, you don’t understand that “completeness” is NOT just features, it is a whole lot of other things that can please the user. Each user needs different things. Completeness is when an app can please 90% of the users with its 90% of its requested features and stability/optimizations etc.
You’ve got to be kidding! I’m using SuSE 8.0 as a replacement for Windows 98, and not just for e-mail and web surfing, either. I have a great money manager (Kapital) and a terrific replacement for Cakewalk called Rosegarden 4, which I use for multi-track audio and MIDI recording. You need to qualify your blaket statements for you go shooting your mouth off like that!
Linux is not for the desktop?!
Nothing!… Nothing!… Nothing in all of the forums that cover Linux and OSS/free software irritates me more than the total copout that “Linux is not for the desktop.”
Read your history, when Linus started the “little OS that could” project more than a decade ago, he WAS NOT making a server OS. Rather began developing a workstation OS that he could use to connect to servers on campus.
(Note that the distinction between a workstation and a server wrt OS is a red herring to begin with. It is only in the closed/proprietary world where this distinction holds meaning — per seat licensing bundled software etc.)
Lets take a little inventory, shall we?
IBM- likes Linux on S390, has also ported Linux to a WRISTWATCH
Sharp- Linux on a pda anyone
Tivo- time shift your TV viewing with Linux
Appliances (not just for the home anymore)- get yourself a turnkey network attached storage device running Linux.
Bigger Network? – give Cisco the boot, run a Linux based router
Sun- check out Mad Hatter project
Motorola- we got Linux cell phones
Need some computational heavy lifting assemble a Beowulf cluster
Want to keep costs down while supporting users who may be ILLITERATE?… what you need is a Simputer!
(I’m serious here, google Simputer and you will find that literacy is not a prerequisite for use!)
It would seem to me that your argument boils down to a fallacy: Because Linux can do everything else, it can’t do grandma’s solitair & email.
B_llsh_t!
..
switch vowels in username to email me
> I like Vim and don’t care for KDE’s editor
You may be interested to know that the Vim KPart will be included in the 3.2 release.
Rich.
We don’t need to dumb Linux down. Linux will get to those that don’t care soon enough.
Hear, Hear!
Mario
Perhaps the two of you missed my point.
GEdit is not Notepad
No kidding, but Notepad has been deemed complete; even though it is pretty useless. Why can you accept that fact and not accept the fact that the current version of GEdit is done, or Linux for that matter? It may not be done to your standards, but since it’s a free program, I don’t recall that being a requisite.
So, using these editors as an example, if GEdit doesn’t suit your needs (as it doesn’t suit mine) then don’t use it. It doesn’t do anybody any good to denounce the effort placed into creating the program. The program may be exactly what the original developer wanted it to be. Do you think by complaining about it the developer(s) working on GEdit is going to come around and program GEdit into your idea of the pinnacle of editors? I don’t. This concept scales very well, as I said, to the OS level too.
Also, I’m not talking about features either, but the product as a whole. In the corporate environment, a program is finished when management/sales/marketing says it’s finished. In open source, it’s finished when the developer decides to stop working on it. That most likely will never change, so there’s no point harping on it, in my opinion. Use OSS if it hoists your flag, or use something else if it doesn’t. Rancorous articles written out of spite and irritation don’t really help anybody, I think.
Maybe I am wrong, but I feel some GNULinux users feel thier OS of choice is superior simply because the ordinary user can’t configure it. “It must be superior if you need superior knowledge to use it.”
With this attitude GNULinux will never gain the market share needed to have the drivers or apps required to compete with MS, as far as being as useful as MS products. Which is what computers are for, using.
How do we win the desktop? Maybe we dont want to. The community in general doesnt want to sacrafice power and freedom just to get grandma to run linux. The ideal situation is for people to know more about basic computer administration. >>>
Winning the desktop is the only way to slay Microsoft. Until Linux comes up with with something that is as easy as Windows but more bulletproof, Microsoft will have the upper hand.
Why the fsck should my 78 yearold blind in one eye, dyslexic great aunt have to learn how to admin her own system to set it up, get on line and send email? I honestly don’t think she can. She can carve wood and make beautiful handicrafts like nobody’s business, but she’s constantly getting her computer borked up.
Also, my IT department at work does not want people to have basic admin skills. They want to go to a box and have a 1 or 2 step fixit. They do what they can to lock the user out of making changes.
Finally, teaching the 200 or so people at my work how to be their own admins takes time and money that we just don’t have. Better to spend that time and money to make us better at using SQL, Dreamweaver, or the Microsoft office suite.
And frankly, some of my co-workers, though smart in their own ways will never figure out how to admin a system. I my job I’m the person who gets called before you call Systems. I deal with PhD’s who need hand holding on Windows. I would never call any of them stupid.
Frankly, I have no interest at all, zero, zip, zilch, none, nada, in learning how to sysadmin. I created an account for myself and my husband in OS X and set up a similar thing in YDL and that’s really as much as I give a frak about. I care more about writing articles or editing video and would rather devote my free time to those.
I despise Microsoft and their rotten business tactics. They are the sole reason I switched to OS X. I want a desktop that is MS free. I am working hard to see that never a penny of my money goes to them.
But at the same time, my interests and priotories mandate that I have an up and running computer for my creative endeavors. So, not only do I not have the interest to be a sysadmin expert, I don’t have the time.
Sure some people probably feel that way, but look at this particular project : he wants to set up a media jukebox server, but setting up any server let alone one with a database backend takes quite some specialist knowledge. Then he whines about how it’s to difficult and incomplete but neglects to mention this is pre-beta software (v 0.11.) To make the affront complete he also wants this for free, and immediately (reasonable chap right ?)
Usability is important, reasonable expectations are more important.
What if you could download the head branch of MS Office? Wouldn’t need the latest IE, comctl32, and dozens of other kernel DLLs? Of course you would. Releases from developers don’t target your Mum, but packagers.
If you want something that works out of the box, buy RedHat or another favorite distribution. Or use the FreeBSD ports system. You won’t get the latest, might you might get the stablest.
If you just want to try the latest greatest, well fine, you can do that, but it might not be easy. And it never will be.
My favourite quote of the article:
So what’s the solution? How about adopting a single word-command – I suggest “install” to compile and set-up a package. This could be done through a Bash script.
Lemme see:
$ which install
/usr/bin/install
Oh dear, this guy is trying to set up a PVR (which is a non-trivial task on any OS, let alone Linux), but doesn’t know that Linux comes with an “install” script already (it’s a simple file installer shell script, nothing special, but has been around on UNIX, BSD and Linux for many, many years).
You can’t ask the OSS to get rid of the project mentality, because many things are just that. Projects. No one asked you to take them seriously. Someone, maybe in their basement, decided to try something or wanted something and set out on it. Maybe it became popular but that doesn’t mean that you get to tell the developer how to act. Thats the joy and curse of OSS. Although if you really cared its entierly possible you could get involved in the project yourself.
And as for his example, while I’ve never used Yellow Dog, it sounds like he realy only tried Yellow Dog. Which seems odd to me because it’s kind of not quite one of the top distros. Its good from what I hear, but for putting together something more experimental, you might want to try something with mroe experimental options like gentoo. Which, also due to its comunity, is pretty big and has a lot to offer.
Essentially it sounds like he tried one package on one distro and determined his experience of linux. What about other packages
surely there are other media packages out there. Xine fills all my movie needs and xmms is fine for music. And there are a ton of other options. Maybe he should have tried something else… something maybe as he wished, more profesional.
Nope, I think in fact you’re the one missing MY point.
You took two applications with differing target audiences, that are in VERY different stages of their life cycle and attempted to use a comparison between the two to justify your arguments.
Notepad’s feature set is AFAIK frozen, and has been for a few years.
http://gedit.sourceforge.net/features.php
Top line:
“Although new features are always under development, currently gedit has these features:”
If they are still contemplating adding new features then the application is NOT complete. It might have reached the same (Or better) level of stability, but in no way can you compare gedit’s level of completion to that of notepad simply because gedit’s feature set is expected to expand.
In any case using gedit as an example simply highlights the author’s point. In what universe does a “lightweight text editor”, as the gedit developers describe it, need a plugin system? It’s symptomatic of current application development (In both proprietary and OSS spheres) that applications gather features they simply do not need. More features = More code = More bugs. The software world would be simpler if occasionally developers stepped back and asked themselves “Do we actually need that cool new feature?” instead of “Yeah, that’s cool, let’s put it in”.
Nope. Gnome has better release cycles IMHO (not that KDE’s are bad), and at the end of the day, there are a dozen other projects that have good release cycles, I can not mention them all. I just mentioned three representatives and thats all. There is nothing much more to it, KDE is not much better or worse on this matter.
———————
Well, your humble opinion doesn’t matter. Why should I care about an _opinion_ ? If you have facts to backup your claims, fine let’s hear them.
As far as “I can not mention them all”. This is a legitimate reason IMO. But are we ever going to see you using KDE as an example instead of Gnome.. just for a change?
Judging by the mentality lots of ‘linux experts’ here that love things HARD to use, I’ll say Microsoft will continue dominate the desktop way beyond 2006.
Many people are describing MS Windows as user friendly.
I would say that a computersystem is “User-friendly” when the user can do what the user intended to do in an easy to learn and productive maner.
Many people argue that a system is user-friendly when installation, configuration and uninstallation is easy to perform. MS has made great efforts in this area and I would say that Windows is the best in this respect. But this is not what the user is supposed to do with the computer is it?
A user-friendly system should be configured and ready out of the box. Further more it should be tuned to do EXACTLY what the user intend for it to do, nothing more, nothing less.
In this respect Linux and OSS excells because the product developer has so much more power over the system than what he/she would have using MS and non OSS.
Now Linix and OSS may have a steeper learning curve but that’s not really an issue for developers. In the end this enebles them to develop a product that is better tuned for its task than had they used MS and non OSS.
If you decide to build a PVR media server from component software parts, you should at least expect to have problems and be frustrated from time to time. If you can’t cope with that, buy a f***in’ Tivo!!!
Seriously, one of the freedoms that OSS & Linux brings is choice, and with that freedom of choice comes the responsibility of accepting that just because you’re good at the job *you* do, doesn’t mean you’re a naturally gifted software developer too.
I choose to use Redhat 9 & XD2 because everything installs fine – it’s not the latest and greatest but it works well. If I wanted the latest and greatest I’d get Gentoo or LFS and break out the strong coffee!
This guy’s just a normal/hobby user who’s seen a project in it’s early stages and sees the promise of what it will bring tomorrow, but wants it today. There’s a lot of that – I’m guilty of it myself – but sometimes you just have to sit back and wait for the good stuff. It will come when it’s ready, not when you decide you want it. Can’t really ask more than that when it doesn’t cost you a bean!
Eugenia, I didn’t recognize you at first with the “ELQ” instead of your full name 😉
These apps (and hundreds more) are not “finished” in my mind, I find errors in their design, implementation and final output, all the time. I am not happy with them, they have a sour feeling of being “not well designed, not well implemented”, simply because they are released too often, without a greater plan behind them.
I agree 100% with this. I think software completeness includes features AND what you’re describing here. What you said here is why I do not use Linux.
My comments were soley targeted at your statement about software never being finished, as though that were a natural law. It is not. It is an artificial limitation imposed by misdirected thinking.
1) OSS software is developed out in the open, warts and all. What we see from commercial software vendors in the form of betas is their product almost feature complete, but still needs debugging.
If Windows Longhorn, for example was being developed in an open fashion and users had access to the source code, discussion forums and internal email traffic, one could come to the same conclusion simply based on the fact that one couldn’t get the source code to compile and it isn’t feature complete, it obviously sucks.
Sure, a large number of developers make out their little gizmo does more than it actually does and claim that it can replace ayz when in actual fact. The prime example of this would be the constant, “GIMP can replace Photoshop”.
Most developers, however, realise the limitation of their software and clearly label, “UNSTABLE!”, “UNDER DEVELOPMENT!”, “COULD MAKE YOUR MACHINE EXPLODE!”. Good examples of quality opensource software that have matured into very useful pieces include; Pan, X-Chat, Mozilla, Apache etc.
2) A large number of developers have no direction set down. Lets look at GIMP for example, it simply wonders around in circles until someone kicks it and suddenly we get a response, “oh, its almost there!”. Well, lay down what features and things that need to be completed.
When a new version if released, it should be released fully optimised and to the dot of the plan. Lets look at Abiword for example, 2.0dev is one big mess. Coders saying, “oh, don’t worry about that, we’ll complete it some time down the track”.
On one hand these “coders” want people to use their application yet on the other hand they aren’t willing to put a bit of professionalism into the whole development process.
3) The OSS community need to choose a human interface guideline and stick to it, whether it be GNOME HIG or KDE’s, or even Apples HIG, it doesn’t matter.
Having two toolkits is alright however having to different HIGs make applications running on a desktop look like a dogs breakfast.
The user wants to load up Pan and find that tbe about menu selction is in the same place for Konqueror as it is for Pan, or the preferences menu item is in the same menu location it is for Pan.
Same terms need to be consistantly used over and over again. Calling something preferences in one application, options in another and configuration in some other applicaiton only adds confusion to the user. The user assumes that each do a different thing when actual fact, in each application they do the same thing.
4) Less religion and politics, more results. Users don’t care about the religious and political air that floats around the OSS community, they simply want an application that works.
Whether or not they can jump into the source code and tweak the application to make it load 0.001 secounds faster is inmaterial, ultimately, the user has a set of requirements and if your application doesn’t meet those requirements, no matter how much you scream freedom, “freedom” isn’t going to get things done.
You make very good points. The article seemed to me like the author tried to accomplish an involved task and found it too difficult. To get even with the OSS crowd for the failure, the author has written this philippic against the entire OSS world.
By the way, I think MySQL and PostgreSQL are quite easy to set up and administer (MySQL especially). I don’t use any of the GUIs available though, I think the command line interface is a lot easier to use (which surprises me a bit).
Nope, I think in fact you’re the one missing MY point.
I don’t think Iconoclast missed your point, but you are making a point that is entirely off base from what he/she is saying; and the same is true contrariwise.
Iconoclast, correct me if I’m wrong, is stating that the article is pointless because whining about a program not having this that or the other feature is a complete waste of time. I tend to agree with this whether you are talking about commercial software or OSS software.
Also, the editor comment/comparison was to illustrate that if GEdit, Notepad, or whatever, doesn’t suit your needs, but something else does, then for heaven’s sake stop whining and use the product that works for you. It seems you’re reading far more into the comment than was intended.
In the case of this particular article, it is whining about Linux, MySOL, etc. So, Iconoclast is basically saying if Windows suits you better, then instead of whining that Linux isn’t Windows, why don’t you just use Windows? (Why he/she couldn’t have just said that instead of using convoluted analogies is a head scratcher though – sorry, no offense intended).
Notepad’s feature set is AFAIK frozen, and has been for a few years.
And the current release of GEdit is frozen. Completeness doesn’t mean that a program will no longer be improved in future releases. It means that for a given release, all goals and plans have been satisfactorily accomplished. It’s the same with commercial applications.
If they are still contemplating adding new features then the application is NOT complete.
By that argument, then Windows is not complete. Neither is IIS, IE, Office, or most other MS applications (except for Notepad – which I wouldn’t tout as a major accomplishment.
)
It might have reached the same (Or better) level of stability, but in no way can you compare gedit’s level of completion to that of notepad simply because gedit’s feature set is expected to expand.
Again, I don’t think that GEdit’s completion was the point per-se, but rather that complaining about GEdit when there are a thousand other editor’s to use is kind of pointless.
But since you mentioned it, I disagree with you on this point. I have worked in commercial software for a long time and completeness is judged at the version level, not the overall future plans for a product. In fact, software should never be complete, by your definition, if you want to sustain a company.
In any case using gedit as an example simply highlights the author’s point. In what universe does a “lightweight text editor”, as the gedit developers describe it, need a plugin system?
But that doesn’t matter. The developer wants to put one in and that is the controlling force and the only one that matters.
If as an end user you don’t like it, the productive choices left to you are to a) use something else, b) write your own software. Of course, you can write vitriolic articles and piss and moan, but the developer will most likely just mock you and call you names (I know, I work with several of them) so that is a waste of time. With commercial software, if enough market stink is produced, the company will sometimes comply with market demands, but with small OSS projects, the market simply isn’t a driving force.
It’s symptomatic of current application development (In both proprietary and OSS spheres) that applications gather features they simply do not need.
I agree with you here, but I still think in commercial applications, those in control of the company decide and in OSS applications the developer decides. As end users, our role is simply to pick the app that best fits our needs. If one doesn’t, and we can’t write our own, we settle. That’s just how it is across the board, not only in the OSS space.
More features = More code = More bugs. The software world would be simpler if occasionally developers stepped back and asked themselves “Do we actually need that cool new feature?” instead of “Yeah, that’s cool, let’s put it in”.
That will never happen. With commercial applications that would be the death of your company. No new features == no new revenue. It’s as simple as that.
I don’t want you to think that I don’t agree with the points you’ve made; some of them I do agree with (in fact, I think I agree with everything except your thoughts on completeness). But, the truth remains that combative whining is not a productive way to improve your end user experience. On this note, I have to strongly agree with Iconoclast; it is a waste of time.
Defintly. The author has a point.
This also goes for apis and libraries. Beeing a developer, it’s just a plain pain to produce applications on linux.
New/update libraries appear every month, and they’re often incompatible, so you have to relearn the api quite often.
And you never know which library the user(his distro) might have..
I’ll give windows one thing, feature finished and stable ABI/APIs in the core libraries. Write a gui app, compile it on windows NT and it will run just as good on winxp and vice versa.
(no, gcc -static is not an option.. )
…or, should I say, hundreds of people got together to prepare this potluck feast for you, and it’s OK if you just pay at the door instead of CONTRIBUTING SOMETHING……
…so, one of you comes in and starts screaming bloody-murder that some of the recipes were STOLEN from you, after you yourself contributed half of them, and let your partners contribute the other half….
…and another one of you comes in and starts whining that you don’t like the sweet taste of this, the sour taste of that, and the salty taste of the other thing…..
…and another one of you comes in, from the local restaurant across the way, hollering about how BAD it is that people have the NERVE to do something like this and ruin your business….
…and another one, from the frozen-dinner factory at the other side of the country, comes in and starts putting up a comparison chart on the blackboard, showing how much cheaper this would all have been, had you all just gone to thestore and bought a frozen dinner…
…but me, I don’t complain. I just brought some crackers, anyway. They go nicely with the cheese Mike brought.
Oh…I’m sorry….was I not supposed to bring any crackers, because the box has your name one it? And is there a fine for that? Why?
I develop for Linux and Windows and I really don’t see that one is more difficult than the other. Linux may have changing libraries that are incompatible, but I’ve experienced that on Windows as recently as this month.
On the other hand, Windows API and all the available libraries, wrappers, etc. are, put simply, bloated, sometimes incompatible, and in dire need of a spring cleaning.
It’s easy to be backward compatible when you never get rid of your old garbage.
Go to the library and read:
The Last Battle
by C. S. Lewis.
Take special note of the scene near the end, where the dwarves are offered a beautiful feast.
It’s the best example of what’s really wrong with Linux, and it’s us, all right. But not in the way the author claims.
>Most people don’t want to “get inside the system and learn >how everything works.”
That’s the problem. If you plan to use a system for more than word processing, web surfing and email, you HAVE to “get inside the system.” This is reguardless of what OS you use, no matter how “polished and robust” it is.
>Why is that people such as yourself seem resistant (don’t >mean to be insulting, but there seems to be resistance to >idea of making opensource software more polished and >robust) to the idea of making application installs >polished/intuitive/easy?
I’m NOT against this. In fact I’m all for it. But when people say “I just want it to work” they all too often mean “I just want it to work like windows.” Which makes me ask, “then why are you using linux?”
I guess my comment was a bit argumentative, especially harping on about the gedit vs Notepad comparison so much.
Part of the problem here, IMO, is that complete probably means different things to different people as far as software goes. Someone intimately familiar with an application (Eg A developer) is going to find that the application is complete much sooner than someone who isn’t familiar with it (Eg Grandma :>).
From a personal point of view (Others probably won’t agree) an application is complete when the only thing left to do to it is minimal maintenance for rarely found bugs and yes, by that definition all those applications you mentioned are incomplete.
Perhaps there’s something stranger going on here. I think most programmers can testify to the seductive allure of the “idea” stage of a project. Namely that part at the start of a project where you’re generating ideas/dreams of what it will eventually be. A lot of projects get dropped right here (Think there was an article on OSNews sometime in the past that described something like this) when the amount of effort required to put those ideas into practice is realised.
I’m thinking maybe there’s another stage, one that appears after a project is completed. Namely the “comfortable” stage. At this point the project has mostly switched to maintenance mode as it approaches completion. Perhaps there’s just less incentive, once the “comfortable” stage has been reached, to start a new project and instead the preference is to keep busy by tweaking the finished one or adding new features.
In a commercial environment this might be because the company is unwilling to risk developing a new application, or try entering a new market. So they keep adding features to the old application and hope to sell it as being improved.
However I think in the OSS world it might be different. For the non-guru programmer I’m guessing that joining the development team of an application that has already reached the “comfortable” stage is a more attractive proposition than joining one that is just starting out. Less work to do, actual code to play with and an working application that can show you results are a few of the reasons. My guess is that few of those incoming programmers won’t want to insert a feature into the application, just to place their mark upon it (Kind like graffiti artists :>). So you get inevitable feature creep beyond the designs of the original team, forking only exacerbates this.
Of course extra features aren’t always a bad thing.
When people say they want it to work “just like Windows”, they’re not saying that they want to run Windows and they’re not saying that they wished Linux was Windows. What they are saying is that they want a simple installer that doesn’t require them to read a 100-page readme just to install the app. Windows apps, for the most part, install and work without tons of reading and downloading of 3rd-party libraries.