Linus on SCO, and the Desktop Being 10 Years Away

An anonymous reader writes "In this interview from last week's Linux.conf.au in Australia, Linus Torvalds talks about how the SCO lawsuit 'riled' him and led him to spend a week writing an application to archive his email, and how he think Linux will take 5 to 10 years to become mainstream on the desktop."

Comments

It's all about the desktop journey

[LibrePensador]

I don't know about you folks, but for me, when it comes to Desktop Linux, the journey really is much more rewarding and interesting than the destination.

I guess, to some degree that is because I started using Linux as my main desktop close to five years ago, but also because I am aware that profound social changes take time.

I think the key to the desktop is preloaded machines by big-vendor being available at retail stores. Only when the vendors have a stake in the success of Linux will they make sure that the peripherals state on the box that "it runs on Linux".

Re:It's all about the desktop journey

[Microlith]

And for some of us the journey, while interesting, is not nearly as good as the end result.

I'd love to have an easy to use system that I could handle without much difficulty while still having the power of Unix at hand should I want it.

This is not Linux.

Apple has it down pat, but that requires an investment in their hardware. Mandrake, Redhat, and SUSE have the install process down pat. The issue comes in just general responsiveness (behavior with hardware, plug & play, getting software installed/uninstalled.)

The question is when we will see something like this for the PC. Who will create the PC equivalent of MacOS X?

Different interpretations?

[Mazzie]

There seems to be a lot of different interpretations of Linus' views of the future of Linux floating around. There was a recent post on /. entitled "Linus says 2004 is the year of the Linux Desktop" or something like that. That seems to be a bit of a conflict with this article.

Can someone clarify his view for me? I don't follow Linux very closely, but am genuinely curious what Linus' real thoughts on the future of Linux for the desktop are.

Re:Different interpretations?

[rcpitt]

2004 is the year of the desktop as far as Linux people are concerned. IBM is reportedly pushing all their people to put Linux on their desktop by the end of the year and there are major governmental pushes all over the world to adopt Open Source (which in most places means Linux but in some means putting Open Office on Windoze).

The point is these are somewhat captive and specific-use oriented desktops, not those of the great unwashed public which account for upwards of 90% of the market. I don't know that it will take 10 years but it might - M$ won't sit back and allow the erosion of their virtual monopoly to take place without a fight and this will include everything from economic incentives for game producers (can't do that for hardware OEMs anymore can they? but the judge didn't say anything about software producers) to "lowering" their prices. Yes, I predict the "value" to the consumer may increase because M$ will bundle more and more into their "base" offerings as they have done in the past - to the point where on a new system the fact that you get "everything" included (OS, game applications, mildly hobbled office apps, etc.) for only a mere $300 over the cost of the hardware will turn people's heads. Problem is you wont' be able to purchase these things individually anymore (have you tried to purchase just Word lately?) so the real value won't be known - marketing M$ style 101.

There are still lots of apps that people "must" have that we (Linux) don't quite get right. That's a lot of inertia to overcome.

On the other hand, I see a ray of sunshine in M$ move to a new OS that is not backwards compatible with much of the software out there. Personally I think they're shooting themselves in the foot, and it remains to be seen just how incompatible they will be - but this coupled with some more work on binary compatibility stuff (WINE, etc.) will make moving to Linux that much easier for some.

It's going to be an interesting decade.

Linux becoming commercial?

[anarchima]

...it's not organised in the commercial conference kind of sense. But that just means it's a lot more relaxed, the people just talk about technology, they don't try to sell stuff. And these days in the US it's unheard of, you can't make money with this kind of conference, so I go to the Australian one and I go to one in Canada (Ottowa Linux Symposium). So even Linus admits that the Linux "project" is moving away from its earlier, non-commercial roots. I wonder what effects the increasing commercialisation of Linux will have, through businesses like Red Hat trying to make a profit and so on. Hopefully it won't be all bad, but I'm worried that Linux will just turn into another Microsoft (obviously with open source, but still)...

Linus commenting

[xant]

The interesting thing about his comments about desktop Linux are that he's making them at all. He used to have a position of "Linus is what it is, I don't care where it goes, it's just fun to watch." He's not doing that so much now that it appears to be actually getting the places people imagined it would go 5-10 years ago. To make a specific claim, even one as flexible as that, is out of character for him and shows that he's starting to become interested in seeing his work succeed commercially (other than in the areas he works on directly).

The Board is set, the pieces are moving...

[ttldkns]

Alot depends on how secure m$ "secure computing" model is. If they do what theyre bragging about and allow pages of memory to go unchecked even by the OS itself i think u have the beginnings of the recipe for a super virus.

The next version of windows and how they move to get it mainstream (new standards, no forward compatibility for older windows, whatever) will be a big factor in how the desktop 'game' plays out...

Linux is developing for desktop with Lindows OS , its M$ turn, we need to wait for their move.

Re:I agree

[Durin_Deathless]

Why are we all so focused on cloning something we all agree is awful? Almost everyone I know agrees that overall OS X is a better interface(of those that know both, those that know one don't count here). So why not clone the best instead of cloning the worst?

Seriously, the whole hiding the apps from the user thing ticks me off. I like the OS X solution better. You can have an optional start menu if you like, but make the apps as easy to add/remove as OS X and Be OS and NeXTstep. All GUI programs should be this way. None of this "Program Files" you're too stupid to look here, and don't mix the GUI apps into the same dir with the command line ones.

OK. I'm done. Do I need to don a fireproof suit?

about that email archive program...

[abde]

I hope it's open source. Maybe Linus will release it? I'm drowning under ten years of archives, spanning email clients from Eudora-Mac v1.0 to Thunderbird and almost everything in between. I'd love to have a cool program that could organizde my scatterred archives ...

Linus' point

[zr-rifle]

Linus is very coherent. He often says that the kernel isn't being developed as a competitor to Microsoft's own thing. That's why his typically relaxed, hackerish timetable is very extended, while most agree that _now_ is the time for the Linux desktop to emerge.

That's why Redhat, IBM, SuSEa re investing in companies like Ximian who focus on the desktop dark-side of Linux.

Longhorn won't be out till 2005 if I'm correct and many users are very insatisfied with Windows XP, from Sobig/Blaster outbreaks dragging down productivity levels to random annoyances like messenger popups and a full suite of internet blockers/virus stoppers/software firewalls needed to surf the web.

Users are keeping an eye open for alternatives, that's why Linux desktop development needs to become desirable, marketable, usable and thus a replacement for the Windows desktop.

Desktop

[Anonymous Coward]

Linux on the desktop will happen when businesses can switch all their machines to linux and not miss anything. When Jim-Bob and Betty-Sue are forced to use, and are taught how to use, linux at work they won't be afraid of it at home.

I think everybody understands the lack of an exchange type collaboration server hurts business adoption, but it's not the only thing keeping people from switching.

My business wants to go linux, but we can't. We use an ERP system called Macola. It makes heavy use of VBA and soon will support only MS SQL Server. There is nothing we can do short of writing our own manufacturing and accounting packages.

Before you point me at compiere, let me inform you that I've done research into that. I'm not a big fan of the lead developer. He's dragging his feet on database independence (when few people want real independence, they just want an open database supported) because he wants to get paid for it. Many people have brought forth suggestions and were willing to get started only to get no response from him. Development companies were willing to put people on it and they get no feedback as to the status of the project. So still the whole system is tied to oracle and there's no feedback at all as to when that might change. For the lead developer of an open source project he is VERY stingy with the information. Let's not ignore the fact that there is no current manufacturing module. There are, however, 3 separate development projects that aren't working with each other because of petty pride issues. The lead dev does nothing to stop the pettyness. So fuck compiere. I'll check up on it next year. I don't expect it to be usable then either at the pace it's moving. You have no idea how many people hit their forum gung-ho ready to start working only to leave again after getting no answers to their questions.

There is nothing else out there that is as close to production ready as compiere. There are erp systems that run on linux, but those are for the big boys. My company is very small, the cost of buying those erp systems would be more than the savings switching to linux would create.

Re:Five to Ten

[tjwhaynes]

We produce a lot of floating point intensive code that depends critically on the underlining OS calls, and while the code may run, it becomes quite a chore to justify to the customer (government) why the results may differ from earlier versions. This tendency for code to be brittle with compiler and OS upgrades is not something we observe under IRIX and SunOS, the two other platforms we support, and have supported for longer than LINUX.

Ahhh the joys of floating point. There are days when I wish that floating point was banned. Customers have a nasty tendency to assume that floating point means totally accurate. Very few really understand the limitations of floating point and comments along the lines of "what do you mean I can't store 20 significant figures in my database?", "I entered 1.10 and now it's 1.0999999", "I've been running my simulation through a billion iterations using a 'float' type and the answer is screwy" are not only common but rife.

That said, within the limitations of the floating point code I've written, I've not observed changes on Linux between versions. I do observe differences between the results on Linux, Solaris, HPUX, AIX and Windows in the least significant digit, but that doesn't suprise me.

I wonder therefore whether you are being burned by standard flags on the compiler with respect to mathematical optimisation. If you are suddenly using --fast-math that will definitely screw your results, as will any of the other flags turned on by that setting. Ditto check -m128bit-long-double -m96bit-long-double or similar settings that might alter your precision and throw new answers out.

To be quite honest, if you are seeing changes in behaviour and you have test cases which demonstrate these changes, you should inform the GCC team via the mailing list and try and determine what has happened. GCC vies to be compliant (often more compliant than other compilers) with IEEE and ANSI standards, and useful bug reports can go a long way to maintaining that compliance.

If you haven't logged such problems, well, nobody else knows that that problem exists.

Cheers,

Toby Haynes

Re:Nonsense

[ceejayoz]

Technically, KDE/Linux has been good enough for the desktop for 1 or 2 years already.

For someone already computer savvy, perhaps.

For your average non-techie, it's not. Hell, even I had issues with Mandrake 8.0 - and I'm doing PHP coding for a living at the moment.

The 3d-modelling niche is a very good example on how fast Linux can take over a market when the apps are there.

The 3d-modeling niche is a very good example of Linux running not on the desktop but as a processing cluster (in this case, rendering graphics).

Some things to consider

[NtroP]

I, personally, think 2005/2005 will be critical for Linux on the desktop for serveral reasons.

First, with Microsoft EOL'ing support and bugfixes this year for NT4 and 98/SE, I see many users and organizations casting about for alternatives. IIR, about 25% of the Internet-connected users are still using 98/NT. With XP being expensive and probably requiring new HW as well, they will be forced to consider Something New(tm). This may mean looking at OS X - since they need new hardware anyway. Or, more likely, they may consider "trying" Linux on their current equipment - especially if they have a friend, or know someone, who can install in for them for cheap or free.

Second, and this ties in with the first, public schools and many businesses are really starting to feel the financial crunch of constant HW/Software/License upgrade. Many public schools (like ours) cannot lease equipment due to board policies against "incumbering subsequent administrations" (or some such nonsense) meaning that new equipment is cash out of pocket and old equipment, which can no longer be used/supported, is surplussed at a total loss. Businesses, as well, face the fact that upgrading older equipment in order to run the new OS from the Beast, simply to be able to have 10 more unused features added to Word, is stupid and wastefull.

When you sit back and think about it, for most schools and businesses, 95% of computer use is for what? Email, Internet access, basic word processing/spreadsheets/"powerpoint" and maybe some IM or connectivity to a "mainframe" for financials, records, etc. which generally means some sort of TN5250/whatever emulation. ALL of this can be done with Linux as the desktop - with the added bonus(?) of increased productivity due to end users not being as able to install Webshots, Kazaa, Trojan-loaderPro, or VirusOfTheHour 6.0. This means work can be done.

But there is still a huge hurdle. Most companies and schools don't necessarily have the technical know-how or confidence to roll out Linux on the desktop. I think this is a pretty big hurdle, but not a showstopper. First, I see a lot more advertising from big players ("no one ever got fired for recommending IBM") on prime-time TV for Linux. Second, I see that Sam's Club is selling a $300.00 Linux box with Linux pre-installed and (in our store) an entire row of monitors demoing it sitting next to the XP boxes selling for hundred$ more. This is bringing Linux into the conciousness of the public (although as geeks we seem wonder how anyone could have missed it for so long :-)

Let me speak from personal experience for a second. Last week we had an engineer from a software vendor show up to install an expensive, high-end HW/SW solution. Unfortunately, it runs on windows only, so we had to buy several Win2k3 Servers and have their engineer set it up for us (lot's of custom tweaks, lots of $$$). I asked him if there were any plans for porting it to Linux, especially considering that he recommended checking with their company first before applying any MS patches to these bexes as some of them have broken their software in the past (eek!). He turned and looked at me and said that over 80% of the places he's been to have asked the same question. So they've begun porting. It should be available next year sometime. Score one for the good guys.

Along those same lines, I took him around to some of our installations to test the new system on our workstations. Wanting to start with the possibility of having the greatest success, I sook him to one of our "newer" labs. His first comment was "You're using Dell GX110's still? Those are, like 4 years old!". I didn't bother to tell him that, as Systems Administrator, I'm still waiting for my GX110. In fact, we still have IBM 340 workstations deployed. Those are 6 or 7 years old.

We are facing a huge budget crunch. Because of this, we are being forced to do a close eval of possible ways to cut costs and squeeze the most out of our current investments. Af

Re:I agree

[BigGerman]

My problem with Windows XP search is not the doggy thing - it is easy to turn off.
The problem is that search does not work - it is not intuitive, makes assumptions about file types, and most importantly it is DIFFERENT from Windows 2000 (MS best OS ever).
With Linux I am comfortable that all the tools work the way they are advertized.

People get that Microsoft is garbage now

[IncohereD]

I think with the lower end of Linus's statement (5 years), the use (and awareness) of Linux will become much more noticeable. I've noticed recently that the SCO lawsuit has made some waves in UK papers, where previously you'd be hard pushed to find a mention of Linux whenever a computer-related article is published (Microsoft, Microsoft, Microsoft!). Possibly something to do with the fact that the big name of IBM is involved, but surely this is a good thing - getting the Linux name actually recognised!

My roommate was working tech support in the summer, and when blaster hit he definitely started noticing angry people saying stuff like "Windows is bullshit!", who had probably never thought about it that way before (i.e. previously they just blamed computers in general, or themselves). People are starting to blame Microsoft for their failures. And that can only lead to them looking for another option.

Re:I agree

[Anonymous Coward]

I know people keep saying over and over again that MS use focus groups for their GUI. But there are so many examples of weird limitations and designs that illustrate this must be a big fat lie.

A few simple examples:

1. Environment settings dialog. You go try edit the "path" environment variable and then tell me that a focus group said this was nice.

2. Windows XP control panel. In particular the changes they did to dialogs like Services, with an "Advanced" and "Simple" tab in the bottom of the window. You honestly believe groups and surveys recommended that?

3. File sharing in Windows XP Home edition. First version of Windows where I had to give up on helping a guy sharing a simple directory (without passing along floppy disks between computers with some weird setup).

I seriously doubt Microsoft used usability tests for anything but a couple of versions of Windows back in the 90'ies. Its not something they do on all their GUIs, thats for sure.

Interface testing

[aashenfe]

Here is an idea for your local LUG.

Nothing to do on a weekend?

Head down to a mall and set up a user interface test. Call the mall first and ask if they will donate an area to the activity. Take machines down and set up tables.

Ask passers by to take a survey. Give them a task to complete. After they try it, have them fill out a survey about the experience. Collect the surveys on a website so open source developers can access the info.

Sound like a good idea?

Re:I agree

[spectre_240sx]

Hear Hear. One of the major problems I've had with most gui's is that they aren't configurable enough. I want options. I love a lot of things about OSX, but I hate not being able to maximize a window! Oh, and don't get me started about the home key not taking the cursor to the beginning of a line. Before anyone flames me telling me this is already possible to change, what about public computers? Can't change those. How about some sort of removable device one could carry that would save settings such as those? Unfortunately, that would mean setting up standards which we all know Microsoft will completely obliterate :P

It seems to me that there also hasn't been enough inovation so far. However, I think Apple is heading in the right direction. Expose ROCKS! I just wish it was easier to control. The F keys are a little too out of reach IMO.

P.S. I apologize for any incoherance in this post. I have so many things I want to say and just not the room to say them here.

Re:And the number is ....

[otis wildflower]

That doesn't take into account the appearance of the keyword multiple times on a line.. You literally need something more like...

wget -O - http://australianit.news.com.au/articles/0,7204,84 07881%5E15841%5E%5Enbv%5E,00.html 2>/dev/null| perl -e 'while() { s#literally#$i++#eg; print "$i\n"; }'

On Linux usability

[0x0d0a]

even baby joe can use it.

I don't know about that. However, it's been pretty clearly established that, five or six years ago, a tech hobbyist could use Linux as his sole desktop. He might have to use care in purchasing hardware, and he might have to deal with LaTeX instead of a word processor. He might have to re-request documents in a different format. He might spend an awfully long time getting things up and running. However, Linux was usable alone.

KDE and GNOME and other projects steadily got easier to use and were cleaned up. Windows compatibility improved. Companies slowly started to throw their weight behind Linux.

Two or three years ago, I'd say that a power user could reasonably start using Linux. There were still some annoying issues. Antialiasing wasn't in use, and many folks noticed this, if they were accustomed to Windows-style antialiasing. Sound drivers at the time were usually OSS/Free, so distributions used software sound servers to do sound mixing, which frequently resulted in poor-quality-resampled sound that broke up. XFree86 3.3 was still around, and 3D support in 3.3 was pretty bad. You still had to use the command line for a reasonable number of things (probably looking online for someone having the same problem), though folks were working hard on frontends.

Today, I think that a power user can comfortably run Linux, without any of the old drawbacks. 3d support is generally roughly on par with Windows. Audio is much better -- most distributions use ALSA and take advantage of hardware mixing, though more unusual features like hardware reverb generally aren't supported. Things like support for cheapo printers and reliable Windows filesharing support are in place. Most Windows productivity programs have an acceptably usable equivalent, and while document compatibility still isn't perfect (OpenOffice isn't identical with MS Office), it's good enough for most people to comfortably get work done without making an annoyance of themselves. Things are *not* equivalent to Windows. While most unusual hardware can be made to work one way or another (for example, I have a SmartHome USB X10 controller that can be made to work under 2.4 by compiling and installing modules myself...though 2.6 support is not in), it's still not flawless. The typical Linux distribution has gained weight -- GNOME and KDE are both quite heavyweight. Games are just not there -- this may not be an issue for the business desktop, but it's a huge deal for the home desktop. Binary software distribution (and no matter how nice it would be for everything to be open source, it just isn't going to happen) is a phenomenal pain in the ass, even in the presence of the LSB. I have Loki games, games that I purchased perhaps two years ago, that already do not run on current distributions. There is no existing technical solution, short of using Java bytecode and taking the performance hit that doing so entails.

I find that XP Home's multiuser workstation environment is much more accessable to a typical home user. Jane can log on, then she can switch to Bob, then he can log off and Jane can continue using her software. While I have run multiple X servers before on my box, I don't believe that there are any major distros that support such a setup nicely out of the box, and I remember running into all sorts of interesting bugs at the time -- run OpenGL software or something, and freezes started coming up.

Two of the major players in the Linux productivity world are OpenOffice and Mozilla, requred for MS Office and IE equivalence. Both of these use oddball widget sets. They are usable, and generally operate roughly like other applications on the system do. However, they are still disconcerting to the user. I *know* when something is using Athena or XUL or whatever OpenOffice uses, and I adapt my behavior accordingly. It's still confusing, unintuitive, and looks unprofessional to someone just trying to do work, however. By comparison, the Qt-Gtk differences are much mor

GNOME is Windows, but slower

[thockin]

I was at LCA, and saw a few interesting presentations on GNOME. Here's the revelation:

THEY'RE RE-CREATING WINDOWS.

No, really, they are. That's not necessarily bad, but it is a bit scary. Look:

GConf == Registry
Nautilus == Explorer shell
Bonobo == DCOM
GStreamer == Direct Show
DBus == (something they do now) ...

Much of the same duplications are being done for KDE, too. Re-inventing, re-inventing, re-inventing.

Furthermore, they're doing it worse. Or at least more slowly. Nautilus is SLOW. GNOME is much slower on equivalent hardware than Windows XP is.

I'm fine with re-implementing something that is the rigth answer. I'm not convinced all of these are, and I'm *know* we're not as fast or stable as XP in the GUI.

I want to see Linux and free/open software succeed. I really really do. I don't particularly LIKE OS/X, but it is a better experience than GNOME is, still.

I once more suggest that either the KDE team or the GNOME team concede to the other. Stop duplicating or triplicating efforts. We're still pretty far behind, and it doesn't seem to me that we're catching up (except on the simplest of desktop tasks).

Re:I agree

[ma_sivakumar]

Linus says that on the technical side, Linux is ready. Only the commercial space has to be created. That might happen sooner than what he predicts.

If you look outside the English speaking world Linux has a greater chance of reaching the desktops within next couple of years.

I am involved in a project to bring out a Tamil desktop for tamil speakers (Zhakanini. Only less than 5% of the population has access to computers now. One of the main reasons the majority do not use computer is the lack of tamil interface. Microsoft is not going to support Tamil interface anytime soon (for an unknown market demand) and the open source applications provide great support for localisation.

Combining these two factors this project aims to bring out a desktop for firts time computer users. they are not bothered about existing applications and we will be selling them pre-installed systems. Once we make the usage rate to say 20% of population with zhakanini, Linux desktop will be the default for tamil speakers (about 80 million).

I am sure there are many more communities like this in the world.