Technology

"Solitaire is all anyone will ever need"

Christopher Wright — Wed, 10 Jan 2007 09:49:04 -0600

One Laptop Per Child is, in my opinion, a laudible and worthy goal... to create low-cost laptops that are then distributed for free to children in developing countries.

This is a worthwhile effort. There are people who feel the project is a waste of time, and that people should be focusing on other, more basic problems, but I believe that focusing on any one problem to the exclusion of all others won't solve anything. OLPC won't save the world but that doesn't mean it won't do good things.

That said, sometimes I worry about what exactly people involved in the project expect the children are going to get out of it. Take, for example, the following quote:

"A child doesn't want to play the latest video games. He wants to be able to read a book."
-Michalis Bletsas, OLPC official, as quoted in Linux Today

This quote is offered as an explanation as to why the laptop isn't powerful enough to play the latest and greatest games available on the market today, and it makes me want to bang my head against the wall until I lose consciousness, just to give me a moment of sweet respite from the silliness of the idea.

There are good reasons why the laptop in question shouldn't be able to play the latest and greatest games. The most significant is that it's designed to use as little power as possible so that you can use a hand crank (actually, it seems to be more like a rip-cord) to power it up for short periods of time -- this automatically rules out the power hungry high-end processors on the market today. And it's perfectly reasonable, given that limitation alone, that it be designed for a more utilitarian purpose. Add to that the attempt to make building such a machine as cheap as possible, which once again rules out the latest and greatest processors on the market, and you have a legitimate case for focusing on your primary purpose -- building a machine to be used for education -- and eschewing fancy extras like supporting cutting edge 3d graphics capabilities.

But to suggest that children wouldn't want to play games, given the opportunity?

You know, when I was growing up I loved books. I was what some would describe a "bookworm." I read as many books as I could find. But I was a lot more likely to stay up until 3 or 4 in the morning playing Ultima III than I was reading a book. Given the choice between playing a new game and reading a book, I could have gone in either direction.

Maybe children in developing countries are substantially different in temperament from children in developed countries, and they hunger for knowledge while scorning digital entertainment, but I'd be willing to put money down that this isn't the case. Games are fun. I don't know why anyone would think that children aren't interested in having fun.

It should be enough, when asked why their laptops won't play World of Warcraft, to say "we're building an educational machine, not a gaming machine. To make it a good educational machine that works in any number of challenging environments, sacrifices had to be made." Implying that children would be uninterested in such games if they were unavailable reflects a philosophy about human beings that strikes me as unrealistic at best.

What Bubble Are We Talking, Here?

Christopher Wright — Fri, 22 Dec 2006 17:48:10 -0600

General Protection Fault, by Jeff Darlington. Now that's when you'll know there's a bubble...

This will a gentle cut, a minor nick if you will -- both because of the approaching holiday, when my thoughts ought to be focused on Love, Joy, Peace and GoodWill Towards Man, and also because I think the transgressor in this particular instance is more guilty of being too close to the perceived problem than he is of any kind of egregious journalistic excess.

That said, this caught my attention because it made the lists on Slashdot, and even the briefest episodes innocuous hyperventilation can suck all the air out of a room if you stuff enough people in it and get them to start hyperventillating at the same time.

So, in brief:

Slashdot posted a news article entitled "Has the Desktop Linux Bubble Burst?" which refers to an article on OSNews.com entitled "Has the Linux Desktop Bubble Burst?" writen by a guy who wonders, essentially, if maybe the Linux Desktop "bubble" has in fact "burst."

The article points out that development for the GNOME and KDE desktops have slowed since 2002, when the latest full dot releases for each were released (3.0 for KDE, 2.0 for GNOME) and then follows that up with "not much has happened since then," notes the introduction of very sophisticated desktops in Apple's OS X and Microsoft's Windows Vista, and bemoans that Linux may be falling behind in the desktop world.

Specifically he uses the terms "bubble" and "burst" -- the very combination of words that was used to describe the crash and burn of Dot-com speculation at the turn of the 21st century, and is being used right now to describe the eventual downfall of Web 2.0 speculation, and will inevitably be used to describe the downfall of any kind of popular technology trend you can think of.

But bubbles and bursting are metaphors used to describe hype-driven phenomena. The hype causes the bubble to expand, and expand, and expand, until the hype exceeds the bubble's capacity to contain it and then it disperses into nothingness. The Dot com bubble burst because there were too many stupid people doing stupid things on the web, and investors could no longer ignore that they were throwing money at idiots. The Linux desktop, on the other hand, is not driven by hype -- it is driven by necessity. People who develop for GNOME and KDE don't do it because they want stock options, it's because they want a good, stable, working Linux desktop.

And unlike the Dot-com fiasco, where talented programmers were paid too much money to develop ecommerce applications that allowed you to buy shoes for your neighbor's goldfish online, development has been focused on mostly realitic goals and has borne fruit. KDE 3.0 was a substantial improvement over 2.0, which was itself a signficant step up from 1.0 -- but KDE 3.2, 3.4, and 3.5 were all important dot releases (sorry 3.3, I don't remember a thing about you). Linux is the only thing I use on my laptop (my primary computer) and the problems I have with it have nothing to do with the desktop and everything to do with device drivers.

I don't see the problem the author is describing, and I wonder if he really sees what he thinks he sees, or if he's seeing something else entirely. There is a difference in kind between "hype" and "excitement" -- I'll freely grant when each new version of a Linux desktop is released there's a fair amount of excitement in the community (and a fair amount of sniping -- but hey, that's computers for you) but we have yet to see any of the silliness that is so prevalent with hype. We get "hey, we created a dialog that lets you set up a printer!" instead of "now you can use your wireless network application to turn off your refrigerator!"(1) We get "our built in browser has one of the most advanced and complete implementations of CSS-2 currently available" instead of "this shiny, glossy title bar will make you want to lick the screen!"(2)

There are no Linux desktop startups buying $500 chairs, there are no Linux desktop-themed knicknacks being sold on eBay for thousands of dollars, there are no Linux desktop board games. No one has sold their stock options associated with GNOME or KDE in order to become ridiculously rich, and none of the Linux desktop programmers have become firmly ensconced in the world of the online punditocracy. Neither KDE nor GNOME can realistically be considered a "lifestyle choice" -- all KDE/GNOME flamewars are fought over old ideological differences and current conflicting development goals. And there is no sprawling economic infrastructure that depends entirely on the success or failure of the Linux desktop. In short, there is no bubble, and so there is nothing to burst.

Is it reasonable to worry about GNOME's loss of support and manpower? Yes. Is it understandible for people to wonder about the effect SuSE switching from KDE to GNOME will have on KDE development? Yes. But taking those concerns and turning them into a doomsday scenario where Linux desktop development collapses due to an overinjection of rabid hype -- no. Not reasonable in the slightest.

--------
(1) I'm not saying those applications don't exist for Linux, I'm just saying they're not used as a selling point for a user interface.

(2) Oh Mr. Jobs, why did you endorse such poor hygienic behavior when we are already so challenged in that arena? Thousands of computer users licking their monitors... my soul cries out in horror.

The Incoherent Values of Technology Journalism

Christopher Wright — Tue, 19 Dec 2006 16:28:56 -0600

From Superosity, by Chris Crosby.

Over at Ubersoft.net I took Infoworld to task for publishing a list of technology predictions for the coming year that were, as near as I could tell, mind-bogglingly lazy in scope. But Infoworld, it appears, is only a minor-league player in this vast wilderness of hyperbole-riddled pablum, and they have been trumped by another publication whose recent proclamations are so egregious that I was compelled to take Eviscerati.org out of the mothballs a little early (I'd planned on a 2007 revival) in order to do them justice.

CRN bills itself as "Vital Information for VARs and Technology Integrators." One assumes, then, that the information it chooses to publish on its site is information that VARs and people who spend their time integrating technology absolutely must have. The truth of this I leave up to those VARs and technology integrators who actually read the publication, since I am neither -- unless compiling the most recent version of ndiswrapper on my Kubuntu Edgy laptop in order to get wireless access counts as "technology integration." Still, after picking up on this little tidbit from Slashdot, I have to wonder if perhaps the VARs and technology integrators are getting their money's worth when they read this publication.

Like all publications that attempt to convince their loyal readers that they have their finger on the pulse of whatever part of society they are covering, CRN engages in end-of-year navel-gazing. Of particular note this month is their 2006 Products of the Year, a Top Ten list that purports to tell you the ten most important products that were released in 2006.

One of those products? Microsoft Vista.

That's right. Microsoft Vista, the operating system that was released to businesses eighteen days ago, has been chosen as the top operating system of 2006. Eighteen days of availability for this operating system that businesses aren't even that enthusiastic about adopting is enough for them to crown Vista the top operating system of 2006.

And when you read their article, you have to wonder if these poor VARs and technology integrators might be better served getting their information elsewhere. When you read their article, it becomes increasingly clear that CRN's view of the technology landscape doesn't exactly mesh with a little something most people like to call "reality."

From the article:

While debate rages over whether the five-year wait was worth it, the truth is Vista is pretty much the only game in town.

The above quote is 100% factually correct as long as you ignore OS X, Solaris, BSD, BSD, Linux, Linux, Linux, Linux, Linux, Linux, Linux, Linux, Linux (and Linux), and a few other BSD's and Linuxes I missed along the line.

We might be able to take the article's claims more realistically if they changed their justification slightly -- from "pretty much the only game in town" to "pretty much the only game in town that buys enough advertising space for us to consider saying nice things about them" -- but as they haven't done that, we can only sit back and shake our collective heads as we wonder what they were smoking when they reached this decision, and whether or not it would be worthwhile to find that stash.

No, the real reason Vista made the list is because it is new and it is shiny and they thought "hey, nobody ever got fired for writing about Microsoft" (unless you're writing bad things, of course), and it didn't matter that it hasn't been in use in the corporate world long enough to determine whether it meets any or all of Microsoft's claims.

Essentially, dear friends, Vista was a "Product of the Year" not because it's good -- it's too early to tell that yet. No, it was made a "Product of the Year" for 2006 because it was released in 2006. If it had been released in 2007, it would have won a spot in the 2007 "Product of the Year" article, even if someone else released an operating system that never crashed, was immune to viruses, ran every program ever made flawlessly and was not only free, but came with a money-back guarantee.

Microsoft's standing in the industry of tech journalism is set in stone. Microsoft will cease to be a darling in the industry of tech journalism only after it has already been ground into dust by some other company, and then that company will become the new media darling that can do no wrong.

Why? Because with a few notable exceptions, technology journals aren't interested in actually reporting technology-related news, or seriously researching new technology trends. By and large they exist for one reason only: to reassure the CTO that all the new technology he just bought was a good investment. Whether it was or not is irrelevant, of course. In fact, as long as he can point to a media article ebulliently lauding the latest and greatest turnkey solution, it won't matter how much his employees in IT complain that it just isn't working the way it's supposed to.

Still, it's hard to blame them. Who wants to read a publication filled with articles that tell you that your software doesn't work correctly, you paid too much for it, and it's a looming security hazard that will bring your company to its knees? Well, there is a market for that, but it's much smaller than the market that will pat companies on the head, tell them they're doing fine, all the while getting money and presents (oops -- I mean "evaluation copies") from software and hardware companies.

In fact, I've changed my mind about the entire thing: if companies will start bribing me (oops -- I mean "sending me evaluation copies") then I'll be perfectly happy to come up with a top ten list and feature each bribe (silly me, of course I mean "rigorously evaluated and extensively tested product") in it. If I get more than ten items, I'll make it a top twenty list. Or a top one hundred list. Whatever works.

Sure, it'll probably take some of the charm out of the site... on the other hand, I'd probably post more.

A Little Story of Things Going Horribly Wrong

Christopher Wright — Mon, 31 Oct 2005 14:27:00 -0600

El Goonish Shive, by Dan Shive.

I currently have two computers. Alex, my Windows XP machine, is used for games, file backups, and studio recording. (Mostly games and file backups these days, because my home studio is in disarray, but it's slowly getting sorted out.) Mark, my Linux laptop, currently runs Xandros Linux and is what I use for writing, web design, and my work on Help Desk and Kernel Panic. Of the two, Mark is currently FAR more important to me.

Which is, I believe, why it winds up breaking all the time.

I meddle, you see. I'm a computer meddler. Most of what I know about computers comes from me "messing around with stuff" until it breaks. It's not that I'm particularly talented when it comes to computers -- I'm actually rather thick -- I've just been blessed and cursed with a certain amount of fearlessness when it comes to trying new things. Blessed, because that means I get try new things that are, on the whole, pretty neat. Cursed, because I'm really not talented enough to do things right the first time, which means I spend a lot of my time trying to fix things that I've broken in a spectacular fashion.

But I meddle. Any computer I sit in front of I'm compelled to muck about with in some fashion. And so when I use a computer all the time, the chances of me meddling with it to the point where it rolls over and dies increases to the point where I have turned system recovery into an art form.

Last week, for example, I was trying to compile drivers from the Linux Wacom Project in order to use a Wacom Graphire digitizer tablet I have sitting around the house (not the Graphire 4, which the link displays -- an older model). I did not succeed in getting the Graphire tablet to work. I did succeed, however, in changing the read/write/execute privileges of the entire /usr directory so that only the root account could access it.

It took me some time to figure that one out, but eventually -- with the help of some guys I know who know more about this stuff than I do -- I was able to figure it out.

One week later, I decided to try to download the latest version of AbiWord, a pretty neat word processor that has a few features no other word processor for Linux has: specifically, a word count feature that updates as you type, and a grammar checker. I was curious about the grammar checker, mostly, and that was my undoing.

The latest version of AbiWord is available using Autopackage, which is an interesting new way of installing software for Linux. It's supposed to be able to see what dependencies are needed in order for the application to install, and then go get those dependencies and install them for you. Inkscape, the vector graphics application I use to create Help Desk and Kernel Panic, uses autopackage to great success.

I didn't have as much luck with AbiWord. Autopackage refused to install the damn thing. First it needed library X, then it needed library Y, then it needed something else entirely... I spent a few hours on Saturday typing "apt-get install [something]" in order to get past one dependency problem, just to run into another. And another. And another.

Finally I came to Link Grammar.

Link Grammar is what AbiWord uses for its grammar checker, and autopackage couldn't find it on my system. Which makes sense, because it wasn't there. But there was no way to get it easily ("apt-get install link-grammar" informed me that no such thing existed on any of the Xandros-friendly download sites I'd set up on my system) so I had to do something that I should have known better to avoid: I had to compile the damn thing myself.

Compiling the Link Grammar file available on the official site was pretty easy, but ultimately useless... it created a program that ran locally, from my user's account, but didn't actually put itself anywhere that the AbiWord autopackage could see it. AbiWord had an updated version of the program available, so I downloaded that.

It was substantially different from the original version... with exactly the same install instructions.

In other words, the install instructions said "go to the directory with the Makefile, and type 'make' in order to product the binary file you need to execute to run the program."

The latest version of link-grammar hosted on the AbiWord site didn't have a damn makefile. No, you actually had to run a configuration script in order to create the damn makefile... and no instructions as to how to do it.

But hey, I'm a man of the new millenia. I have an internet connection, the internet has Google, so I went to google and typed in the phrase "link-grammar abiword install problem" and what do you know -- someone was having the exact same problem I was! And they posted it to an abiword mailing list!

Unfortunately, the stated solution to the problem was something I'd already tried (which hadn't worked) involving the original release of Link Grammar, not the modified one available on the AbiWord download site. In fact, I could find NOTHING helpful detailing how to install the modified version on the AbiWord download site...

So what do I do when documentation fails me? I ask one of my friends who knows more about this stuff than me. So I went over to a forum that I frequently go to when I Don't Know What The Hell I'm Doing, and I asked them what the Hell to do next. And I got an answer... one that worked... and autopackage saw link-grammar, and moved on to the next thing...

... and the install failed again. I needed to download something else. And once again, it went a few yards further, and failed, needing something else.

At that point I gave up and turned of the laptop. That was last night.

This morning before I work I turned on my laptop to check my email. The laptop seemed to boot just fine, right up to the point where the logon screen was supposed to display, at which point the screen went completely blank.

Yep, that's right. I managed to screw up X Windows yet again. Oh, and National Novel Writing Month starts at midnight tonight, which means that this evening I will not be trying to troubleshoot the problem, I will instead be trying to re-install Xandros on my laptop... because it will take less time. I'm an old hand at re-installs, you see. I know exactly how to do it, exactly how to recover my information... it'll take all of three hours. A three hour re-install to get me back to where I was, compared to three hours of troubleshooting to figure out exactly how badly I've screwed myself over... there's no contest.

I have no-one to blame but myself, see. It's tempting to blame AbiWord simply because they updated link-grammar without actually updating the install instructions, but really, I should have given up the very first time the autopackage installer told me I needed to install something else first before it could continue. That should have clued me in that I'd wandered in too deep, that I was way over my head, and that to go any further would seal my fate. It's not like I needed the grammar checker. But as soon as it reached that point, the point where I should have turned away... the entire thing became a shiny red button that I couldn't help but press. So I pressed it, and it blew up in my face. I am an arrogant creature, full of hubris, and I got mine.

So anyway... Mark is broken. With any luck I'll fix in time for tomorrow's comic and the 1,667 words minimum I want to write between the hours of 12:01 and 2:00 AM.

This post? Therapy. Nothing more to see here, please move along.

Maybe it's just a "Practice License"

Christopher Wright — Mon, 22 Aug 2005 03:47:46 -0500

From Superosity, by Chris Crosby.

The Free Software Foundation has announced that they are working on an update to the GNU General Public License. The proposed revisions aren't available for review yet (and won't be until early 2006) but they're already talking about a few of the changes, and one change in particular is provoking a lot of worry and speculation. The FSF is trying to downplay the extent of the changes, but at the same time they come across as just a little bit defensive when it comes to defending their right to update the license.

One particular change is causing some concern both within and outside of the free software community, and only time will tell if this change is going to go into the actual revision, or be abandoned as part of the FSF's own "practice movie." If it does get included, however, it suggests that the FSF is narrowing its focus. Some of its ideals aren't co-existing very well, so one of them is getting downgraded from "core belief" to "looks nice on paper."

Ideological triage: the unfortunate result of a successful revolution.

For the purpose of this discussion let's ignore the differences within the free software community. Let's ignore the practical and ideological differences between free software and open source software, and the spats between the various factions within each group, the personality conflicts, the feuds, the betrayals, the meltdowns, and all the other things that make gossip so much fun. That's much too complicated to get into at the moment, and this article can't afford to get any longer than it already is. We're talking specifically about the GPL, the GPL was specifically created to further free software, and all the other groups, whether they like it or not, exist because the GPL came first.

I don't care what kind of peg you are. This article is a round hole, and I have a very big hammer.

The terms of the GPL are a little difficult to understand in its native tongue, but to its credit it speaks a much clearer dialect of Legalese than most software licenses. The license spells out the rights you are granted concerning the use and distribution of a program that uses it. Under its terms, you are allowed to examine and modify the program's source code... but if you distribute that code for any reason, you must make the modified source code publicly available to whomever you distribute it to, free of charge.

In other words: if you change something for your own use, you are not required to make those changes available. If you change something and allow others to use those changes, you must provide them access to the source code so that they may do the same.

But, claim some, there is a fly in the ointment. A new business model has appeared that seems to skirt around these requirements, and -- they claim -- subverts the very essence of what the GPL is trying to do. This supposed loophole is that the GPL puts restrictions on the distribution of source code, but not on how that source code is used. So if a company modified a GPL-Licensed web application -- an application that ran on one machine, and one machine only -- they could charge people who wanted to log into that machine and use that application, without making the source code available. They aren't distributing the program, you see... they're just using it.

As it happens, this fly in the ointment has turned into something of a viable business model today. Companies like Google, Yahoo!, and Amazon.com -- companies that provide services over the web, as opposed to software -- all run software licensed under the GPL. They all have modified that software, optimized it to fit their specific needs... and as far as I know, none of those specific optimizations have made their way back into the free software community. Why not? Because those changes aren't being distributed. They're running on servers sitting in each company's headquarters.

This, some say, is a loophole that absolutely must be closed. When it was created, no-one ever envisioned that "web services" would be as prevalent as they were today. No-one understood that it could actually be profitable for companies to make money in such a manner. This unintended consequence is a result of the fast-changing face of business in the computer age, and it needs to be fixed before it renders the GPL irrelevant.

The problem with this view is that, well, it's completely wrong.

This supposed "loophole" in the GPL has been known about for at least six and a half years -- that's when I first became aware of it -- and at the time it was very specifically described as not a loophole at all, but as a consequence of liberty.

And who described it this way, you ask? I'll tell you: it was Richard M. Stallman -- the man who wrote the GPL and fathered the entire Free Software movement.

Let me take you back to January, 1999. A developer of a web application is considering using the GPL as the license for his product, but he's concerned about something:

But there is a problem... Imagine this...

1) We distribute our web software under the GNU copyleft

2) Big company X, takes our software adds extensions to it, builds a web site around it, and they puts our site out of business.

This is exactly the "loophole" that everyone is talking about today, and that the changes in v3 of the GPL are supposed to "fix." But on January 14, Richard Stallman posted this response on gnu.misc.discuss:

It is definitely possible for someone to make a modified version of a free program, and not release it.

However, it would be misleading to describe this as a problem with the GNU GPL in particular, because the same is true for any free software license, including the MIT Scheme license that you use now.

In some cases, I would agree that it is antisocial to make a useful modification and not release it. But if a license made that illegal, if a license did not allow people to make and use their own private changes, without publishing them, I would consider the license unacceptably restrictive. People should always have the right to share with their neighbors, but compelling people to share can be wrong, especially when it conflicts with privacy rights.

Still don't beleive me? You can read the entire discussion right here.

It's not that they didn't think it was a problem. There are always problems attached to freedom. One could argue, very persuasively, that one of the essential costs of freedom is that you have to put up with people using it in ways that piss you off. Richard Stallman's post made it fairly clear that he wasn't in favor of the situation, but he was also very clear that, in his view, there was a greater principle at stake.

So what changed?

A lot of things changed. Software products licensed under the GPL or another free software license have become an integral part of the internet, are working their way into the corporate world, and are even being officially adopted by governments. At the same time, companies who are reluctant to fully embrace the ideals of that movement are trying to find ways around it.

Some are beyond reluctant -- some are positively desperate to try to find a way to "break" the free software movement, and the GPL in particular. SCO went so far as to try and (unsuccessfully) claim that the GPL was Unconstitutional.

Companies don't like it when you tell them they're not allowed to profit off of your work without compensation. They don't like it because it's one of their most profitable business tactics. The GPL is, in essence, beating these companies with their own schtick.

So the FSF can be forgiven for paranoia where their baby is concerned. They don't want to rest on their laurels, and quite rightly so -- complacency makes it more likely that a dastardly villain will find a way around the license that renders the entire thing moot. The next thing you know, the GPL is tied to the railroad tracks, the villain holds the mortgage ransom, and there is no square-jawed hero on hand to save the day.

On the other hand, it's hard to shake the notion that if they do go down this route, they are essentially trading liberty for security. The FSF has always been very up-front about the fact that they are an organization that pursues an ideology of freedom rather than a development model. Is the situation so dire, then, that they must sacrifice a piece of that ideology in order to save a greater part? Or are they simply distracted by the annoyance of someone using that freedom to their benefit without sharing their good fortune with the free software community as a whole?

I admire the FSF, and regardless of what they ultimately decide to do I suspect I'll continue to admire them. That said, I find I agree with the Richard Stallman of 1999 more than I agree with their current rumblings concerning this "loophole," and I think calling it a loophole is inappropriate considering the author's previous stance. Liberties of any kind are prone to misuse precisely because they are liberties, and the freedom to behave only within rigidly defined boundaries isn't really freedom. In order for that word to have meaning, it must anticpate and, at least to a certain extent, tolerate offensive behavior.

Closing this "loophole" doesn't seem like an act that promotes freedom. It seems a lot closer to the behavior of proprietary, non-free organizations who consider any use of technology that interferes with their goals as an act of piracy.

I'm going to resist the temptation to quote Ben Franklin, since I've already sort of alluded to the quote in question and it's become somewhat cliched of late (though not a whit less true for all its over-use). Instead, I'm going to run off somewhere to hide while the teeming masses of angry free software developers marshall their forces against me. And as I run, gibbering in fear, certain of the knowledge of my impending doom, I ask: if this change goes through, what freedom is being protected? Under the current terms of the license, what freedom is being lost?