The Year of the Free Software Desktop

Uncle Entity

Thu Jul 17 12:59:05 -0700 2008

manage

Well, first of all there was the AT&T-BSD lawsuit that 'allowed' linux to gain a foothold in the first place.

Then you have the hardware compatibility issue, you have a better chance of finding a linux driver for any random bit than with the BSD—don't know about solaris though.

Finally, you have the anti-capitalists that object to someone actually being able to perform useful work on the backs of the volunteer developer community (Apple's Darwin port anyone). As I understand it the GPLv3 won't even let you use GPL code on a public facing server without releasing the source even though you aren't distributing anything except the result of the GPL code that is running. This leads, as far as I can tell, to less developer commitment to the truly 'free' BSDs as only the True Believers contribute to them.

Isn't every year labeled 'The Year of the FOSS Desktop' anyway?

The Year of the Free Software Desktop

N. Dan Smith

Thu Jul 17 13:15:42 -0700 2008

manage

As I understand it the GPLv3 won't even let you use GPL code on a public facing server without releasing the source even though you aren't distributing anything except the result of the GPL code that is running.

I believe you are referring to the Affero GPL, which is distinct from the standard GPL v. 3.

As for the drivers, there is no reason another kernel could not have as many drivers for desktop hardware as Linux.  In fact, as I intimated, OpenSolaris might even facilitate easier driver development for third parties (though that is debatable).  Still, I think the reason Linux has the most drivers is because it has the most momentum.  However, with Sun becoming a huge open source player, and with OpenSolaris possibly going GPL 3, we could see a shift in momentum.

The Year of the Free Software Desktop

Uncle Entity

Thu Jul 17 13:46:12 -0700 2008

manage

Maybe Solaris is the wave of the future but I really don't feel like learning a new system. Linux and freeBSD I can get around in just fine with openBSD being not that big of a leap from freeBSD that when I experimented with it on a few occasions I didn't have too much trouble.

Ubuntu 'Just Works' these days and if I ever get around to getting my cellphone to tether with it again (got a new phone a while back) I can de-louse my hard drive of Vista and maybe try out Solaris...though I'll probably put openBSD back on that partition since I'm determined to get a fully functional Gnome desktop working under it as I've never actually done anything except play around with it on an ebay surplus server I used to have up and running.

small point about labor

Thomas Lord

Thu Jul 17 14:48:57 -0700 2008

manage

You have the anti-capitalists that object to someone actually being able to perform useful work on the backs of the volunteer developer community

Prudence based on reasonable-person approaches to fairness with regard to labor is not "anti-capitalist". Quite the contrary. Reliance on volunteer labor, which must be perpetually recruited via misleading marketing -- that is anti-capitalist. In capitalism, your procurement tool is you wallet -- not your good looks.

-t

small point about labor

Uncle Entity

Thu Jul 17 17:29:55 -0700 2008

manage

Prudence based on reasonable-person approaches to fairness with regard to labor is not "anti-capitalist".

Well, I was thinking more along the lines of the 'tivo' clause or the web server code than the psychic profits people derive from contributing to a fundamental paradigm shift or whatever you want to call the FOSS Revolution. You want to make an omelet you have to break a few eggs, the trick is to be the chef methinks.

Not that I have any problem with the other new stuff like the anti-patent part or...honestly I don't have a problem with it at all, keep up the good work guys and gals.

It's the whole hiding behind copyright law while claiming to be free/libre that I take issue with.

small point about labor

Thomas Lord

Fri Jul 18 09:33:48 -0700 2008

manage

I was thinking more along the lines of the 'tivo' clause

That isn't "anti-capitalist" either: quite the opposite.

A consequence of software freedom is that it creates a competitive marketplace for modifications to programs. If I am given a copy of a program but forbidden from modifying myself or hiring someone of my choosing to modify, my software freedoms have been infringed. It's "modus tollendo tollens": if the software is free, the market exists. Therefore, if the market does not exist, the software is not free.

Digital restrictions, such as Tivo, are anti-capitalist. They exist precisely to prevent a market from arising. Since that in turn implies that the software so distributed is not distributed in freedom, it's appropriate for GPL to declare explicitly that people are not permitted to distribute the code in such ways.

or the web server code

I assume that you mean the extra provisions of the Affero variation of GPL.

I don't think I can defend them. The reasoning behind them has always struck me as a bit shaky. The FSF license FAQ offers:

It is essential for people to have the freedom to make modifications and use them privately, without ever publishing those modifications. However, putting the program on a server machine for the public to talk to is hardly “private†use, so it would be legitimate to require release of the source code in that special case. Developers who wish to address this might want to use the GNU Affero GPL for programs designed for network server use.

I don't quite buy that because I can't find a coherent interpretation of this distinction between public and private use or a clear definition of what it means to "talk" to a server. Absent the Affero clause, the GPL draws a bright line: if any right granted by the GPL is transferred to another person then, in effect, all rights must be transferred. If "talking to a server" constitutes a transfer of GPL rights (and thus requires source availability) then shouldn't the public also receive the right to modify the code running on that same server? Without that bright line, I have trouble seeing a principled reason for the Affero clause and I tend to think it was a mistake.

And I would agree that Affero is (at least arguably) "anti-capitalist" because it diminishes competition as follows:

Suppose that I run a successful web service, writing and owning the server code myself. For whatever reason, I put in links to source code under Affero. The resulting competitive situation is:

Initially, I fully own the code and I do not need a license to use it. Indeed, at any time, I am free to make modifications to my server but not include those in the source distributed (though I should not keep the fact I've done so secret, for that would be fraud). My competitors who download and start using my code enjoy no such freedom. They must share all of their modifications. The playing field is only ever again level if I decide to merge in 3rd party modifications to my own copy, under AGPL, after which I too must share all of my local modifications.

That isn't simply the "dual licensing" advantage that sole-owner copyright holders inevitably have under copyright law: it's an asymmetric grant of rights explicitly created by AGPL.

The same logic that supports the Tivo clause draws AGPL into question: AGPL removes a degree of competition from the market for modifications, therefore, it is not a free software license. (Obviously the FSF disagrees.)

There is a better solution, I think, to the problem AGPL tries to begin to address -- but it is a technical solution, not a licensing solution. We don't actually need nearly as many "centralized services" as we have. For many, and perhaps all needs, we could get by giving users general purpose accounts on virtualized servers, with users installing their own software there. For example, instead of an account on a blog program, a user would install their own copy of a "blog node" program, that programming working in P2P fashion with similar nodes elsewhere (more like Usenet, less like Wordpress). The AGPL isn't needed in that case: ordinary GPL will do. It's also a better architecture for user privacy, etc.

-t

small point about labor

Uncle Entity

Fri Jul 18 18:32:53 -0700 2008

manage

A consequence of software freedom is that it creates a competitive marketplace for modifications to programs. If I am given a copy of a program but forbidden from modifying myself or hiring someone of my choosing to modify, my software freedoms have been infringed. It's "modus tollendo tollens": if the software is free, the market exists. Therefore, if the market does not exist, the software is not free.

I don't know.

I don't think the market's existence is somehow linked to the freeness of the software. The market existed back in the day when RMS started on this journey but the software wasn't free before he and others created it out of the Void.

And, as I pointed out in another post, the fact that Tivo doesn't release the full software stack to make a functional replacement OS doesn't in any way change the freeness of any of the individual software packages. Even the much talked about Linus has a similar view on the 'kernel module theory' that the new GPL addresses in the tivo clause. OK, maybe I didn't specifically mention that in a different post but at least implied it.

Digital restrictions, such as Tivo, are anti-capitalist. They exist precisely to prevent a market from arising.

My understanding was they didn't release enough information to build a fully functional OS and that was the issue and not some DRM. I haven't really kept up on the issue though.

Nothing is stopping someone like Apple or a bunch of hackers getting together and making a competing product. I won't provide links because that would just be silly.

I think the root of this issue is that people are upset that tivo uses GPL code *and* tries to keep control over the hardware as a business model. If people were able to bootstrap mythTV onto a specialized piece of hardware like a tivo then that would cut into their profits and no one likes that. Give away the razor and make it up on the blades.

You don't hear (much) complaint about Apple locking people into their hardware with their implementation of an open source OS. They were smart enough to start out with a truly free base and throw the devs a few bones here and there to keep them happy. I can't compile a cvs version of darwin and add the apple bits to get an OSX super-beta and I'm perfectly fine with that...I just don't buy their computers. Well, other than my ipod that is.

But, anyway, I'm just kind of rambling. All copyright is anti-capitalist in my book if you use the laissez-faire definition.

slight OT

duhavid

Sat Jul 19 20:59:08 -0700 2008

manage

Just saw this on Groklaw.  Just in case you are interested.

http://www.groklaw.net/article.php?story=20080719184549151

The Year of the Free Software Desktop

Bruce Perens

Thu Jul 17 15:11:46 -0700 2008

manage

BSD had other problems besides the ATT lawsuit. Early versions wouldn't use the Seagate disk interface that most PCs used, and Linux would. But I think that BSD licensing was what then, and now, continues to kill BSD. It just doesn't make sense for the developer.

The Year of the Free Software Desktop

Thomas Lord

Thu Jul 17 16:44:48 -0700 2008

manage

I think that BSD licensing was what then, and now, continues to kill BSD. It just doesn't make sense for the developer.

I happen to agree but would you mind elaborating on how you got there? ("I'll show you mine if.... You first." :-)

-t

The Year of the Free Software Desktop

Bruce Perens

Thu Jul 17 16:48:21 -0700 2008

manage

BSD licensing is like being some big company's unpaid employee. You give them all rights, they make it proprietary, they give you nothing back.

Now, there are some cases where you might want to do this. For example, if someone else pays you to make your code BSD licensed, you've been compensated. Or if you are promoting a standard interface and want everyone to use the example code for it. But in general, it makes sense to license so that the Open folks can do what they want, and the Closed folks have to pay for the privilege.

You will notice that technocrat.net code is under Affero GPL.

The Year of the Free Software Desktop

Uncle Entity

Thu Jul 17 17:50:06 -0700 2008

manage

Just to expand on this a bit, you were behind busybox if I'm not mistaken.

Some evil router company comes along and changes it around a bit and doesn't release the source, how does that effect the work you put into it in the past?

All the labor you put into it that you donated to the community is still in the official release and they are in no way going to take away the reputation you have built up over the years because of your work on busybox among other projects. I'm really not seeing the justification for using the force of law to compel others to also donate the proceeds of their labor against their will.

No victim, no crime and all that.

But I in no way hide the fact that I'm against copyright and other forms of IP as a general rule.

The Year of the Free Software Desktop

Bruce Perens

Thu Jul 17 21:47:24 -0700 2008

manage

I'm really not seeing the justification for using the force of law to compel others to also donate the proceeds of their labor against their will.

Think of it as payment. As an Open Source developer, I have declared that I will accept payment for my work in another form than money. I will accept the contribution of additional software to the world as as Open Source.

Anyone who doesn't want to do that is welcome to pay me money.

The Year of the Free Software Desktop

Thomas Lord

Thu Jul 17 17:51:25 -0700 2008

manage

Well, in keeping with my incitement, I guess I should quote Mel Brooks' Blazing saddle: "Excuse me while I whip this out..."

You give them all rights, they make it proprietary, they give you nothing back.

Oh, you mean like Canonical did with GPL'ed Arch to me? ("Launchpad")

GPL is no obstacle to similar shenanigans to what you describe although, indeed, I agree strongly with you that GPL is less vulnerable and that is why it is more popular in kernel space.

Our difference is this:

I think it was the GNU/Linux vendors, not the programmers, who wanted GPL for the kernel. They received the greater benefit. It theoretically excluded MSFT. It absolutely excluded (then) Sun (their main target).

The BSD lawsuit ruled it out for commercialization as free software, except for the brave-little-to-lose crowd, for about 6-12 months just around the time when the Linux kernel was getting off the ground just about as convincingly as the Spruce Goose. Obviously the Linux kernel has matured nicely in many ways but, at that time, it was a question of just a few months difference and uncertainty during those few months -- so much of it was undisputed even early on -- it was avoided by the big money investors because it didn't exclude, just as you say.

And, at that time, BSD was way, way, way out in head by technical metrics (still is, by many measures). But for the legal issues for those few months, it was a no brainer choice and, if licensing could be ignored: Linux was a clear waste of time. (Well, almost -- see below about x86.)

Yet, licensing mattered to investors and, although GPL doesn't exclude anyone, it does exclude proprietary derivs -- hence the big push. (I should qualify this: Linux also drew attention for being originally a very x86 unix when the othes weren't quite ready for that -- that helped make the case that it was a "Sun Killer". But, again, that was only during the same few months of difference after which the x86 BSD ports disambiguated. It was Augustin that first drew the attention of the VC community to the potential of an x86 unix to chomp away at Sun's market. They pounced.

That VC pounce created a marketing push, big time, and propelled Linus into a much larger spotlight, ultimately leading him to a pretty cushy career (with all do respect -- I speak not only from critique but also jealousy). The VC pounce led to RHAT's strange "let's make stock-option gifts to some of that community" play. It led to economic "momentum" and a largely BS promise of a career path for volunteerism. Make examples and handsomely reward a few volunteers and many volunteers will step up and try to emulate the earlier ones even though it takes years to catch on that, no, those early "rock stars" are exceptions -- not really how things work.

And then it's all network effect with stuff on the margins. The cheerleading squad (aka Andover.com) appears. Hegemony gets imposed. And, yeah, a lot of newbies (a full generation now) get drawn into volunteering for Linux in, a way they don't fully appreciate, mainly in service of the vendors. It's "the latest style".

The simple fact of the matter is that the majority of GNU/Linux contributors don't get compensated, licensing or no, and don't have enough awareness to care whether or not some freedom-robbing use of their code is made. It's just herding. It's driven by vendor interest in the exclusions afforded by GPL, not so much by hacker interest. It's driven by marketing. It's a damn shame, is what it is.

Freedom fighters don't want "Closed folks [paying] for the privilege." They want labor justice among those vendors who use free software licensing.

Meanwhile, until then, it's a scheme. It's a rip-off for hackers. It's fraud.

-t

The Year of the Free Software Desktop

Bruce Perens

Thu Jul 17 18:23:53 -0700 2008

manage

Well, I don't know what happened between you and Canonical, and I'm not sure you want to discuss it here, but I thought you were being paid.

And you've done this same rant about Open Source being some scheme to rope in developers for big companies a large number of times now, and pretty much everyone else seems to not be buying it, me included. And I wish you'd quit, because it's already very tedious.

Yes, if I were Red Hat I'd want to use GPL for the work done by my own employees. I use AGPL for the code I'm making now.

The Year of the Free Software Desktop

Thomas Lord

Fri Jul 18 10:20:25 -0700 2008

manage

Well, I don't know what happened between you and Canonical, and I'm not sure you want to discuss it here, but I thought you were being paid.

Various generous individuals, to whom I am eternally grateful for help when help was badly needed, gave a mix of small gifts and some wages via the FSF. These were life-saving but amounted to less than minimum wage. Burning savings, taking on debt, and my (now) wife's wages kept us alive, got us to a state where we could escape an expensive lease, etc. Unless it was done pseudonymously, Canonical did not offer help. Towards the end of that, another generous friend to whom I'm grateful gave me some larger assistance / consulting work -- that was around the time I stopped busking -- but that was not "for Arch" per se. Since that time I've had two consulting gigs in biotech and a non-programming stint at a veterinary clinic; the debt load was relieved by uncontested bankruptcy proceedings, and if we were half our actual ages we'd be in ok (workable) shape for now.

The point in this context was in response to your description of BSD-style licensing as a worse than GPL because 3rd parties can take your code proprietary, leaving you high and dry. GPL code can also be taken proprietary whenever it is useful to build trade-secret systems out of it. Canonical did precisely so with Arch, spending a pretty penny on it, not only failing to compensate for the R&D benefits they absorbed, but also hiring away the core Arch volunteers and sponsoring a situation in which they became fairly hostile to the project. All under GPL.

It's similar but not identical to what happened with GCC, back in the day. There, too, volunteers were hired up and hostility to the core project rose. There, too, trade secret protections were part of the formula for making money from the GPL code base. The difference in the two cases is that, in the case of GCC, there was commercial incentive not to abandon the core project but, rather, to put it in different hands and change its priorities.

So, again: GPL licensing does not prevent proprietary plays on projects and doesn't prevent leaving projects high and dry as vendors muscle in.

GPL is popular with, for example RHAT, because it excludes proprietary variations on their product lines. A competitor can snarf the source and enter the market against them, but only (for most practical purposes) in ways that leave RHAT free to snarf right back.

And you've done this same rant about Open Source being some scheme to rope in developers for big companies

I'm sorry it bugs you but it is also right.

The GNU project began as a political action whose aim was to eliminate reliance on proprietary software. It has devolved, now, into building license-fee free platforms used mainly to run proprietary software.

-t

The Year of the Free Software Desktop

maethorechannen

Thu Jul 17 13:47:15 -0700 2008

manage

Why is it that so many are focused on promoting the Linux kernel on the desktop as opposed to Free Software or Open Source Software on the desktop?

I think you've got it kinda backwards. I'd say the focus is on a free desktop and as Linux is the free/open kernel with the greatest momentum it becomes the natural choice. In the same way that most free/open desktops use gnome or kde and not something like fltk or enlightenment.

Linux have to be the kernel of the F/OSS desktop when there are now other available options which may be better suited to the task?

Do you have a particular alternative in mind? If so, which one and why?

namely OpenSolaris

I keep meaning to install OpenSolaris to have a play, but after looking around the OpenSolaris site it gives the impression of being about as ready for the Desktop as Linux was five years ago with immature package management (though at least it's being done by the right guy for the job) and a lack of drivers. OpenSolaris on the server is a different story.

frenetic development process of the Linux kernel and its lack of a stable API for drivers are two possible disadvantages

I'd say that most of the frenetic pace in kernel development has little to do with the desktop - Linux runs pretty much everywhere and it's the everywhere bit that's making up most of that development. And is the API really that unstable? Personally, I'd take a kernel with an active developer base over a stable yet moribund one anyday.

The Year of the Free Software Desktop

Charles E Hill

Thu Jul 17 14:18:03 -0700 2008

manage

The AT&T vs BSD lawsuit delayed BSD at a critical time during adoption, giving Linux a leg up.

On a personal note, I was weaned on SysV.  BSD is a corruption of the true essence and alien.  :-)

Why not the BSD family?  People constantly whine about how many Linux distros there are, that they are confusing.  But it still boils down to the same kernel built around the GNU foundation.  Other differences are minor and fairly easy to work around: run levels, package management, etc.

There are 3 main BSDs on the other hand, and they aren't wholly interchangable.  What works on Free doesn't necessarily work on Open or Net.  Which to choose?

And then there is Sun and OpenSolaris.

Why did Sun create OpenSolaris?  Why not contribute the nifty parts to BSD or Linux?  Why an incompatible license with Linux?  Because in the end Sun is a hardware company.  They need a certain level of control because it is in their blood.  Sun isn't to be fully trusted.

Why didn't Unix ever make it to a mainstream desktop?  Two reasons.  First is price.  The second is Solaris, HP-UX, AIX, Unixware, Ultrix, Irix, AT&T, A/UX, Amix, DecUX, Monterey, Coherent and a dozen others affectionatly known as the Unix Wars.

The ability to make proprietary, private modifications to the kernel and basic OS opens the door to vendor lock-in and essentially, Unix Wars II.

There is a fundamental difference between BSD/CDDL and the GPL.  Yes, the GPL is viral and that is a GOOD thing.

You can build proprietary applications on top of the GPL standard OS to your heart's content.  But the core OS *must* remain free and a standard or you end up handing a great deal of power to the proprietary OS vendor.  The perfect example is Microsoft.  Windows doesn't do anything that special that isn't done as well or better by other operating systems.  The ONLY reason Microsoft can charge as much as it does for the OS is because it is shrouded in secrecy and is popular.  Take away either factor and they may charge $25 for it, but good luck.

Don't get me wrong.  I have nothing against proprietary applications nor paying for software.  I do, however, have a great distaste for a proprietary operating system.  Just like having to use and AT&T phone on AT&T wires.

The GPL is very simple: we'll share if you will.  If you want to stand on our shoulders, you will reach back down and lift us up with you.

The BSD/CDDL license is simply: we'll share.  Feel free to step on our heads on the way up.  Appropriate in some places, but not in an OS.

The Year of the Free Software Desktop

N. Dan Smith

Thu Jul 17 14:40:59 -0700 2008

manage

Well, that is why I am hoping that OpenSolaris ends up GPLv3, which some folks at Sun have hinted at.

The Year of the Free Software Desktop

Charles E Hill

Thu Jul 17 14:57:13 -0700 2008

manage

I've heard that as well.  It will be interesting to see if OpenSolaris survives what would essentially be cannibalization by Linux.

Schwartz

Thomas Lord

Thu Jul 17 16:49:04 -0700 2008

manage

Schwartz seems pretty interesting and I kind of assume a lot of the recent openness of Sun is his influence. Deity help him the first time they have a few truly dismal quarters but, overall. I think the board will have trouble pegging any losses on this openness and then the next question is what wins can be attributed to it. I tend to like Sun generally because of their heritage and, in many areas (yeah, of course there's plenty to kvetch about) -- because they tend to be some of the better sw and systems engineers in the room. It was obnoxious, dumb, needless, and social-value-destroying for RHAT to attack them so.

-t

The Year of the Free Software Desktop

Champs

Thu Jul 17 14:35:06 -0700 2008

manage

Most users are not like me, but I'm pretty sure it's the applications they care about, like I do.

I enjoy using Macs -- there's a great stack of free-as-in-speech desktop apps, and a UNIX subsystem for "the tricky stuff."  Windows is almost as good, and I use a lot of both.  They have served well as desktops over the years, and I haven't seen a compelling reason for change.

drunk looking for keys

Thomas Lord

Thu Jul 17 15:52:56 -0700 2008

manage

You know the old joke -- staggering drunk wandering around below the streetlight -- cop asks what he's doing -- "looking for my keys... I dropped them" -- "oh yeah? where were you?" -- "Well, over there [points to a dark section of the street]" -- "So, what are you doing looking here?" -- "The light is better." Yes, well, that's what it is to question kernel choices when contemplating the desktop.

The kernel is not the reason free software hasn't conquered the desktop. In reality, it's all Miguel's fault, so to speak. Or more likely, Michael's, but that's an inside joke.

Kernel choice does make a difference to a desktop but all of the current choices are perfectly adequate, modulo driver support. That isn't the problem.

The problem is the relative paucity of free software applications.

That paucity exists because the "toolkits" and "frameworks" we have use a bum architecture. I mean Gnome and KDE -- bum architectures. The problems are hard to sum up briefly but I'll try:

Both are architectural copies of MSFT's framework, itself a copy of the X Toolkit / Motif / and more directly the Andrew UI toolkit. Those, in turn, are not quite copies but approximations of the Smalltalk approach. That's the lineage. Important points are: (1) In the transition from Smalltalk to the X Toolkit / Motif / Andrew a critical element was lost -- an interactively programmable, dynamic system was replaced by a gawkward static-compile, mostly-early-binding system. The skeleton of the OO framework was parroted, but not the power and flexibility of the development environment. (2) The ported architecture works mostly OK, sorta, but only at huge developer-time expense. Work hard enough with this architecture as realized in C, C++, C#, etc. and you can eventually do what you need to get done but it is freaking expensive when measured in programmer hours. If you need to go back and revise stuff -- that is also hellishly expensive.

"Sprawling" is the right word for the kinds of applications this architecture leads to. "Abstraction" is anathema -- the toolkits lead to a proliferation of subsystems and components, each highly specialized so that re-use is exception, not rule.

Now, if your are microsoft, circa 1988 and forward for about 15 years, that works to your advantage. You've got plenty of cash and every hour of successful developer time generates at least three orders of magnitude, probably more, of return. Writing intractable code and writing code "the hard way" is no obstacle to you because you can triple your developer hour burn rate and still pull money in, hand over fist. All you need to do is develop a corporate culture that gives you enough control over "that many" programmers and you win. You win because you can bring out features by shear brute force application of labor and no competitor with less labor hours to bring to bear can possibly compete. All well and good for MSFT to use this architecture because if anyone tries to clone it they're doomed -- the stupid labor-inefficiency of this abstractionless, every-step-painfully-hard architecture is, from the MSFT perspective, a barrier to entry against competition.

Along comes the early GNU/Linux vendors and punk Miguesl and betwixt them they decided to bet their whole desktop wad on COPYING THOSE ARCHITECTURES.

That is where they shot themselves in the foot. Way back then is where the GNU/Linux vendors ceded the desktop. They can't (and shouldn't be able to) afford sufficient labor for that software architecture. Yet they banked on it anyway.

It's not like there weren't better ideas around. Architecturally, Emacs had a great deal right and Joel "Don't Fidget with Widgets" Bartlett had at least the essence of display right. That was the right direction. It was the direction of the original (ahem) Guile project. And it was starting to come together right around the time it was maliciously disrupted and a junior hacker, Miguel, was thrust, way too soon, into the limelight for all the wrong reasons.

You might say, if you "get it", that "Well, Tom, In that case isn't the architecture of a modern A-list browser, with it's DOM-based display and Javascript extensibility exactly what you want?" Indeed, it's kinda close and I would point out that the one area where free software actually does compete on the desktop is exactly there. Details matter and there's much sloppy and wrong, overly standards-constrained, etc that holds things like Firefox back but - yes, not only is it abstractly the right architecture but it is, confirming my view, the main success.

The kernel is not that important. It's the framework for GUI that matters. The vendors screwed the apocryphal pooch on that one.

-t

The Year of the Free Software Desktop

Matt Fulkerson

Thu Jul 17 21:48:25 -0700 2008

manage

I see a huge problem with OpenSolaris catching on with the general public.  Linux distributions have gone through a phase where they didn't always "just work" right out of the box.  But on the other hand, they have always been easy to learn to adminster.  Take any hand-me-down PC, and you have the perfect platform for learning to use a Linux distro while not worrying as much about breaking anything.  In contrast, who can gain experience with Solaris?  Until very recently, only IT professionals.

As a long-time Debian user, if I have to use Solaris, the first thing I do is install the GNU utilities.  Then I run back to my Debian desktop and log in remotely so I don't have to deal with Solaris' GUI and terminal.  In Debian, the GUI and terminal are ideal because you have your pick of your favorites easily installable via apt.  Don't like Gnome or KDE, then go with XFCE, which is insanely easy to use and customize.  So I have to agree with Nathan's point that the focus should be on Free Software on the desktop.  Of course, the GNU/Linux kernel itself is free software.  Now along comes OpenSolaris wanting to be free, too.  Debian could very well support another kernel.  So my question is, why might Debian/GNU/Solaris be a compelling alternative to Debian/GNU/Linux, either for free software developers or desktop users?

Kernels?

zogger

Thu Jul 17 22:55:38 -0700 2008

manage

I don't get why the hurd project is so routinely dismissed and ignored. What made that particular endeavor just falter forever?

The Year of the Free Software Desktop

steelbytes

Thu Jul 17 23:55:52 -0700 2008

manage

ok, my comment is not why linux instead of bsd etc, but ...

re 'year of foss desktop':  where is the "Killer App"?

as a developer I write c/c++ for windows, and an increasing amount of php/html (Still avoiding js/ajax/etc).

if this trend away from desktop apps towards browser based continues, then the OS and hecne the kernel has very little todo with desktop experince.

killer app

zogger

Fri Jul 18 08:14:46 -0700 2008

manage

Well, the killer app now seems to be social networking (like this forum for instance) along with advanced search functions (search combined with maps and pics and GPS and so on, plus connections to connections with friends, online, texting, chat, etc), with a segue into mobile computing, being able to be connected with a small portable device and have the experience *not suck*. Business interests are well covered, there's just so many ways to type up a report before it gets to the ludicrous phase of features I think, and spreadsheets are old hat.