Slashdot Mirror
What Linux Must Do To Survive...
mgoodrum writes "Emily Dresner-Thornber has posted an editorial/rant about Linux's viability as an end-user OS over at Netslaves. An interesting mix of criticism and her history as a Linux user." I think she's on the right track, but most of the places she says "Linux" I would substitute "one distribution". GNU/Linux need not be a monolithic entity to be adopted, there just has to be one user-proof distro available.
There are *tons* of word processor programs for Linux, including Word Perfect.
.doc format. None of them can handle this function (I use Staroffice clunkily anyway, but Word does work better). Here's why:
Yes there are. As a journalist, I need to write articles that conform to page limits (400 per page) and submit then is
* StarOffice won't let me do a word count on a selecton
* OpenOffice isn't stable, and didn't do sectional word counts anyway
* Abiword isn't finished
* Applix doesn't even HAVE a word count.
* KWord can't output stably to Office 97
* WordPerfect Office 2000 works great on Windows and is a pleasure to use. However, the WINE absed Linux version is inconsistently unbreably slow in its GUI and crashes rather often. I (meaning Cybersource) purchased a copy and we got burnt because we can't use it - it's that bad. A QT of GTK version would be wonderful and well worth the cash. A WINE version is not, at least at this current stage (perhaps a service pack might change my mind).
Yes, it is easy to mock those views. They are pretty much exatly against that which linux stands for. Choice. However, consider this:
1. Provides one window manager.
I see no reason a standards body can't decide on a standard default for the window manager. We aren't saying don't make the rest available. Just pick one for the default, regardless of distro, aside from specializied distro's that need to be different. Most non-tech users will stick with the default which would provide a commonality accross various people's installs.
2. Provide one shell.
See above answer. Hopefully the user wont' need the shell, but if they do, it can't hurt to have a standard default.
3. Provide a unified Linux "look and feel"
This one really WOULD be nice. Lack of consistancy in user interfaces drives me nuts in linux.
4. Remove options.
Fine. Make two different sets of config screen. Advanced and simple. Make simple the default. Make it so you can change the default to Advanaced if you so chose. There, done. Newbies and non-techies don't have to worry about it, but with a magical check box we can forever unlock the "Advanced" options. That doesn't sound unreasonable.
5. Stability stuffs...
Yeah, duh. Of course we want stability. But, duh, the command line needs to be a last resort if all else fails. Seems like she was stating the obvious there...
Justin Dubs
Bullshit. There is something close to 5,000 drivers that come with a default installation. When I installed Windows 2000 on my main box, absolutely everything was recognized. Correctly. And the sound works. And the printer works. Etc.
I don't know what she's on, but the default Mandrake install, which boots into KDE, looks remarkably similar to other *cough*Windows*cough* GUIs.
Wrong again. You argue that the woman doesn't understand the entire point of "Free Software", then you should you no absolutely nothing about Pay Software. KDE is in no way like Windows. The closest thing it has is the K menu, which is a Start Menu rip off. You can't position things on the K menu by clicking and dragging them like you can in Windows. A majority of the control panels aren't functioning fully yet with most cards (hello Sound) And you have to go to a command prompt to get more critical things done.
Contrast that with Windows, which gets everything right on the first try AND is easier to boot. I tried that bullshit with "teaching kids Unix at an earlier age so they would understand it" for a high school project. You know what? They couldn't understand a damn thing. But when they saw a Mac GUI they understood it immediately.
And I would suggest not calling a female journalist "babe", unless you want a wrath of feminists breathing down your neck.
- I don't care if they globalize against free speech. All my best free thoughts are done in my head.
Actually, I'd argue that even among the various *NIXes, Linux leaves a bit to be desired. My own experience: About a month ago I needed to set a personal webserver up for a friend. He needed to set up three sites with three domains on the same machine. Since he was short on money, we pieced together a P2-333 server, and I proceeded to load up RedHat for him. Total time to get the OS installed, configured, and talking to the network: 90 minutes. Then he decided that he actually wanted the machine to double as his development environment, so I had to get his sound, printer, and higher video modes working. Time? Another 40 minutes of patching, tweaking, manual file editing, and crawling through man pages. Then I reinstalled and configured Apache to run his sites. Web server and patch installs took another 60 minutes, start to stop. Finally, I installed Forte and his other dev tools on the machine, which required several more patches and manual file edits to get working...adding another 40 minutes of tweaking to the project. Total time? Ignoring the huge amount of time I spent searching for various documentation, it took nearly four hours to get the system running.
Of course, the system bombed the next day when he tried to update Java 1.2 to the 1.3 J2SE, so I got to repeat the whole thing again. And then it bombed again a week later when he tried to shut it down, and the whole damned filesystem corrupted (as far as I'm concerned, ext2 is just plain evil).
So, being a good friend, what did I do next? I grabbed my Solaris 8 Media Pack($70, unlimited license), drove back over to his home, and went to work. Solaris was installed, configured, and fully functional with his hardware and network in less than 30 minutes...and I NEVER ONCE had to edit a g*ddamned configuration file in vi to do it. iPlanet FasTrak Edition was installed and running all three of his sites 20 minutes after that, and his development tools were installed without a hitch or a patch. He figured out CDE in no time and is now happy as a clam.
The problem was not one of familiarity, I regularly use my Mandrake 6 box for development and built my own distro for my DNS server. I've used Linux since 1995 and am just about as familiar with it as one can get. But I'll be one of the fisrt to tell you that Linux has flaws. It's not that Linux isn't easy enough to use, it's that it often seems like some of the developers went out of their way to make it difficult. I like Linux, and I doubt that I'll be repartitioning my Mandrake box anytime soon, but I certainly wouldn't recommend it to anyone who isn't a serious chiphead. Although we've come a long way, the major distros still need to do some more work on the ease of use issues.
There is nothing so pathetic as seeing a beautiful young theory roughed up by a tough gang of facts.
Easy. Nothing.
Hold on, Captain Mods-Me-Down, I have a good point here.
Linux will survive no matter what. First, ask yourself: "What is Linux, anyway?" There are many ways you could answer that, but this time, I'll answer it this way: It's an operating system that's written by a dedicated cadre of highly skilled, super-intelligent, uber-geeks. They create it for themselves, because they need it for themselves.
Now ask yourself: How can you stop them? I don't think you actually can. Outmarket them? They don't care, the kernel-hackers keep on creating. Make strategic alliances, meta-conglomerate mega deals that lock linux out? The kernel-hackers keep on creating, still ceasing to care. In fact, short of taking away their computers, the uberhackers will continue to hack no matter what the rest of the world does.
The funny part is that in spite of their lack of caring about market success, Linux has become a huge market success. Now that I think about it, that might even be the REASON it's become successful. Ironic, isn't it.
In any case, I think that Linux-based companies have to be worried about survival, Linux will survive simply because there is a group of people who will never stop working on it.
" Another thing, and the last thing I'll talk about, is the fact that good human factors design (I'm talking about useability here, not happy dancing stuff, or eye candy) makes applications more useable for all humans, including hard-core coders. It shortens learning curves by allowing us to see and fiddle with just as many features as we want to at any time. We aren't all experts in every area of computing, or at least nobody that I know is, even those people whom I would consider computing gods. A little useability goes a long way there. Don't tell me you don't need that stuff unless you toggle in boot code on the front panel every time you boot!"
You seem to be confusing ease of use with ease of learning. Programmers care about productivity and are willing to put up with steep learning curves. Lusers care nothing about productivity and just want to learn the thing enough to do a couple things.
As an example consider VI or Emacs or ZOPE. All of these products are notoriously hard to learn but once you know them you fly!. You become so productive you feel like superman. Would any serious programmer be using word to write a program? I don't think so.
Sometimes ease of learning and ease of use conflict with other. Productivity often means not using the mouse and memorizing complex keyboard commands. Your average luser would never put up with that.
War is necrophilia.
To get to that point, you have to match *client* expectations. Er, that's Joe Sixpack again. Plug it in, it asks annoying questions once, you're done. Pay more money, answer a few more questions, and you get a spreadsheet.
Linux is not there yet.
I want more, you want more, but most of the world doesn't. They want porn and instant messages and letters to mom.
I agree with dear Emily. Removing options, at least in one distro, will do wonders for client adoption, and that's where you start attacking the cost of an operating system. Young's vision won't work until end users see Linux as a viable option. -j, a rabid FreeBSD user
I forget what 8 was for.
I guess the point I'm really trying to make is, why are we (as a community) trying so hard to compete with Microsoft's Evil Empire? It doesn't really matter if Linux or BSD or Apache or KDE or GNOME becomes mainstream. What is important is that the option is there, and people seem to be flocking towards what is being developed.
Another way of looking at it is this: Free Software does not have to mimic Microsoft and other commercial software in order for it to be worth the effort. As it stands, the Linux desktops that I use are plenty easy enough for average users, and if there's any sort of a demand for it to become easier, some company or group is going to suppy it.
The free software movement isn't about replacing commercial software, it's about providing alternatives. Merely by the organic way that the movement has been growing (more people get involved, more people are contributing, the products are getting better and easier and more refined etc.) An analogy for free software could be the markets: the best projects are going to attract more users, develop reputations and therefore attract more developers.
Trying to use a heavy handed coalition to guide the development of the free software movement won't work, and definetely won't be accepted. The "vi or emacs?" question is quickly being replaced by "gnome or kde?" with a handful of sideline contenders. No consortium could ever successfully select between GNOME and KDE... it sounds like too much like central planning, which is shortsighted and can never keep up with the diversity of an open market.
If GNOME and KDE continue to develop seperately from each other, and both refuse to adopt any sort of consistency between the two (ie. do we need another aRts to be developed for GNOME?) the end result will be that one of the two will join the sidelines, or perhaps the market will produce bindings or wrap arounds to compensate for the inconsistencies.
Where is free software heading? Only time will tell, and the "open ideas market" will be the judge.
But at the end of the day, it doesn't matter of Linux never replaces the Evil Empire. It'd be nice, but the option is already there for people, and the free software alternative keeps getting better everyday. Stuff that doesn't keep up will be left behind naturally, not by some clear decision. Standards will develop de facto from user acceptance.
Finally, it is perhaps exactly the diversity that the author points to as a flaw, that actually gives the free software movement it's vibrancy. It's great because it can meet so many needs, from regular computer users to geeks who post on slashdot.
Jeremy McNaughton
------ Live simply so that others may simply live.
The thing is, this non-unity is inherent to the design and sould of Linux. Linux can never be user friendly precisely because it is a hacker's machine. Who else but a hacker would be happy that almost all programs are distributed as source so that they can be compiled for any platform that Linux supports?
One of the things that this article touched on but didn't go into detail about is the installation procedure for most software. With Linux, if you're lucky, there's an RPM or a DEB somewhere out there (but those are frequently distribution specific, are usually written by someone other than the author of the software, and are frequently poorly done), but really it's practically impossible to properly maintain a Linux box without gcc or some other compiler. And what about Windows? Simple installer. Sometimes these installers aren't very good - but almost always the software will be installed and will run. With MacOS it's the same thing. There's never a library that you need to get. Even Linux applications that have been ported - jed, for example, has a Windows version. Do you need to install slang on Windows before jed? No, of course not. This is at once Linux's strength and it's fatal flaw.
I absolutely agree that Linux needs standard hardware specs. So far, Linux's general attitude has been "we'll run on anything!". But only sort of. And there's not always full support for a specific piece of hardware. And there are always bugs. I have yet to find an ethernet card whose MacOS drivers had "issues" where it couldn't do certain things, or crashed the machine under certain circumstances. Why? Because Linux is almost like an art form here - it's NEVER done.
So anyway, I absolutely agree with Emily's assessment of Linux. And though I love Linux to death (I'm a computer geek - and computers are my hobby), Linux is the Peter Pan of operating systems. It refuses to grow up - because with a little fairy dust and a happy thought, it can really fly.
Except that "the fastest growing" at a particular time is pretty meaningless when you're starting off with a small number. For example, Microsoft had an 89% desktop marketshare for the year before last. Even if they went to 99%, the rate of growth wouldn't sound that impressive because they started out with such a large chunk. Soon after MacOS X comes out, it will become the fastest growing OS because it's going from practically 0 to however many people get it — a big rate of change.
Maybe that "fastest growing" title gets you all tingly, but I'm much more impressed that Microsoft went from 89% to 92% desktop marketshare (where the real volume is at) while Linux is way down there at 1%.
Cheers,
with 128M of RAM and a 220 MHz processor ...Office is zippy on those specs? Please.
:) But anyway, I see no problem with her feedback and these are all valid criticisms in my opinion.
It is. X takes a lot of RAM, and while KDE and GNOME are growing slimmer by the second, they do too. Office 97 and Windows 95 will run on a P133 w/ 32 Mb of RAM. Not great, but okay. KDE (itself) will take 4 minutes between logon and desktop.
Huh? It's been around since '91 and has been "hot" for the last two years at least... hardly "flavour of the week".
Flavour of the week in an expression. Yes, Linux has only been popular for a relatively short time compared to Windows or netware or MacOS.
Upgrade this driver", "fiddle with this registry setting", etc
There's complexity in installing Windows apps, but the above comments are completely out of touch. NT and 2K Administrators touch the registry quite often. Regular users don't. Very few apps require driver upgrades.
However, I do agree in that a solid packaging system (which needs much more work on standardized package names, capabilities, granularity, naming conventions, etc) combined with a decent utility like apt-get (prolly on a RPM distribution - its the LSB and a more popular system) would provide an incredibly easy to use installation system.
But in the meantime its Gimp needs LibGimp needs GTK upgrade needs Bonobo upgrade needs GlibC upgrade. And that sucks. Still. there's only a couple on months before an easyto use apt based distribution (Mandrake 8) is released.
Duh. Welcome to Free Software, babe. That's the whole *point*.
I thought the point of free software was that it was more ethical to use Free Software than closed source software? In terms of Open Source, that's the exact opposite of Open Source - remember release early, release often? ESR has the same beliefs as this guy, and the B&B emphasises making stable, useable released as frequently as possible (and treating those who give feedback with respect as well, by the way). Besides, some people use Linux for the same reason they use MS Word - because for their task, its the best tool for the job.
I don't know what she's on, but the default Mandrake install, which boots into KDE, looks remarkably similar to other *cough*Windows*cough* GUIs.
Agreed. Mandrake would easily have to be the closest thing to getting Linux going on the desktop. But (for now) things like software installation are still headaches (lacking apt-get till version 8).
I'm salivating at the thought.
I'm truly glad somebody has taken the time to express these very same sentiments that I have had during the past week as I try to install linux for the first time.
It's about changing screen resolution. Mandrake has a control panel that lets you change every little detail of the window in look n'feel, but the panel has no mention of how to change the screen resolution. (I know about Xconfigurator now, but they were laughing at me in the slashdot irc forum)
It's about clicking a box to turn on the sound without having to install additional software / or visiting the command line as "super user". It's about not having to "compile" applications for your particular distro.
I know that in time I will become proficient at linux, but that's because I have the technical inclination to do so. Forget those people who just want to "get something done" (like my mom).
I agree that linux is terrific for it's endless configurability, but that is it's death knell.
McDonalds is not a success for it's "endless selection," but it's consistency of product. You can go to any McDonalds in the world and get the same exact thing. That is one of the primary driving factors to it's success.
The success of Linux is not due to its configurability and quality of the kernel, but instead a testament to the microsoft hegemony, and that a few people will take _any_ half baked (escuse me, beta) alternative.
Take the cheese to sickbay, the doctor should see it as soon as possible - B'Elanna Torres, "Learning Curve"