[Mon May 23 09:54:44 CDT 2005]

Slashdot publishes a story today about how Google will be offering personalized homepages soon (check here for an example). I truly like the way these Google guys have managed to settle on a clean and simple design that is nevertheless popular. In other words, what I like about Google is that they have managed to prove something I have always believed: users care mainly about the functionality, and they hate to see cool designs getting on the way when it comes to computers. Just compare the design linked above to the humongous personalized pages other services have set up. Sure, Google only offers a fraction of the data but we all know that too much information just ends up being noise. {link to this story}

[Tue May 17 11:49:26 CDT 2005]

Much has been written about Debian's eternal problems to release new versions of their Linux distribution. It just takes years for them to release a new stable distribution. Well, it appears as if they are not the only open source project experiencing these problems. Scott Long wrote a short article about FreeBSD 5.x and the future telling us about their own release problems:

By the middle of 2002 is was very apparent that we needed to start focusing on getting 5.0 released. Unfortunately, we fell into the trap of wanting to finish more features in order to feel good about 5.x. We kept on ignoring the fact that 5.x already had a lot of good and needed features, and that the number one goal needed to be to get it stabilized and turned into 5-STABLE. Instead we drew up a road map document that dictated releases based on features rather than on stability and, even more importantly, timeliness.

So, as it has been the case with other open source projects lately, the FreeBSD folks followed the lead of the KDE and GNOME developers and have decided to move to scheduled releases. From now on, there will hopefully be a new -STABLE branch every 12-18 months and new point (or minor) releases every 4-6 months or so. I would find it quite surprising if the Debian developers do not adopt a similar plan sooner or later, especially after Ubuntu proved it is perfectly feasible. Of course, the problem with Debian is the sheer amount of software that is bundled in their repositories (and, therefore, officially supported), not to talk about the incredible amount of architectures they code for. Still, a two-tier approach like the one suggested a couple of months ago simply makes sense: allow certain key architectures to be a part of the core Debian distribution, and then support the rest unofficially via something like Ubuntu's Universe packages.

By the way, since we were talking about FreeBSD, Daemon News tells us about a new service called BSD Updates that make it really easy to download and install binary patches to your FreeBSD box. I know many BSD'ers hate this, but it definitely was one of BSD's shortcomings. I ignore if the service is free or not. I still have to give it a try to update my server at home. {link to this story}

[Tue May 17 11:16:20 CDT 2005]

While reading a piece on ZDNet speculating about the idea of Microsoft buying Red Hat, I encountered the following thoughts:

The biggest threat Red Hat faces right now is that IBM could settle with SCO and then release its own Linux along with workstations and servers based on the Cell processor. With SUSE essentially out of the picture, Linspire in a world by itself, and Debian not getting the press it deserves, such a move by IBM would leave Red Hat with nowhere to go except a suicidal head-to-head competition with Microsoft in the x86 marketplace. Given that Cell outpetforms x86 by an order of magnitude and doesn't have the security weaknesses built into the x86, this would leave them fighting to hold an ever decreasing share of a shrinking market.

Yes, that is what I thought too. It sounds like a lot of hocus pocus to me. Microsoft buying Red Hat? IBM building its own Linux distribution? A new processor taking over the world and destroying Intel? Linspire? Huh? In any case, I had never heard of this new and revolutionary Cell processor. It appears to be the product of Sony, IBM and Toshiba to power the next version of Sony's PlayStation video game console. Sure, some of its features do sound quite exciting:

"The Cell architecture is designed to make grid computing almost universal," Halfhill said. "It makes distributed processing part of the design. If you have several of these machines on a network, the work can be spread across a network."

Cell also implements a number of on-the-chip security measures, mostly aimed at preventing unauthorized copying or distribution of copyright content, Halfhill said. Such functions typically are handled by software that sits on top of the chip, but Cell bakes security into the silicon with innovations such as a memory design that allocates memory into secure chunks. That way, only an authorized application can access a protected piece of content.

Still, we have been here before, right? Whatever happened top MIPS, Alpha and so many other exciting platforms? It sure could be different this time around, but for the time being all those predictions of Intel's demise sound at the hands of this new processor sound quite exaggerated. {link to this story}

[Mon May 16 12:41:06 CDT 2005]

Onto another issue also discussed in the Debian mailing lists. Martin Schulze sent an email to debian-security asking what to do with PHP applications that have include files placed directly under Apache's DocumentRoot, therefore allowing users to read them from their browsers. Needless to say, this provides some nice information on how to exploit the applications themselves and leaves them wide open to cross-site scripting attacks. Yet, it truly is a common practice in the case of many PHP and Perl applications out there. Henrique de Moraes Holschuh simply answered that configuration files should be under /etc and it is OK for packagers to change that. It is quite frightening, but the amount of apps that do this is just amazing. Just spend some time to look into the directory structure of your own server to find out. {link to this story}

[Mon May 16 12:34:03 CDT 2005]

Talking about silliness. Adrian Bunk sent an email to the debian-devel list wondering whether the all GPL'ed programs would have to be moved to non-free since the GPL does not allow to modify the license itself. Huh? Talk about mental masturbation. Thank goodness, Glenn Maynard clarified that including license texts in packages is an unavoidable requirement of distributing any software, plain and simple. I am telling you, sometimes Debian's legal discussions truly go too far. {link to this story}

[Mon May 16 11:56:23 CDT 2005]

I understand people may like one or another Linux distribution for whatever reason, including just personal preference. However, there is something silly about the idea of Slackware not supporting NPTL until its most recent -current release when most other distributions have supported it for years now. Mind you, we are not even talking about a distro that takes a considerable amount of time to come up with a new version, a la Debian. These guys are already in version 10.1 and had not included NPTL support all this time. As I said, I can see this being an issue of personal preference, but then one comes across these aggressive postings by certain advocates who want one to believe that their distribution is the best thing since sliced bread and... well, that is when it gets sort of silly. {link to this story}

[Mon May 16 11:02:00 CDT 2005]

Anybody who follows the Linux mailing lists out there must have seen a thousand times those annoying messages people send asking how to unsubscribe from the lists or simply attempting to unsubscribe by sending an email to the list itself. Of course, had they even bothered to read the lines at the bottom of any email sent to the list they would know how to unsubscribe, but who would even think about doing that, right? In any case, Adam Hardy recently came up with a quite humorous proposal to one of these "unsubscribe" emails. :) {link to this story}

[Fri May 13 10:15:20 CDT 2005]

The Java versus C# debate has been going on for a while now in the GNOME community, but now Havoc Pennington has provided some more information as to why he does not think Mono/C# is the right choice:

Enough on technology. On the crux of the issue, what I wish GNOME would understand: Red Hat not shipping Mono is currently a can't rather than a won't. Making it worse, we are not able to spell out all the facts on why we can't. In fact, none of the Red Hat engineers (including myself) knows all the facts (and we couldn't share them in public if we did). We continue to try to get this issue resolved and there are serious efforts in progress, but, well, they have been in progress for quite a while. At some points it felt close to resolution, at the moment it doesn't really, at least not to me.

Please don't ask for more details here. I can't give them. You can take this as evidence that I'm lying, and that's fine, but I can't do anything about it.

I believe we have legitimate and non-evil reasons why we can't ship Mono. And I think open source Java looks plausible and a lot nicer than C; Java and Classpath will even run on Mono, and if C# becomes more viable later, experiments such as Graydon's or the Lucene port show that it isn't hard to do a Java to C# conversion. And guess what, we need open source Java in the desktop anyhow for OpenOffice.org and the browser plugin at minimum.

I don't know what people expect Red Hat GNOME developers to do. We can't roll over and say "OK, we'll start hacking in C#, even though we don't see a path to shipping any of the stuff we're hacking on" - does anyone seriously expect that?

Again, the problem with Mono is not (and it has never been) technological, but rather political. I know there are many people out there who would rather not deal with this type of issues, but that will not make them go away. The reality is that betting the future of the open source community on C# will leave Microsoft in charge. Yes, they can talk all they want about making it a standard, but if history is any help Gates & co. will never let pass a chance to use a product of theirs to screw the competition. It is as simple as that. Do you trust your future to Microsoft or not? Do you truly believe that Gates had a sudden change of heart and has now decided to play nice? I sure do not, which is why I would rather see open source developers using Java. {link to this story}

[Fri May 13 08:47:08 CDT 2005]

Extreme Tech publishes an article on how to set up your own Linux home theater PC. It definitely sounds like a lot of work right now, no matter how attractive the idea is. My wife will have to wait until someone comes up with a nice and easy setup... either that or she will have to give me a week off to do it all by myself. :) {link to this story}

[Thu May 12 14:28:23 CDT 2005]

GNOME Desktop published today a short story on the latest discussion by the GNOME developers comparing the different source management systems freely available out there that is well worth a mention. Apparently, James Henstridge took the time to discuss CVS, Subversion and Arch, including a quick comparison of the key commands for each one of them. But there is something that called my attention from the discussion that took place in the gnome-hackers list whn Luis Villa wrote the following:

FWIW, for those not there, the big argument made by the pro-arch camp at the boston summit was that svn was merely incremental improvement over cvs while arch had the potential to really change how we do things. I can't really speak to that one way or the other, not really having used either, but that is just the background.

{link to this story}

[Thu May 12 13:53:40 CDT 2005]

Oh, surprise! Computer Economics has carried out a very informal survey on its website inquiring about the advantages of open source software, and the "less dependence on vendors" category won by a good margin. It sure should not shock anybody who knows anything at all about open source. The "free as in beer" story is the one that people remember the best, and it is also the most anecdotal and easier to publicize in the headlines. On the other hand, the true advantages of open source tend to be a little bit more hidden and less talked about: reduced dependence on software vendors, customization, lower cost of ownership... {link to this story}

[Thu May 12 13:24:43 CDT 2005]

I had to help someone today setting up a restricted account on an IRIX system so that certain users could only use sftp to download and upload files without being able to gain shell access to the system. Along the way, I came across a little tool, called rssh, that makes it quite easy. I only had to download it from their webite and follow these directions:

• After extracting the software, compile with GCC using the following flags: --prefix=/usr/local --with-sftp-server=/usr/bin/sftp-server (again, this is on an IRIX system, and it may differ on other flavors of UNIX)
• Edit the /usr/local/etc/rssh.conf file to enable only sftp by uncommenting the line that reads: allowsftp
• Add /usr/local/bin/rssh to the /etc/shells file
• Edit the entry for the newly created account in /etc/passwd so that its default shell is /usr/local/bin/rssh

Well, that is it. It seems to work as advertised, and the user who logs in has a very restricted account that can only be used to upload and download files from sftp. Needless to say, and since the tool is not included in (and therefore officially supported by) any major Linux distribution, one should always watch for potential security vulnerabilities and keep an eye on their mailing lists for possible updates and patches. {link to this story}

[Tue May 10 16:14:40 CDT 2005]

We sure read some interesting news today. First of all, Novell has announced that they are buying Linux vendor Immunix. Not many people know them, but the Immunix folks have been working on building a highly secure Linux distribution for years now. They were founded back in 1998 with some funding from none other than DARPA, and ever since they have been involved in some excitement development projects such as the Linux Security Modules. So, it sounds as an intelligent decision by Novell, especially since they are eyeing the high-end enterprise market thanks to their acquisition of SUSE. Of all the Linux vendors out there, it seems pretty clear that Red Hat and Novell are the only ones that are coming up with a nice and comprehensive strategy and line of products to attract the enterprise customer.

On the other hand, we also read today that Sun has bought Tarantella for a total of US $25 million in cash. I am not sure what to think of this other deal. What is Sun going to obtain from Tarantella? As far as I can remember, these guys were a part of SCO and now compete directly against Citrix, which is the most important vendor of thin-client software. {link to this story}

[Mon May 9 11:41:17 CDT 2005]

Ars Technica published an interesting history of the GUI by Jeremy Reimer that takes us all the way back to its prehistory (Vannevar Bush's musins about Memex, a device that would look like a desk with two touch screen graphical displays, a keyboard and a scanner) and its true first steps when Douglas Englebart performed a public demonstration in 1968 (the article includes a screenshot of the oN-Line System) to our days. All in all, it is a pretty good summary of the history of the GUI. {link to this story}

[Mon May 9 10:27:59 CDT 2005]

Just read about yet another interesting Mozilla extension. In this case, it is WebMailCompose, which allows you to automatically load your webmail's compose page when hitting any mailto link. The extension supports most services I can think of: GMail, Yahoo, Hotmail, Squirrel... {link to this story}

[Mon May 9 08:54:44 CDT 2005]

Well, it is Java's 10th birthday, and its school card is definitely a mixture of successes and missed opportunities. On the one hand, it has become quite predominant on the server side, to the point that a simple look at the job postings out there makes one feel as if getting into Java has some future. There is a whole branch of the technology industry making products both in Java and for the Java developers (just think of vendors such as IBM, BEA or Borland), and a thriving community that has given us, among other things, Eclipse. Today, there is very little doubt that Java is a mature and stable technology that can be used for key enterprise projects without sweating it. In addition, it is indeed playing a central role in the market for embedded devices. Yet, on the other hand, Java has not lived up to its own hype. Back in 1995, it promised a world of "write once, run everywhere" that would take over the desktop market and dislodge Microsoft's semi-monopoly. It never happened. The alliance between Sun and Netscape did not work as well as expected, Microsoft played a couple of its usual dirty tricks to water down the Java standards and, later on, came up with .NET, a clear copy of Java that only runs on its own operating system, as usual. True, Sun has benefitted very little from Java when it comes to its bottom line, at least in appearance, but as Scott McNealy points out:

Imagine we hadn't done Java 10 years ago. Where do you think Sun would be today? It would be all Windows. We'd be done. If people aren't writing Java Web services, they're writing to .Net. If they write to .Net, they write to Windows. If they write to Windows, they don't write to Sun equipment.

We all benefit from Java's existence. McNealy has a point. {link to this story}

[Sun May 8 16:36:30 CDT 2005]

While much has been written lately about Microsoft's final release of 64-bit Windows, the reality is that most companies are deeply invested in 32-bit versions of the operating system and it will take some time before they switch to 64-bits. Nevertheless, I think it is safe to assume that with the arrival of the 800 pound gorilla to the new architecture we will hear more and more of it in the future. Still, for all the excitement about the promises of 64-bit (better performance and, above all, a higher ceiling for the maximum amount of system memory addressable by the kernel), it is AMD that appears to be growing beyond Intel's worst nightmarish expectations. I read in ComputerWorld that Andy Bechtolsheim, one of the founders of Sun Microsystems, has returned to the company and has beeen working on a secret project code-named Galaxy to develop their next generation of serves based precisely on the AMD processor. Needless to say, if Sun were to abandon its SPARC processors in favor of AMD it would be big news. I do not think it is something we should expect anytime soon, but it seems clear to me that McNealy is already positioning the company for that event. The end of life for SPARC is on the cards. {link to this story}

[Sun May 8 16:18:50 CDT 2005]

ComputerWorld interviews John Markoff, author of What the Dormouse Said, an analysis of the roots of today's technology in the worls of the counterculture. Quite a few conservatives out there are shocked that the technology world is choke full with liberals, but it should not surprise us so much once we learn about the origins of the personal computer. After all, while the first computers were born in environments strictly controlled by the Government and were used mainly for military purposes, the freaks who invented the personal computer grew up in a completely different milieu: the acid trips and countercultural protests of the Bay area in the 1960s. In other words, to a great extent, the personal computer was invented as a libertarian reaction to the excessive control and rigidity of the computing world up until then. The people at the Homebrew Computer Club wanted the freedom to build and experiment with their own computers, regardless of whether the coporations at the time believed that the product made sense or not. {link to this story}

[Fri May 6 16:24:22 CDT 2005]

Matthew Szulik, Red Hat's CEO, has warned us against the dangers of software patens again. Way too many people out there feel that this is an issue only free software freaks in their foolish extremism worry about, but the reality is that software patents truly are a danger to the entrepeneurial spirit. As Szulik points out after explaining that Red Hat already has 17 full-time attorneys on their staff:

It's very possible for someone to work for three or four years developing a process only to turn around and find he faces a cease and desist order. It is an enormous threat to the entrepreneurial spirit.

{link to this story}

[Fri May 6 16:19:03 CDT 2005]

Dana Blankenhorn discusses in his blog the problems of coming up with a good open source business model, and ends up offering the Little Red Hen business model as a possible solution:

Companies that really want the software may pre-pay for support. Future support charges might be discounted to those who provide help. This can come in many forms, staff time, physical resources, even testing.

Those who grow the wheat and bake the bread should get to eat it first. I suspect many open source projects already work in this way, and it makes a lot of sense.

The idea is not bad, but I seriously doubt it will be the solution so many people have been searching for. {link to this story}

[Wed May 4 16:46:38 CDT 2005]

Debian Sarge has entered freeze!. However, to give you an idea of how screwed up Debian's release cycle is, by the time this new version releases (by the end of May, if we are all lucky enough), there will be a few distributions out there already running GNOME 2.10. Yes, I know, Debian prices stability and most people who run it do it on servers anyways. Still, it tells us quite a bit about its own problems. {link to this story}

[Wed May 4 10:20:00 CDT 2005]

It is nice to see that the mainstream technology media is keeping a close eye on what happens in the Linux world beyond the two or three major vendors. Today, eWeek published a short review of Ubuntu 5.04 with lots of praises for this new Debian based distribution.

Ubuntu is an excellent choice as a desktop or notebook operating system: It includes the latest GNOME and KDE versions, and it requires less post-install configuration than does Fedora Core or Debian on its own. For instance, we found that, unlike with Fedora, laptop hibernation worked out of the box.

{link to this story}

[Tue May 3 16:46:03 CDT 2005]

Scott Robert Ladd ran a few benchmarks comparing the performance of GCC 4.0 to GCC 3.4.3 on the Opteron and Pentium-4 platforms, and the results are not very good for the latest release of the GCC compiler, to be honest. Its performance is just so-so, and it does not look as if it could be recommended just yet. {link to this story}

[Tue May 3 15:43:49 CDT 2005]

MLAgazine publishes a short article on A/UX, Apple's little known older brother to their flashy Mac OS X. Few people know this, but there was a time back in the 1980s and early 1990s where UNIX was so well respected that even Microsoft released its own flavor of the operating system, called Xenix. In any case, the article I am referring to here contains a couple of screenshots of what looks like a nice example of traditional Mac GUI on top of an UNIX core, not so different from the approach they recently took with Mac OS X. A/UX could only run on a Macintosh II with 4 MB of RAM and an 80 MB hard drive, which at the time was a high-end system, believe it or not. Among other things, the OS supported NetWare, TCP/IP and AppleTalk, and it also emulated X Window to allow users to run UNIX GUI applications in a seemless manner. {link to this story}

[Tue May 3 09:59:26 CDT 2005]

It is just amazing how much we have advanced over the last few years in the level of awareness of mass media about the open source world. Not so long ago (well, a few years ago, definitely before 1998), few analysts out there would have heard of "free software" or "open source", not to talk about the distinction between "free beer" and "free speech". But things have changed, there is no doubt. Neil McAllister writes a piece in InfoWorld about Linus' snaffu with BitKeeper, concluding:

Torvalds seems to have fallen for the "free beer" argument: He didn't have to pay for BitKeeper, so he figured it was good enough. But not having to pay is not, and has never been, the real purpose of free software. The point is to avoid the situation Torvalds eventually found himself in: McVoy didn't like how his product was being used, so he took his ball and went home. Could you afford to switch gears in the middle of a project if one of your key software vendors did the same?

The business community likes to distance itself from the ideological debates surrounding free and open source software, but the BitKeeper case is a prime example of why enterprise IT management can't ignore software licensing issues. You don't want your PBX vendor telling you how to use your phone system, or your printer vendor telling you what to print. Wouldn't you prefer software that didn't tell you how to run your business either?

He is absolutely right. To me, this has always been one of the major selling points of open source software, even though too many vendors out there are just too afraid of using the argument because they consider it "too political". Listen, not all politics has to be considered dogmatic or extremist. In this case, I think the argument makes sense, and any business executive who thinks it through will see that it is a powerful reason to support open source software and avoid proprietary solutions like the plague. {link to this story}

[Tue May 3 09:53:33 CDT 2005]

The Inquirer published today an interesting piece on the window of opportunity that Linux currently enjoys to displace Windows from part of the desktop market.

A window of opportunity to displace Windows XP is based on three factors: the declining price of hardware, Microsoft's current pricing and philosophy on software licensing, and the general lack of innovation in the hardware arena. AMD and others are beating their chest about supplying $100 PCs to the downtrodden around the globe, yet First Worlders can get a $299 to $399 PC out of the local Big Box or via mail order that includes a $100 Microsoft Windows XP load "tax."

At hardware prices under $200 bucks, it becomes a no-brainer to throw in a Linux distribution onto the hard disk rather than having to mess with Microsoft OEM licensing. It may be left to the consumer that if they really want Windows XP, they'll have to pay and install it themselves.

The only problem I see with this argument is that it has been said before, and it did not happen. On the other hand, it is also true that prices have never fallen this low, and the lower the price of computers falls the higher the percentage represented by the Microsoft tax. There has to be some point where the OEMs rebel against the high cost of the OS. {link to this story}

[Mon May 2 14:22:56 CDT 2005]

Ed Dumbill tells us why he does not find GNOME programming fun anymore. It boils down to the old C versus Mono/Java debate:

It seems to me that if the GNOME project cares about existing in the same way it has in the past, it needs to swallow some bitter pills.

Forget about ideological purity at the base of the stack. Find ways to make Mono, Python and even Java work when mixed into the same developer platform. They have good docs and a strong following.

We don't need a grand unifying scheme. Mono and Python just need glue where it counts. Stuff like easy widget reuse is overrated. Really.

Forget also about bloat. It's really not the deal we think it is. Ask any OS X or Windows user. The mass of Mono plus Python is still small. Besides, any distributor is going to want both of these in anyway to get the best applications.

Forget about some grand project to completely replace everything in a big bang in however many years time. Instead, slide new technology in incrementally. The price of making new stuff work with old is worth paying. The alternative may well be that everyone's gone to sleep while we were busy making GNOME 3.

Maybe I'm wrong in this analysis, but it comes down to the fact that GNOME needs a critical mass again to move forward. If it doesn't get it soon, developers will carry on moving to where the fun is.

And, in conclusion, perhaps that's right. The future desktop will be probably forged elsewhere than in the GNOME project as it is now.

It is definitely starting to look as if the exclusive reliance on C is becoming a serious issue for the GNOME project. As Dumbill himself explains, the real innovative development is happening on the Novell and Canonical fronts and both vendors chose to bet it all on Mono and Python respectively. In other words, down to the metal C development may have become by now more of a nuissance than anything else, at least when it comes to applications development (needless to say, system development is a completely different thing). The adoption of a high-level scripting language may be necessary to revitalize the GNOME community but, as Dumbill states, we all may be better off if instead of "choosing a winner" we just allow a thousand flowers to blossom and then pick the most beautiful ones at a later stage, or at least that is how I read his piece. {link to this story}