Wow, I can't believe I'm posting before Beer! ;)

Great post, Asa. I really hope those who want Linux to remain an elitist o/s will take this to heart and put effort into sharing Linux with the masses.

I have to second Asa's beliefs.

I am simply glad you, Asa started this discussion. I really am. I am traditionally a MS Win user, with a lot of interest in Linux, so that I actually run it in several of my servers running critical applications. I am not using it in my regular day desktop, because I have few acquisition that runs only on Windows. Otherwise I'd probably (but not surely) jump to the full linux experience. As it is I am confortable with GNOME. People talk a lot about the deployment of Linux in desktops and way to improve it. It's not really Linux, but what the people really interact with that eventually needs work (GNOME, KDE...). I definitively would like the distributors to provide a simple accessible, starter edition version with everything that is needed, but nothing more. Make some choices, not bundled 3 word processors, 5 browsers, people get confused. As you said, Ubuntu is doing a fairly good job at it. However plenty still needs to be done. For example I scraped Ubuntu because in the current version the drivers for the centrino bundle are old, and they don't work. Now if you expect desktop deployment you also need to avoid such problems. I liked Ubuntu, but it was simply unusable. I know it's a difficult task, but on the other end, linux developers can really pretend average Joe to go and recompile the kernel with the latest dirvers.

This is my simple experience. In a computer world that seems taking an authoritarian direction (DRM, very restrictive licenses), I am really looking forward of an open, democratic OS for everybody to benefit from.

Nicola

"Linux has a window of opportunity right now -- before Mac moves over to Intel and can reduce its pricing, before Windows Vista is released,"

To suggest that Mac OSX and Windows Vista offer something so new that once they're released people will dump linux for them in favor of the UI is pretty lame.
X11 translucency and composition extensions have been around since before Microsoft even thought of adding those capabilities to Windows GDI.

"Let's get RP Linux out before people are wow'd and bedazzled by a much nicer windows UI"

Nope, I'm sorry, that's bunk. Luminocity, cairo, x11 compositing, their managed wrappers for mono, and other projects related to xdamage, xcomposition must be completed, and they must get funding.

############# START INTERLUDE ##############

There was recently a change in the libcairo API that broke gtkcairo, and it still hasn't been fixed. There needs to be some dedicated funds and developers going to fix that.

gdkcairo.c: In function 'gdkcairo_init':
gdkcairo.c:37: error: too few arguments to function 'cairo_create'

http://lists.freedesktop.org/archives/cairo-announce/2005-May/000014.html

"API changes
===========
Here are some of the API changes that have occurred:

~ cairo_create(void) -> cairo_create(cairo_surface_t *)

This is the big change that breaks every program. The ability
to re-target a cairo_t was not particularly useful, but it did
introduce a lot of muddy semantic questions. To eliminate
that, cairo_create now requires its target surface to be
passed in at creation time. This isn't too hard to cope with
as the typical first operation after cairo_create was often
cairo_set_target_foo. So the order of those two swap and the
application instead has cairo_foo_surface_create, then
cairo_create.

" and before those two operating systems and their primary hardware vendors start building the next level of draconian DRM into the platform -""

Half the libraries built on cairo no longer compile.

Would this happen over at Microsoft?
No, because they have money and developers to update the libraries that depend on the base libs.

I could easily fix gtkcairo, but my changes won't show up in the CVS people can download to develop with. I've fixed device drivers, libraries, and other elements crutial to users, and posted fixes on messageboards and NNTP.
I even patched ralink's 2500 driver and posted the fix.

############# END INTERLUDE ##############

These are the types of real world problems that we face, before we can bust the Gnome and X UI into the next level of competition with other systems.
These problems simply need to be fixed. The related projects need funding.

These are the things that stand in the way of Linux FreeX becoming as nice and as easy to develop for as Avalon, and Mac OSX.

The goal is not to hoodwink people into using linux before the nicer Vista and Mac OS's are released. The goal is to get the sh_t together and make Linux UI nicer than OSX and Vista. I think Novell is doing a great job at that with it's mono project and managed Cairo, and Xcomp stuff.

Put some MoFo developers on the case to fix some of the issues with XDamage, do something to help for crying out loud. Sun contributed tons of advancements to FreeX recently. They contributed most of the Open Office stuff.

################## NEXT ################

"Mac moves over to Intel and can reduce its pricing, before Windows Vista is released, and before those two operating systems and their primary hardware vendors start building the next level of draconian DRM into the platform"

http://developers.videolan.org/libdvdcss/

I can only believe there will ultimately be a libbluray-decode, and a libhddvd-decode as well. Only 1 key has to be leaked to free all media, and they wouldn't dare deactivate a decryption key because it would mean thousands of firmware players would immediately cease to function.

Beer, maybe you should give it a rest.

Nowhere did I say (or even suggest) that " Mac OSX and Windows Vista offer something so new that once they're released people will dump linux for them in favor of the UI"

Go re-read what I wrote. If you've got something relevant to what I said, which was that Linux _now_ has an opportunity to gain _more_ users before the next set of Mac and Microsoft releases -- and _not_ that current Linux users would dump it, then please do post that response.

If you're going to set up strawmen and knock them down, please find somewhere else to do it. Thanks.

- A

Great post Asa I totally agree, as for Beer claims about not blacklisting leaked keys for HD-DVD/Blu-ray they most certainly will do it and then issue new keys to the disc manufacturers so all new discs will not play on cracked players, the last thing in their mind is giving the consumer want they want and will lay the blame at those who leaked/crakced the key.

"Go re-read what I wrote. If you've got something relevant to what I said, which was that Linux _now_ has an opportunity to gain _more_ users before the next set of Mac and Microsoft releases -- and _not_ that current Linux users would dump it, then please do post that response."

I'll post it then. What's the point of getting people to use linux now before the next set of Windows and Mac releases? They are simply going to switch back anyway when those new OS's are released.

Windows users are going to switch back to IE7.

I don't think linux should aim for pleasing users of legacy software, and should rather look to the future and making a different and better system for it's own users. Nobody's going to use linux, if it seems like a "fall back" or 2nd hand system.

Beer, the point is that many of them won't go back. Linux is a compelling alternative with a lot of great features that many millions of users would appreciate. It's also current a great safe heaven from the malware and viruses that are plaguing Regular People on Windows. The trick is getting them to try it long enough for those sticky features to grab ahold of them. On top of that, getting some of them now (in decent numbers) helps to ensure that there will be opportunity to get more later.

- A

I think that's underestimating Microsoft a bit. They know about the problems on their platform, and this has been an especially long time between OS releases for them.

People have to want to use linux, *because* it's linux, otherwise it won't stick.
People use Mac, *because* it's mac. Because of Steve Jobs, and the whole turtle neck, coffee shop pseudo intellectual culture behind it.

People use firefox on windows because they think it will keep them safer. That's a whole other story. If and when microsoft addresses the security concerns and the ability to easily create browser extensions, there isn't much left.

If most people don't go back to IE once the release build of IE7 comes out, you can be the first to email me and tell me "I told you so".

I think linux should and has to be very different than windows, and should be very usable in it's own right without chasing down binary and or feature compatibility with other systems.

Beer, I've composed that email and it's in my Drafts folder. You'll be getting it shortly after IE 7 is released :-)

- A

I hope you're right.

Sorry for staying on this off-topic, but one concern for Firefox on Vista, once it's out, would be the UI. From the screenshots I've seen, Vista looks a lot different, with menu bars under the toolbar and other weird conventions. If Firefox fails to look native, I think a lot of users will get the same kind of feeling you get by running a Windows 95 program under Windows XP.

Functionality-wise, though, I'm pretty sure people will be too comfortable with Firefox to switch back to IE.

Anyway, Asa has a point in that now is a great opportunity to attract new Linux users. Regardless of how many would switch back to Windows/Mac again once they release their new beasts, the end result would be a greater Linux user base and as a result, better vendor/driver/software support. Linux would, simply put, become a major player in the game.

I, myself, am learning Linux simply because I don't want to move the the newer versions of the DRM-friendly OSs (Mac & Windows) when they come out. And there's no way I just gonna stay on XP forever.

Please disregard all the typos above. I'm tired.

Asa and Beer

A simple question ...
(1) will any version of linux ever be able to run the top 100 window programs,not interested in being told that "x" is a replacement for quickens or "x" is a replacement for M$ Office ... I want to keep using these programs ... I and the reqular folks and the companies that they work for have spent lots of $$$$$ on training and time on learning to use these programs ... regular folks don't want to reinvent the wheel again ....

(2) when will I be able to get 100% support from these 100 top programs support staff for linux? .... like I can now because I am using windows

Untill one and two happen, linux as a desktop is a "joke"

My2cents - 3-5 years after Mac 10.x runs on Intel .. Apple stops making computers and stays with the ipod and sells its Mac OS as a replacement for windows ... and by then linux will be a 1-3% user base ... and you all will still be discussing why linux failed ...

later, Richard

I am a Linux user. I have since more than a year no Windows OS on my desktop PC.
I didn't miss much, because some windows apps run with wine. But a few weeks ago, Google started "Google Earth". It doesn't work with Linux (nor Wine nor crossover plugin). But it seems to be fantastic. If I would switch back to Windows, then because Google Earth is Windows only and it's a killer application.

I am also a Firefox user - on both my PC and my laptop (Windows/Linux dual boot). Not because it's safer, but because it is more comfortable and because of all those nice little extensions which are easy installable or removable without touching the underlying operating system.
Should I really use MSIE 7 when I would go back to windows? Without adblock? without diggler? Always in danger to make my system unusable by installing something new? A file manager is not a browser. Neither if its name is Konqueror, nor if its name is Explorer.

Asa, I've followed this series with interest, but I'm a bit disappointed with the conclusion. What's needed is not more articles and presentations on what Linux gets wrong; believe it or not, you haven't presented a single new idea anywhere in this series. What's needed is for people to dig in and help out.

I can only speak for the GNOME desktop, because that's what I have experience with, but here are some things to consider:

* There's a GNOME HIG document detailing sound, consistent principles of interface design; developers need to be made aware of that and follow it.

* There's a task force called the GNOME Usability Project that scours the desktop for usability issues and works to resolve them; people need to be made aware of it and how to help it.

* Distributions like Ubuntu are trying to focus on "regular" users and put a huge emphasis on usability; people need to be made aware of that and how to help out.

As beer pointed out, the bottleneck is resources. Linux usability needs people who are willing to try the software, report problems and help fix them. As you know from your work with Mozilla, that's the hardest part of any open-source effort. And for some projects it's even harder, because they don't have the backing of an organization like the Mozilla Foundation.

So if you want to help Linux usability, you can; there are already like-minded people working on it and doing their best to publicize the effort. The progress that's been made in the last couple years has been amazing (oh, how I wish some of your readers had been using Linux four or five years ago, so they'd know just how far it's come), and if more people will pitch in and help things will improve even more quickly. That's really all that's needed at this point.

Richard,

check out http://www.codeweavers.com/ for your windows apps. It supports some versions of MS office, and Photoshop, and Dreamweaver and other such software.
It's basically a low cost commercial version of Wine, with an easy install.

It will also install IE6 out of box if you need to view pages in IE on linux, it's handy. While it doesn't support all windows software, it supports some major titles. I use it for adweb, which I use to submit my sites to search engines. There is no linux version yet, so I have to use this Win32 API interface layer.

If you need a windows gaming abstraction layer, you can check out http://www.transgaming.com/ for a product called cedega.

With the advent of managed code, .NET and mono, there will no longer be a need for such compatibility layers, and winforms applications should work seemlessly across linux and windows.

http://www.mono-project.com/WinForms

Novell is planning for full winforms support in the 2.0 version of their Mono product, and now since Microsoft is pushing developers to use managed .NET code, this should ease the pain of future application migration, as more and more developers use .NET MSIL code for their apps.

I'm not sure what the ramifications of people using avalon and Novell's legal ability to provide that in their own mono implementation will be, but overall, it looks like interoperability between windows and linux applications could be a few short years away.

Hey, I've been following your little series and agree with it all. Linux can be a system that everyone can use, however your first sentence in this post sums up the Linux community attitudes, and that needs to change.

They advocate Linux as the best kernel to use - yet whenever someone treads in the area they are alienated. If someone makes an idea for a good software program, but mentions "closed source", they get blasted.

If they want more companies to support Linux, then they'll have to embrace the closed source. If they want more users, they have to embrace the "Regular Users," the ones who probably don't spend every waking hour in from the computer.

The Linux-guru attitude irritates the hell out of me.

Richard: Why expect these programs to run on another operating system? If you buy the Windows version of Microsoft Office, can you install it on a Mac? Of course not, you have to buy the Mac version! Now it just so happens that, some years back, Microsoft decided to produce a version of Microsoft Office for the Macintosh, and they have yet to change their mind. But you still have to re-invest money in getting a different version of Microsoft Office for the Mac, and take into account the differences in the way the Windows and Macintosh versions work.

On the other hand, you can install, say, Adobe Reader or Opera for Linux, because Adobe and Opera have decided to produce Linux versions of those programs.

That's the key thing: the vendor has to decide it's worth releasing the software. It's a chicken-and-egg problem: if the market's small, few vendors will release products, thereby keeping the market small. That's not something that "Linux" can do on its own, though. It means business deals, it means lobbying. It means convincing (for example) Adobe that, yes, some people really will be willing to spend several hundred dollars to buy Photoshop for Linux even though Gimp comes free with the OS.

That said, many high-profile Windows apps will work via Codeweavers' Crossover Office (as Beer pointed out). Implementing full Windows compatibility is kind of like trying to hit a moving target in the dark, but it's something that's being refined continuously -- and Codeweavers has been a major contributor to the effort.

I would love to see a simplified user experience for Linux but I do not think you can cheat and take some kind of short cut through a morphed version of any of today’s distributions. Even if it where not a lost cause the results would be less than expected. I think that is what we are already seeing with some of the so-called “user-friendly” distributions out there.

Look at your own case history - Mozilla vs. Firefox. In order to make a true Regular People browser, the Firefox team had to start a whole new project with a totally different philosophy of application design that necessitated a complete rewrite of the user experience. If you want an Linux based OS for Regular People, you have to advocate the creation of a whole new distribution with a totally different philosophy of OS design that will necessitate a complete departure from the traditional Linux user experience. Firefox couldn’t be derived by reworking the Mozilla Suite. Likewise, a Regular People distribution cannot be derived by reworking today’s Linux distributions.

Those that use Linux today are a dedicate group. They like their system and it servers them better than anything else could. They will continue to use it and develop it despite its limited appeal to the masses. And I’ll hazard to guess that quite a few would never be motivated to use or help develop a Regular People distributions regardless of how much you urge them. They are just not into scratching other people’s itches. (see Seamonkey project)

In short, don’t ask people to change. Shout out and listen for like minding individuals who can make it happen. Help them to “just do it” with no discussions necessary!

I agree wholeheartedly with all the articles in series. My only concern is the slight suggestion that Firefox is a "job done" when, in my opinion, there is still much to be improved upon, extended and refined. I am sure this is an unfair criticsm, as most previous posts seem diligently focused on the path ahead for future releases, improving the feature set and so-forth. So I suppose my only real point is that I find a tone of humility appealing! :-)

Richard Martin

> A simple question ...
> (1) will any version of linux ever be able to run the top 100 window programs
> [...] I want to keep using these programs ... I and the reqular folks and the
> companies that they work for have spent lots of $$$$$ on training and time on
> learning to use these programs ... regular folks don't want to reinvent the
> wheel again ....

Never. (reinvent the wheel again? ;-))

> (2) when will I be able to get 100% support from these 100 top programs
> support staff for linux? .... like I can now because I am using windows

Never.

>Untill one and two happen, linux as a desktop is a "joke"

I think you are seriously misjudging what constitutes "desktop" and what applications "Regular People" use.

Desktop is not simply office machines. Home desktops also count. The whole dynamic behind businesses and what OS they use is different than that of home users and, as such, should be treated completely separately. Businesses are already switching to Macs in droves, mainly to get out from under the malware and virus yoke of the current Windows platform. There are also strong indications that they actively trying to avoid having a monoculture environment.

As I understand Asa's comments, home users are the ones he is predominantly talking about. Home users are the only "RegularPeople", since business users almost always have some sort of IT staff that help with any problems that arise.

So the question is, what applications do "Regular People" use on their home computer? The vast, vast majority don't use Quicken, I can tell you that. My bet would be that the VAST majority use the following apps *only* (in order):

- Browser (IE)
- MSN Messenger
- Word processor (Word)
- Outlook Express
- Excel

(+ Hearts, Mine Sweeper etc)

Most Regular People probably don't even use Outlook, they use Hotmail or another free web-based email provider.

Depending on who actually manages that Regular Person's computer (and I realize that I am somewhat contradicting my previous IT staff argument, but there's a big difference between dedicated IT staff and a friend/family member who's only around a couple of times a year), IE might have been exchanged for Firefox (which then doesn't gets updated, since this doesn't happen automatically and no *real* notification is given) and Outlook Express for Eudora or possibly Thunderbird. The really ambitious friend/family member *cough*geek*cough* will have installed OpenOffice instead of Word and Excel, although that is guaranteed to cause a problem that once or twice a year when the Regular Person decides to email a document to someone.

----------------------------------------------------------------
The problem, as it stands, is not that Linux is actually too difficult for Regular People, but the Linux is too difficult for those who manage Regular People's computers. These geeks may use Linux themselves, but they know that installing Linux on the Regular Person's computer will cause endless headaches and a massive increase in required support.
----------------------------------------------------------------

As such, I actually disagree with Asa's stance about moving settings and such over. That's not anywhere near top priority: I, to use an example, am perfectly capable of moving my mother's or sister's bookmarks and documents over. And, having already switched them to Firefox, Eudora (way before Thunderbird became useable) and OpenOffice, the lack of Microsoft products isn't the problem either (with one exception: My mother's tennis club refuses to talk with Firefox. But even solving this, even if it required working with the club to change their site, isn't a major problem in the scheme of moving her to Linux).

The, by far, largest hurdles are:
- Attaching hardware (digital camera, scanner, flash drive etc) doesn't cause that windows "what do you want to do, windows action handler thing" to launch. It's annoying to me and I generally try to kill it (to the extent that Windows allows, but since it's broken as h*ll...). Attaching a digital camera in Linux these days is a peace of cake. It just shows up in "My Computer" (Konq) in SuSE. Great for me, worthless for my mother or sister.
- No decent, simple photo-editing application.
- My med-student sister's anotomic atlas on DVD? Yeah, right.
- Video codec-of-the-week playback, incl. oldies like Quicktime.
- All those spyware riddles downloads, e.g. every single poker room I've tried, won't work.
- No background "idle" downloading of updates (esp. important for modem users).

If either my mother or sister actually used MSN (and my sister probably does), then add that to the list. Kopete? Gaim? Neither is more than a half-measure.

For my brother, the problem is GAMES.

And the problem is, these hurdles really are virtually INSURMOUNTABLE by an "ordinary geek". Most of these simply cannot be solved by a "normal geek", or even hoardes of them, and the rest would just be a major increase in the already annoying workload of supporting a friend's or family member's computer.

Beer: From the way I understand it, Microsoft typically don't HAVE to go around dealing with the mess of broken APIs, because once an API has been released and is utilised, it doesn't change. This is one area that Microsoft really seem to excel at -- keeping a stable set of well-defined APIs. If they need new functionality that can't be added without changing the API, then a new API call is created (eg. the various assortment of FunctionEx() or Function2() ) and the old one becomes depreciated -- BUT still continues to work.

While one could argue this pollutes the API space, the benefit in backwards compatibility is (potentially) much larger. Same goes for changes in data structures that require everything to be recompiled...

One thing that irks me about open source is the (seeming) willingness to break binary compatibility on things willy nilly... and in the example you mentioned, even source compatibility!

Antony,

I've read on redhat's site that even gnome plans to base it's gnome and gtk+ on cairo in future versions. I can't imagine what would happen if that came to be true and they just up and changed the init() or create() function.

I've actually been burned alot by changing API in GTKmm, and other linux libraries. That's one great thing about RPM, it lets you list the version of the dependancy.

Even the linux kernel's sys_ functions changed alot, so higher libraries is no big thing.

I used to program for windows, and one thing I miss is stable linkage. I think the functionEx() thing was better than simply changing stack call params. Say you have a func(int a, int b), and it changes to func(int a, int b, int c) Then you have the function popping 3 32 bit ints off the stack, when there are only 2, assuming it's not a fastcall.

That really, really sucks. That's why wrapping linux libraries in mono is such a positive measure. That way you can abstract the bad choices of the library maintainers away from the end developers. And again I think ximian, and now Novell is taking the right steps to solve some of these issues.

If I had to use Cairo right now with this type of breaking, I'd almost rather use it from monodevelop and C# wrapped.

I can't wait until monodevelop is farther along. That should make it easier for alot of people to develop for linux. Anjuta and Eclipse CDT are fine IDE's, except that the disparate libraries are still at the base, and there is the problem of non-steady API.

Asa, I agree with one one hundred percent.

I used RedHat for the first time last year, and was confused by the tons of options available... system settings, config, users... what!?! It just was a cool looking gaming machine to me to play before homeroom.

The thing is, Asa, Linux is open source.

Let me repeat that. OPEN SOURCE.

This means _anyone_ can do _anything_ and share it, right? (I know all the weird complications and such, but _basically_)

Well, *lets* do it. *Us*. The Mozilla Community. Mozilla OS- It's been rumored before. I know it'd be like adding on a ton of people to MoFo and all the complications there, but really- How cool would it be? Something *we* could make. I know you're busy with Firefox. But how awesome would it be to not only be the maker of a fast, clean, and simple browser, but an awesome OS that's small, easy to install, and expandable?

How awesome would it be?

Oh, and a name? How about OSTPBW (OS to prove 'Beer' wrong)... haha. Just kiddin'.

"Oh, and a name? How about OSTPBW (OS to prove 'Beer' wrong)... haha. Just kiddin'."

Joey, I think MoFo should do it! There's nothing wrong with having another linux distro.

Build a new WM/Fancy Shmancy Vector based Widget Set while you're at it wholy based on Cairo and vector graphics libraries, and have it interface Glitz/X. Give Windows Presentation Foundation (ex-GDI) a run for it's money! (don't skimp on the documentation either)

Come on, you can do it!

Hey, I actually I agree with Beer for once. Somebody needs to bring the Firefox effect to Linux, and who better but the Firefox team themselves?

I dont think that People will switch back to IE7 when its out. There are a lot of reasons to keep people staying with firefox. And even if IE7 can adress the GUI and security problems. There is still standards support that will be lacking.

And for the Linux problem. I dont think that more funding would solve all the problems. There are allready enough people who make great effort. There is exa, luminocity, xgl, e17, whatever. Everyone of those projects has something special. But still everyone of them lacks something. I would say none of these projects will succeed because everyone is doing his own thing. That just doesnt work. These projects need to join forces to make one perfect nextgen WM/GUI that is suitable for everyone.

Thats the same problem Asa adressed. There is way to much software out there and the user simply does not know which to use.
For me, it is the same problem which keeps me from using linux. There are simply way too many editors out there. But there is not a single one that is useable for me.

Asa and those who are actively advocating for open source software (exactly GNU/Linux)!

It seems to me that you are all missing the point. Linux has gained a great success recently because
1) IBM lost OS/2 and it didn't want to depend on Microsoft
2) Some other companies saw an opportunity in Linux (EMC, Alias, Crossover, ...)
3) ISPs saw a cheap (even free as a beer) alternative to MS server OSes that has these features: stability, security, interoperability
4) Some people want a safe and open OS which they can rely on. They can be sure it's adaware, viruses and other bullsh*t free
5) Some people want a constructor they can dive into
6) Some other minor reasons

Being a realist you should be aware that most Open Source projects are things no more than "just for fun". I mean if a particular developer loses an interest to his project that projects becomes unmaintained and therefore dies (if no other person takes over it). Also, Open Source developers mainly code for themselves. You can see hundred unresolved bugs in KDE, Gnome or Kernel bugzilla's. Why is it happening? Because those developers don't get paid to solve that bugs. Because they think those bugs are not important, the feature requests are useless in a long term. Usual PC users can't code. Thus problems remain.

Three years ago I have posted a feature request asking whether iframes can be turned off at all in Gecko based browsers. No solution yet. Though opera and IE have this feature for ages.

Three years ago I have posted a feature request asking whether "View image" context menu option could show image where it is, not reloading a whole page showing only this image. This [current Mozilla] behavior simply s*cks. No solution yet. We [Mozilla developers] treat this feature as useless for ourselves. Because we think so we don't give a f*ck.

In Firefox nightly "View Image" context menu option simply disappeared... I don't need it [Firefox developer thinks], who else might need it?

If a developer's life depends on how successive his product is, he/she will solve that bug or implement a feature. He will _listen_ to the users. But opensource developers code for fun.

This is what happening. Linux became very popular in the late 90s. At the same time five years later Linux still lacks a good Office Suite (don't tell me about that coffin named OpenOffice), Linux still lacks a good Raster Graphics Editor (Gimp is far inferior to Photoshop) and so on and so forth.

The only two bright things in Linux World are Linux (kernel itself) and Firefox. What other Open Source projects are well known in the Windows World? Ask any Windows user such a question.

// Artem

Artem:
Your points about Gimp and Openoffice.org are way off. Openoffice is a great replacement for Office for 90% or more of ordinary Office users.

Gimp is also excellent, but just recently became like that. The 2.x series of Gimp is highly usable and actually replaces Photoshop a lot in my office (did you check the pricing for Photoshop??)

--
Denis

No doubt OpenOffice is great application. But OpenOffice is a damn heavy application. You may argue that once being started you can keep it running, but Windows users tend to close their opened MS Word/Excel after they have finished working with _any_ document. Also OpenOffice doesn't share any libaries with other applications except libc and XFree ones. Why the f*ck did OpenOffice developers reinvent UI toolkit number 50? Why not using GTK+/wxWindows/QT? (QT is GPLed if your application is GPLed too).

It takes ages to start OpenOffice even if it was started before (cached).

As for Gimp (even 2.x) I dislike that:
1) All layers are limited in size (PS layers are infinite)
2) I can't assign effect to the layers (i.e. no simple drop shadow on my text)
3) Layout and UI are awful
4) Much fewer effects than PS (except script-foo, but I ain't a coder)

Year, you're right that PS costs much (around $600 if I ain't mistaken), but here in Russia absolutely most people don't pay for licensed software, we buy any software approximately for $5.

// Artem

Three years ago I have posted a feature request asking whether "View image" context menu option could show image where it is, not reloading a whole page showing only this image. This [current Mozilla] behavior simply s*cks. No solution yet. We [Mozilla developers] treat this feature as useless for ourselves. Because we think so we don't give a f*ck.

the ShowImage plugin

The only two bright things in Linux World are Linux (kernel itself) and Firefox. What other Open Source projects are well known in the Windows World? Ask any Windows user such a question.

Well, other than WinGIMP and OpenOffice.org, there's Inkscape, Blender3D, Scintilla and SciTE, Notepad++, and The Open CD, not to mention TBird.... among others...

(they just need a little publicity, that's all... and lotsa luv™)

IMHO the more I use these the less painful it gets for me to move fulltime to GNU/Linux.

Civil, people. Let's keep it civil. (okay so I've been guilty of being uncivilized before :p)

(gotta love Open-Source™...)

The debate is non-sensical. Linux does not need or want anything. It's the linux community consisting of users, developers, watchers, critics, etc that wants things. Likely these things are not mutually exclusive and probably conflicting.

Part of the user community is perfectly happy and doesn't desire change (or even sees change as a threat). This part is very small compared to the potential community growth that might come from your proposals being adopted. This is certainly what a large part of the community would like to see happen. In particular the corporately sponsored developers who nowadays form the backbone of nearly all linux related development would probably be very happy with such a thing (that's what they are paid for). Therefore it is logically wrong to say that linux doesn't need more users because a large part of the developer activity around linux is aimed at growing the amount of users. A majority of the community seems to desire growth, even though a minority doesn't care. Linux needs more users because its community desires it.

Now as to the qualitative aspects. I fully agree that current linux desktops offerings are pretty pathetic right now. It's consistently marketed as "hey, we know you want windows but we're 0$ so please give us a try". Free beer seems to be the strongest argument for linux on the desktop. In addition I see no agenda or roadmap that is going to change this in the short term. An industry wide effort to fix the issues you have pointed out is not present. Worse, many key players are still in the denial phase on many of these issues which means it will take a long time before they are fixed.

Like you I have a long history of trying out various linux distributions and playing with various oss software (I know what I'm talking about). Unlike you I've never been tempted to adopt linux as my primary OS during my decade long experience with the OS precisely for the reasons you point out in your nice "linux on the desktop" series. It's unusable, doesn't do what I need, it's complicated and constantly changing, updating & installing software is a nightmare.

I've written it before and I'll write it again, 2 IDE's/platforms are going to change linux.

http://www.realbasic.com/download/linux/

That one, and this one

http://www.go-mono.org
http://monodevelop.org
http://www.mono-project.com/WinForms

One lets incompetant newbies make full linux applications in less than 5 seconds, and distribute them as BLOATED and heavy single files. This is what made windows boom around win95/98. Newbie-ware.

The 2nd lets newer windows(winforms) software run seemlessly on linux as it does on windows. The 2.0 release features full winforms support and is slated to be released sometime next year.

http://www.mono-project.com/Mono_Project_Roadmap#Mono_2.0

They even have mentioned possible WinFX support. Though I don't know how microsoft will take that.

The first big hit was a couple months ago when RB Linux was released, which is now on it's 2nd beta, and the 2nd bomb will be when mono 2.0 with full winforms is out.

This will go a long way in making the dream of full, seemless interoperability a possibility. Kudos to Real Basic for making the linux version free. I also got an email from them at the beta release saying that when it's out of beta and released, the linux version will still be free.