Yesterday, a co-worker of mine and I had a lengthy discussion about this article posted on OSNews awhile back. My past writing about Linux has centered on general usability and sensible defaults, but his contention was that Linux is the Linux kernel and that anything beyond that is the responsibility of the distribution. The conversation took an interesting turn. Read on for more.
Certainly, the word “Linux” refers to a kernel. I believe, much like the old confederate flag, insistance on what it really is means little when most people perceive it as something else. “Linux,” that is, what it means to most of the word, is a complete, rapidly developing, open source operating system. Each distribution provides a mostly unique spin on it, and generally, most people’s perception of Linux is actually the spice added by their selected distribution. Some people think Gnome is Linux, some KDE, still others XFCE or Windowmaker. As a result, unsurprisingly, the choices a distribution maintainer makes affect not only their users, but the perception of Linux as a whole.
If you ask me, I’ll tell you straight out: I’m all for removal of choice from Linux distributions. Choice is generally good, but too much choice, and worse, uninformed choice, is bad. It’s no different than politics, choosing a text editor in Linux without any knowledge of how they work is like voting without any research into how the candidates stand on the issues. Fine. So let’s talk about choice.
As I said in the past, I believe a distribution should choose sensible default options and applications and leave the rest out. A user who wants specific applications should either a) choose the ditribution which best suits him by including them or b) download and compile those applications himself. Yet, everywhere I go and everyone I speak with tells me otherwise – Linux is about choice and removing the choice is akin to personal insult. It’s a strong voice that argues with me: “Choice is good.” But it’s not consistent.
Why? Because there’s another debate going on, unrelated, but behind the same four walls. That debate is about standards. Microsoft’s Internet Explorer has some amazing capabilities. Sure, I’ll give you that ActiveX and proprietary javascript elements can be terrible and/or security holes, but this is my point: the same community that argues for choice seems to stand behind consolidation and standards. I mean, in the last few years, we’ve seen the rise of open file formats, XML (and XHTML/XSL/RSS/Atom), and most recently, the new release of the Linux Standards Base (LSB2.) Sure, standards make everything easier for application developers to provide a consistent and clean experience, but at the heart of it, aren’t they also a kind of “removal of choice?” If guidelines tell you where you must store user data, or how your code must be written, doesn’t that limit the philosophical choice to do whatever you want? The same way people tell me “If you don’t like emacs, don’t use it,” why don’t people say, “if you don’t like IE’s behavior, don’t code for it.”?
I don’t pretend to be a supporter of non-standardized applications, protocols, and organizations, but I do wonder why there is such heated debate and inflexibility. There are keywords — FUD, troll, zealot — words that people call each other that incite the deepest sense of insult and argument. There is allegiance and faith in hardware, software, programming languages, and interfaces seemingly comparable only to religion in starting all out “flame wars.” Where, I ask, does this come from?!
We accept choice when it suits our needs and reject it when it doesn’t.
Serveral years ago, Linus Torvalds shared his new kernel with an audience employing technologies that would ultimately serve as the building blocks of the internet as we know it today. His kernel, a fully open sourced collection of code, invited developers from around the world to write their own applications and share their code too, to attempt to create best of breed applications. That much is undisputable.
Somewhere along the way, there became a nearly religious belief, some sort of wacky zen centering around “choice” being the key to the success of Linux. The monks who preach this gospel are all over the net. But I’ve never seen it that way. I’ve always felt as though I don’t want to spend my time trying to decide which word processor to use. I don’t want to flip flop between Kontact and Evolution, Gaim and Kopete, Abiword, KWord, Open Office.org Writer, etc. Applications are merely an aside to productivity, the key is a comfortable, cohesive system. And that is acheived with the input and contributions of the masses.
Along the way, I think we’ve gotten confused. Some stand for choice when it’s convenient, like in Linux distributions, but not when it’s painful, like in adherence, or lack thereof, to standards. Some vice versa – they believe in a free for all, with a Darwinistic belief that the best will win out. We rarely permit our OS installations to limit our choice, yet insist on an immediate cease of non-standard behavior. Funny that Linux survives by having a single maintainer deciding what goes into the kernel and what doesn’t, the near antithesis of too many of its followers, who feel as though the removal of some quirky option is unacceptable, and the inclusion of a specific rarely used app is a necessity. How is it that we contradict ourselves so much as a group pushing towards the same goals?
Communally, I feel we should be aware the strengths of Linux. Linux is a superior system, but isn’t always treated like one. Some would say “People won’t use Linux because it’s too confusing,” and the retorts are heard from miles “It works for me.” The trick to Linux isn’t just to get hardware compatibility and drivers, but that’s part of it. It isn’t to have a pretty GUI, but that’s part of it, it isn’t to have a polished, consistent system, but that’s part of it.
No, it’s all of those things, plus one more, ultra important thing: it’s a flexible, adaptable system which can be whatever you need it to be. Linux needs to flex. But more than that, Linux needs to unflex. The people that CAN use Linux often are. It’s the people that can’t that we need to start paying attention to. Linux – in a general sense – needs a good starting point that doesn’t call upon you to make a zillion decisions. And that, I’m beginning to learn, comes from removal of choice.
Users cannot be overwhelmed by Linux or they won’t use it. Applications attempt to become compatible by relying on standards. When Linux doesn’t behave in an expected way, there are dependancy problems, failed compiles, kernel panics, and packages that don’t work across distributions. When a web page fails to adhere to standards, some browsers sometimes misrender the pages, making them only a part of the browsing experience. When an application doesn’t write to open formats, we end up with the disaster we have today – a zillion Microsoft Word documents that still don’t open and format properly in anything but Microsoft Office.
There is a line from the movie The American President that says “The symbol of your country cannot just be a flag. The symbol also has to be one of its citizens exercising his right to burn that flag in protest.” Choice is golden. But part of choice is the ability to select the unpopular path. Part of that choice is the ability to remove choice or not follow standards. And part of it, I believe, is the choice to quietly not use the software that doesn’t meet your individual needs. Unfortunately, it appears that choice complicates things for those who are learning, and adds to the experience of experts. The paradox of technology, it would seem, is to accept choice as both a good thing and a bad thing at once.
If you would like to see your thoughts or experiences with technology published, please consider writing an article for OSNews.
If you look at the trends in the Linux community you see there is a general trend in which explosions of choice follow inadequacies of choice. In short, when there are few good choices, the community decides to bombard itself with many choices, some of which may be good.
Well, what happens next is that the community tends to cull out the poor choices, and consolidate around good choices. In the end, most Linux distros and the community at large agree to follow sets of best-practices that consolidate on a converging set of choices.
So what does this mean? It means that at any moment, the community is busy defining the nature of its good choices and working to improve them. Look at desktops. KDE and GNOME (I’ll not mention the other DEs), while different in many regards, are converging in many fundamental ways. For the forseable future, they will use different core languages and widget sets, but their methods of interacting with the system and together are converging.
It is short sighted to assume that computer users have the same problems to solve. Computer users are as diverse as the human race and the elements within nature. And so are their problems.
The reason I use a computer is very different from the reason my mother uses one. How I use a computer is also quite different from how she, my mother, does. In a computing utopia, the tools my mother uses to solve her problems will vary drastically from the tools I use to tackle mine. Hence, their needs to be a diversity in tools to satisfy the multifaceted nature of our needs.
People advocating no choice, one distribution, one application and so on, are selfish, self-centered and extremely short sighted. More often than not, they are seeing things from their own perspective, rather than accepting the diversity of our problem space.
Linux remains one of the free computing tools that enables entities customize solutions for their problem spaces. Yet people inadvertently seek to destroy the very essence that makes Linux, free and open-source software powerful. Which is the fact that I am free to choose between KDE and GNOME knowing fully well that one size never fits all.
If it weren’t for diversity, the human race would have been extinct centuries ago. Lets learn to take a cue from nature. Asking whether or not choice is good or bad, is a like asking whether or not freedom is healthy. Time and again fields of human knowledge show that choice and diversity is essential for evolution to occur. If there is no choice, consequently, there will be no diversity, and therefore, little to no (r)evolutionary change.
There is a reason why Linux, free and open source software advance faster than their proprietary cousins. I leave that as an exercise for your intellectual capacities.
Hmm, there’s to much choice in women. Some people prefer blondes, some prefer brunettes. Some particularly dastardly fellows prefer red-heads and black-haired women.
Heck some fellows like fair-skinned women, some like tanned, some like dark, chocolate skinned women. Some women are colored somewhat more red or yellow.
And then there are eyes. Green. Hazel. Brown. Blue. Grey. Its all so confusing. Why couldn’t nature just decide on a color and stick with it.
And then some women are tall or short, skinny or full-bodied, small-breasted or busty, alto or soprano. Some are young, some are old, some are even older. Some like guys like me, some like guys not like me, some like other women.
Some like to talk, some like to listen. Some laugh at my jokes (very rare) and some think I’m dull.
Oh, why, oh why, oh why, do we have so much choice.
Why don’t we all agree that women should be blue-skinned with green hair and red eyes, medium-breasted, with long legs and always laugh at our jokes? It be best for everyone if we could just agree to that.
I’m not keen on medium-breasted.
I’m not keen on medium-breasted.
After-market augmentation or reduction is available for the tastes of all couples.
Its for the best if we simply standardize our tastes in women.
PS — for the ladies (boy, do I love geek women): everything applies as well to your tast in men.
binary longevity…. hmmmm never had a problem there sorry, now if he had said .so files, then that would be a different matter.
kde/gnome… the author is mistaken when he says you need to load all the APIs to support kde or gnome apps.. they load as they are needed.
My Documents and Program Files folders… NO NO NO, this is wrong and evil on so many levels….. and another thing, they are not even called folders, they are DIRECTORIES, get it right.
choice is very important, and if you dont like it, use something that gives you less of a choice, Xandros, Linspire etc… and never, ever, run apt-get on them hahaha
this article proves the old saying once again.
“Linux is not ready for HIS desktop ?
No HE is not ready for Linux!” (c)2004 raver31
if the author was to forget all his silly ideas of turning Linux into a Windows clone and trying Linux for the system it is instead, then it would have been a much better article.
I’d love to talk to you about your theory of “software evolution” in 10 years time. I’m pretty sure Longhorn*1.5 will still be ahead. Because that species will develop stronger, to extend that analogy.
The reason will still be the inability of the Linux “community” to recognise the importance of consistency and usability as raised in this article.
That’s one of the fundamental flaws of Linux (as in OS, not in pedantic “kernel”). You’ll never get its contributors/distros/third party open source apps to agree on some standards. (and that criticism’s much broader than free desktop dot org!)
They’ll collectively still be chasing the MS standards.
“binary longevity…. hmmmm never had a problem there sorry, now if he had said .so files, then that would be a different matter. ”
Yes…. And the computing public will love that distinction!
“My Documents and Program Files folders… NO NO NO, this is wrong and evil on so many levels….. and another thing, they are not even called folders, they are DIRECTORIES, get it right.”
Directories gave way to Folders in 1995. If you’re that out of date, maybe you should have second, third and fourth thoughts about commenting on 2004 issues. And exactly how are they more evil than /home/* and /usr/local/* or /opt/* folders?
so the developers, including distributions, need to choose who their audience is.
and the market chooses which distributions they prefer.
all good.
>I’d love to talk to you about your theory of “software
> evolution” in 10 years time. I’m pretty sure Longhorn*1.5
>will still be ahead. Because that species will develop >stronger, to extend that analogy.
A species that have yet to be breed.
>The reason will still be the inability of the Linux >”community” to recognise the importance of consistency and
>usability as raised in this article.
look at hcibib.org and let me know if something stands out to you.
>That’s one of the fundamental flaws of Linux (as in OS, not
>in pedantic “kernel”). You’ll never get its >contributors/distros/third party open source apps to agree
>on some standards. (and that criticism’s much broader than >free desktop dot org!)
standards. there are lots of open standards, which are you referring to kind sir?
>They’ll collectively still be chasing the MS standards.
that sentence is so profoundly wrong. MSFT does not define standards they partake in the standards process.
YHBTHAND
Most of you love having a choice, some of you think Linux should run for president. But your utmost disrespect for every other IT professionals choice in OS expertise is revolting. You disgust the IT world.
Your bandwagon antics don’t mean squat in every other persons and/or corporations world. People use the best tool for the job and/or whatever the hell they feel like. So what if someone else is using another OS. 5 years ago it used to be cool to run a less-than-mainstream OS with the latest toys configured.
Your rants about how cool it is b/c you ‘run’ linux on your pc are entirely stupid. Someone impresses us when we hear that your typing a post from a Linux PS2 or from a gamboy…that’s cool. Did you ever run apache from an as/400? No? Well let me tell you: It’s completely useless but it’s cool. Linux on PC isn’t cool anymore – it’s regular. Who doesn’t do it? You can’t make it through an accredited college without seeing the linux kernel in action.
My only question is: When will you ranters stop talking about your kernel like it’s the 2nd coming. Show me the great apps…that’s how you sell. You Linux trolls are no longer even promoting a hobby OS anymore. Your advertising for free for many different companies one of which is big bad IBM. They’ll find a way to make money from you. Stupid….
And NOTICE: I did not promote any OS. If it works for you fine.
Sorry, I didn’t read all 109 posts, I skimmed most of them….
I don’t think that ‘choice’ is the issue… whether you think ‘choice’ is good, bad, or misunderstood… OK, I’ll stop there.
The point I intended to make is this:
If there isn’t a solution you are completely satisfied with, you create your own solution. This is what happens in the non-commercial software world. I’ve long considered writing my own window manager, simply because after having tried just about every WM I can find, I have specific ideas of what I would like to have in a WM that aren’t all fully implemented in one specific WM. Some would argue there are already too many WMs, and nobody needs another one. Some would argue that my concepts are valid and another WM offering is needed.
“I like feature A from application X, and feature B from application Y. I wish I could have both features in one application.”
And that is why there is choice to begin with. One person or a group of people may have decided that they weren’t satisfied with current offerings, so they decided to create their own. No one forces anyone to use this new offering, but people will use it because there will undoubtedly be others who feel the same way. You may find a text editor that does everything you want, in which case it is not anyone’s fault that you felt you needed to try out 34 other editors before making a final decision except your own. For instance I prefer vim (vim -y actually) and SCiTE. There are many other editors I have never tried, but that doesn’t bother me because I’m completely satisfied with the two that I use. I like the idea that across the world somebody probaly relies on pico and gedit, whether they have ever used my editors of choice or not.
The “Standardization Path” leads to one point, in which nearly all code of a certain genre will perform the same functions in the same way, and looking similar while they do it. I feel this is a “bad thing”.
Of course, some things are needing standardization. I feel file format standards are good (providing they don’t limit programmers or functionality). I feel that filesystem heirarchy standards are good. I feel that standardized network protocols are good. I feel that user-interface standards are bad.
And as far as Linux (the kernel, folks) goes…
This always has been and is to this day Linus’s project, and it is our “choice” to use it. A large userbase doesn’t warrant changes and additions that are contradictory to what an artist wishes of his art.
Just my opinion, and sorry if this seemed to ramble a bit.
If choice is such an issue, then create an installer that gives the user the choice of standard installed packages (0ffice, web browser, IM’s, email, etc.) after DE and OS install. Double click an icon, get a list for each major section of programs normally included in a Linux Distribution; let the user decide which one’s they want.
Have the community come up with accepted descriptions of these standards, then implement them and let the user decide.
That’s what choice is all about.
Or if they don’t want to decide, at the installer give them the choice of an install with one Office package, one browser, one email reader, etc…
Make it so that people who don’t want to choose and people who want choices have an equal decision.
Standardization is necessary, but how do you do that? Linux users are happy the way things are, they can’t care less about next week, much less 5 years from now, you see them in this forum, most are a bunch of yahoos, trolling and missing your point, arguing non-sensically about gnome vs kde. I don’t know…i’m happy with Sarge and icewm, can’t see you wasting your time, trying to illuminate the unwashed masses.
A good idea or not, it was shot down without even being understood. I learned to never try.
>If choice is such an issue, then create an installer that >gives the user the choice of standard installed packages >(0ffice, web browser, IM’s, email, etc.) after DE and OS >install. Double click an icon, get a list for each major >section of programs normally included in a Linux >Distribution; let the user decide which one’s they want.
–snip–
>Or if they don’t want to decide, at the installer give them >the choice of an install with one Office package, one >browser, one email reader, etc…
of the distros I’ve tried, SuSE & Slackware both do this during the install process.. Slackware’s isn’t graphical an arguably requires more knowledge & experience… SuSE 8.2 & 9.2 (haven’t tried earlier) use YaST during the install & after installed, which is graphical (uses QT iirc) or nongraphical (curses iirc).
Some of the “user friendly” commercial distros are narrowing the choice of installable softtware already, and making sane defaults for the users that dont know what or dont want to choose.
This is perfectly ok.
On the other side, the author wishes to mandate choice reduction to _all_ remaining distros, the whole Free Software / Linux community. I dont see a point in his article because there already are not few distros that do exactly what he proposes, and have success with that, within their target audience.
How I understand the author, he seems to think that the other distributions, who exercise their Free Software right to make choice available, encumber the few commercial ones to represent the term “Linux” as a stremlined, polished commercial product which could compete the products Windows and OS X.
The author is sorry that those uncommercial Distros even can be connected to the, in his view, trademark “Linux”, and represent it in a anarchistic way, that in no way can compete with the more professional offerings by Microsoft and Apple.
What he would like to see is to forbid all them to use the well established name “Linux”, and to reserve the usage of this name to the Distros more competitive in the end user desktop market.
He talks about “removing choice” but means removing/hiding the confusion (in form of all the anticommercial community distros) from Joe User about what “Linux” is. You just cant sell Joe User the products Lycoris, Linspire or Xandros when there are Debian, Crux or Slack who also dare to call themselves a “Linux”, and I think that is what in fact the autor is against.
(in form of all the anticommercial community distros)
Non-commercial != anti-commercial.
Nice summary though.
1. Go to http://www.gnu.org and read about the GNU project and GNU/Linux operating system. And you will see that GNU/Linux is standardized. You simply can’t compare GNU/Linux distros with MS Windows. GNU/Linux distros are GNU/Linux variants which is possible and normal because GNU/Linux is an open source project. You don’t have Windows variants because MS Windows is a propriatary and closed source project. If MS Windows was open sourced you would see as many clones and variants as in GNU/Linux world.
2. Yes, GNU/Linux distributions don’t have a standard ‘dummy’ installation. So when you choose a ‘dummy’ installation you get what you deserved. Or better, ‘what you give is what you get’. See my nick: wygiwig. Most of GNU/Linux distributions give you an option during installation process to choose or not to choose particular application. So, if you don’t kick your ass and spend a couple of minutes on choosing what you really want you can end with two or three browsers, two or three IRC clients etc. in your default setup.
3. Yes, on application level GNU/Linux distros and MS Windows are like apples and bananas. What do you get when you buy, for example, MS Windows XP? A kernel, shell/gui layer, a vulnerable browser, a vulnerable email client, a crippled cd-burner, a cd and movie player, and some outdated card games. And you blame GNU/Linux distros because they treat you less dumb then you are. And they give you some choices, because they know that you might have been lazy (or ignorant) not to make these choices yourself (during install procedure).
4. MS Windows doesn’t give you any choices, and you proclaime it better than GNU/Linux distributions which give you more than one choice. And the lack of choices you call ‘a standard’.
5. Although Microsoft Windows is closed source and proprietary, and full variants are not possible on kernel level, there are many variants of Windows on gui level. So called shell replacements. Litestep, Geoshell, Talisman and many, many others. Actualy, there are more shell replacements for Windows than window managers for GNU/Linux. So where is your standard?
6. As far as I know GNU/Linux distributions defaults only on two most advanced desktop environments, either KDE or Gnome. And you suggest that one of them should be removed in the name of standards. No way. Both of them are standard, because they follow the standard paradigm how the modern GUI should behave and operate. Both of them have start menu, taskbar, desktop, icons, file managers, etc. Some minor things are different but not so different that a n user with a little experience wouldn’t be able to manage his/her tasks in both environments easily.
7. You mentioned in your article that your primary operating system is Windows XP. But mostly all applications you use on Windows come from GNU/Linux world (OpenOffice, Firefox/Mozilla, etc.). You are preaching about what standard GNU/Linux applications should appear in every GNU/Linux default setup, you blame them for a lack of standard setup, and at the same time you don’t use the standard Windows setup which comes with Internet Explorer, Outlook Express and other crap.
8. Let’s see. I have Windows XP on my machine and I want to
find an image manipulation program. Ooops, there isn’t one in the rogram menu. Ok, let’s buy or download a free one. But which one is standard? There are trillions of image manipulation programs when I google on them. Let’s try another one. IRC client. Ooops, there isn’t one. Again. I google. There are billions of IRC clients on the internet. But which one is standard? Is it Mirc, is it Kvirc, is it…? So many music players, so many text editors, so many browsers, so many email clients for windows… Which ones are standard? I am lost
Yours,
wygiwyg (what you give is what you get)
I remember reading a while back about how having too many choices in a grocery store can actually decrease sales. I can’t find that particular article, but here is a similar one.
http://www.gsb.stanford.edu/news/research/mktg_consumerchoice.shtml
Since we are talking about choice in general, I think it would be helpful to consider actual market research.
The choice of what software to use on my own machine affects one person: me. And that person is in complete control of that choice.
The choice whether or not to adhere to standards when publishing web sites affects anywhere from tens to hundreds of millions of people, depending on the popularity of the website. Depending on the nature of it, it may be a vitally important part of someone’s life. Yet that choice is in the control of only one person or organisation: the one who provides the web site.
This is why choice in one case is good, and “choice” in the other case is bad.
Was that simply enough put?
You said that the same people for choice supports monolithic standards. Yes we do, we support standards so that you can have a choice as to what program you use to access the information. Like instant messaging you can have any application you like and you know that it will work perfectly talking to someone else using a different one because the standard jabber protocol is in place. Take mp3’s for instance they are the de facto standard for digital music (at least for now) It is for all intents and purposes a standard file format. I can play my mp3’s with: amaroK, Juk, XMMS, Kaboodle, xine, gmplayer, iTunes, winamp, windows media player, my iPod, my mp3 cd players, and even my dvd player. That is just what i personally use in my own home. You see amaroK is not a drop in replacement for xmms they do the same thing at a very simple level “they play music” but the way in which they do this and the method in which they do this is completely different. If you have 40 mp3’s you are going to use xmms no problem it is lighter weight and you can say you like how it looks. If you have 14,000 mp3′ you might want to check out amaroK or Juk. I like the endless features with amaroK but can see how someone would prefer the cleaner simpler interface of juk. Open source is about those programs that do the same things working together even sharing code to reach their goals but to not be the same program offering the same thing.
>Standards are important. What would the world be like if we
>didn’t standardize lights on a car? What would the world be
>like if we didn’t decide that green means go and red means
>stop? What would our civilization be if we didn’t decide
>what’s a G and what’s a T? Standards are what organizes the
>world around you.
As usuall you are uncorrect and comparing apples to pears.
Light on a car are a function wich can be found on any car just like a mouse cursor but have you ever tried to put a Renault lamp in a Volvo light? It does not work…
Linux uses a lot of standards like TCP/IP/NFS/POSSIX/KERBEROS/SMTP/SNMP/POP etc.
Is companies like MS that are actually fighting standards, why do they not give the SMB protocol free so we can on build on it??
ps. Why are you involved in SkyOS translation if you think we need just one standard. Stick to Windows and be happy.
The first 50 posts was utterly confused, so I’ll assume the other 70 is that too.
It all breaks down to language. We just talk past eachother because ther are so many overloaded words in the Linux world.
Just take “Linux” it can mean just about anything from kernel source code to integrated and deployed system.
Choice… in general choice is good but there should be layers. Just like any good software system provides a general api for the lower layers. The problem seems to be that the market hasn’t found its layers yet.
Many distributions just collect a bunch of software alternatives and ship it with some tools to deply chosen software. This is a great product to buy IF you are a company trying to sell computer systems.
Problem arises when the distributors market their product as a consumer product. It is not.
An analogy in hardware would be if a distributor collected a bunch of chips and shipped that with their pcb and soldering tools. It’s just not aimed at end-users.
To remove choice, by law or any other means, is just foolish. To have a sensible chain of decision makeres between the raw material and the end-user would be good.
The article wasn’t about whether choice is good or not. It was an hypothesis on what Linux would need in order to kill MS Windows, from the authors perspective.
Indeed, Linux lacks an uniform look and some functionality, this is bad. But it lacks proper documentation. Sure, existing Linux user-base deals with it nicely. But what about new users? Many times man pages are contradictory, saying stuff is in /etc, or /usr/local/etc, or /opt/appname/etc … Linux already has most of the things the author suggests, but many large distros are yet to implement it correctly.
And some kind of package that makes installation easy would be good. This would mean a standard build environment, and a standard set of tools and libraries — perhaps in a FreeBSD fashion, so that they won’t conflict with your compiler/linker/make of choice. And then a standard package, which probably would include precompiled code and some kernel dependant source to be compiled. And make all distros compatible with such standard.
Distro installation should be made easier also. It involves too many hardware-oriented questions that are nontrivial. At least it’s getting better, pretty much reducing it to “do you have any SCSI stuff?” or so.
Notice that choice would not be lost, and nobody would get hurt by all that, au contraire: many people who has the itch for trying Linux would get to scratch it off, and existing users will survive and move on. Using any unix-based system will never be easy. But Linux should also be feasible for the not-so-geeks — the so-called “power users”, then you have more people, more money, and then industry won’t deny the importance of Linux and will support it.
Also notice that Linux will never really kill Windows. The later shall always exist, for the sake of choice, and for the users who don’t want to learn — or those who like it, learn it and choose it.
hanky wrote: “But we still have a lot of choice when compiling the kernel, between file systems, schedulers, firewalls etc. “
Exactly! Wasn’t the article about what has to be changed before Linux got generally adopted by the rest of the world (not us unix geeks, techs, sysadmins, developers etc.) Most people don’t even know what compiling is, they don’t want to know and thus will never learn. Heck, ask 100 people on the street what a OS kernel is and see how many answers correctly. Leno should ask that one on Jay Walk;)
Also some people argue “Don’t like it, don’t use it”. I think that’s what most people do and that’s also part of the authors point. Most ordinary people wouldn’t like Linux because of the problems mentioned in the article. And they wouldn’t use it. Perhaps they would use Mac OS X which shows the potential of UN*X as an OS for ordinary users.
I agree with the article. If Linux are going to make a big break on the desktop things are going to change. It doesn’t mean we can’t keep our own favorite distributions (Slackware!!!) but there has to be a standard Linux (or a couple of them:). It’s probably up to IBM, Red Hat, Novell to make that work. Then perhaps Linux can be an alternative for more people.
The writer makes many good points, but I think in some respects he confuses lack of choice with orderliness. It’s not necessary to remove choice to bring more order to the “Linux” experience. My preference is to provide maximum choice—an important hallmark of the open source movement, after all—while offering this choice within a context of order. Choice is not some fluff that confuses things. Chaos is the necessary ground of creativity. Evolution requires that experiments are possible. And new experiments add new choices. Once upon a time, Linux was an annoying choice (“Why can’t we just stay with UNIX? Or DOS?!”). In the meantime, pick whatever you like and stay with it. No one is pointing a gun to your head and asking you to switch among four word processors and six email clients, etc. There seems to be defacto favorites now in any event (e.g. OpenOffice.org, Mozilla, etc.).
Now STANDARDS: that’s what the Linux world really needs. Users should be able to try any and all packages they want. However, those packages should be designed for total interoperability within each logical domain (e.g. all word processor file formats should be open and readable by all other word processors—the obvious exception here is that, in the real world, each will also have to deal with Microsoft’s proprietary file formats—bleh!). That’s where standards are valuable. Another “standard” related to “choice within a context of order” is for there to be a “Linux” standard for installing and removing programs, one with the ease and transparency of, say, a Mac. Although there are pros and cons for the many current ways of doing this, it’s a tower of Babel that does nothing to advance the spread of FOSS. There should also be standard directory structures, a standard networking user interface, standard approaches to handling printers, etc. Maybe it’s still too early, and developers need the leeway to move toward a best of breed. But if that doesn’t happen soon, Linux’s big chance at popular acceptance may be squandered.
A word about GUIs. To each their own taste. I subscribe to no particular religion. But there needs to be a way for any program to run easily and smoothly under any GUI. I don’t know how that gets resolved, but basically it seems that either each developer of an application has to create version for each GUI; or one GUI “wins” and the developers have only one GUI to code for. Right now it’s a bit of a nightmare all around.
For all that we have problems with the Microsoft monopoly, their de-facto standardization is a major reason for the spectacular growth of the personal computer. Unfortunately, the “standards” imposed by Microsoft were of the wrong kind (i.e. designed for lock-in; they standardized on one model of automobile rather than standardizing on gasoline, usable by any auto). The FOSS community has an opportunity to create better standards, ones which benefit users rather than monopoly capitalists while allowing for the creativity that advances human civilization.
Indeed, Linux lacks an uniform look and some functionality, this is bad. But it lacks proper documentation. Sure, existing Linux user-base deals with it nicely. But what about new users? Many times man pages are contradictory, saying stuff is in /etc, or /usr/local/etc, or /opt/appname/etc … Linux already has most of the things the author suggests, but many large distros are yet to implement it correctly.
Do you know any project or platform which tries to target this problem (and only this problem)?
Most distros have sane default installations; there may be some redundancy with KDE-based distros also including Mozilla, but I’m sure people can handle that. Anyway, when we follow Adam’s proposal, and remove all redundancy – all choice – from a distro, what will we gain? Nothing at all, IMO. Eighter most distros will use the same choices, thus creating a monopoly of say KDE over Gnome, or every distro looks and works different from the other. You must be very zynical to say “Fedora users who want KDE should download and install it themself”. And there is something you cannot ignore: sometimes you need more than one program of a kind to fullfill all your tasks. For example, OOo writer is probably the most advanced text processor, in my experience the one which opens most MSWord documents correctly. On the other hand, Abiword is about the only one to open WordPerfect files. Since I’ve used WordPerfect for many years before switching to Linux, beeing able to open WP matters to me. So while every distro should make sure it’s default installation is sane, perfectly useable and well-documented, it should not pretend to perfectly know it’s users’ needs and cut all choices. It doesn’t hurt to have the other program on the CD, too!
Now vote me dictator for life, I mean, why should you have choice?
—-
Are you sure you aren’t the Matrix’s Architect?
The article is correct, but worded a bit poorly.
I am new to Linux, and I have tried MANY distros (about 20). My opinion of most of the most common distros are that they are all really about the same. They all have KDE *AND* GNOME *AND* xfce, several browsers, several editors, on and on.
This is too much. Which ones do I select? I tried to install and deselect some of the mail clients and other redundant options, but it broke the distro.
I don’t want 7 email clients, I dont want Links and Lynks and Mozilla and Firefox and Opera and Konquer. I’m not looking to “explore” my options by being drowned by them. Besides, most distros do a poor job at laying out a distro in a newbie fassion. No, Mandrake sucked, Suse is SO much worse.
I want to get my work done. Not try to figure out which program to use, where it is, and how to configure it.
If you don’t want to choose fine, get someone to choose for you. Just don’t mess with my own choices.
The Linux Kernel itself is better off the way it is.
Now, Desktop Linux, that could be a standard worth working for. Too many people gripe about what should do what in a desktop distro, etc.
But the Kernel? Nah, leave it as is.
Horrible to get these issues confused. A standard says: “there’s one way to do X because without a single way of doing X, the benefits of system Y would be useless or unavailable to most users and/or developers.” A lack of choice says: “There’s only one way to do X… just cause.”
So, yes, HTML, CSS, these are standards. Are they suboptimal? Maybe. But if there were 65 different markup languages/style sheet specifications out there, the web would be useless. So a standard was necessary necessary.
That’s why good standards tend to last a long time. Other ones tend to get phased out. For example, HTML is a standard, but XHTML (some might argue) is a better standard, which is phasing HTML out (in the long term).
ASCII was a standard for a long time. UTF-8/Unicode is now considered a better standard, and is phasing ASCII out.
There’s no paradox in saying, “I want to have the choice to use emacs, vim, or gedit, but I also want there to be only one encoding for text files so that I can send those files to my friends or cut and paste their content into other programs.” Again, if there were no text encoding standard, then computer systems as a whole would more or less break down as there would be no application interoperability.
So please be clear on these definitions! I don’t think UNIX developers want to “have it both ways.” I think they are being completely sane about this. We want choice in applications, but standards among them.
Ok having many choices leaves many people confused, and this could lead to less people going over to linux..
BUT, still I see Gentoo, a linux distribution that is all about the choice, and would make any one confused.. And its growing more and more popular!!
And I always hear that people starting using gentoo..
So choice made it more popular…
If you are old enough to use a text editor instead of a word-processor, then you are old enough to find the text editor that suits you…
Hitler or Stalin could have perfectly signed this article.
choice is life.
The author refers to “choice” in several distinct ways:
A) The power to use different, separate pieces of software to create a single system.
B) The power to determine how software will interact with other software.
I think the paradox is resolved if we do with this “choice/choice” debate what we’ve all done with the “free as in beer/as in speech” one.
If we think of standards as language, then the pieces of software we choose are our vocabulary. Some people use very specific, powerful, hard-to-pronounce words. Others use shorter, less-nuanced words. Many people just swear up a storm. But everyone uses different levels depending on the context; and so having the choice available is a good thing.
Adhering to standards is like adhering to a language: it’s an okay trade-off for the expressive power it gives us. Were everyone to make their own language, there would only be unintelligibility.
So yes, there is a slight paradox: we adhere to standards in order to preserve effective choice. Okay–duh. Just as we accept the principle of government in order to preserve an effective (and not just ideal and impractical) form of freedom.