[Tue Apr 27 21:58:11 CDT 2004]

There are times when I do not really know if Sun is a technology company or an opposition movement. It nearly seems as if they cannot live without picking on some other company, as if they cannot build their own identity unless it is in opposition to somebody else. It used to be that Scott McNealy could not climb on top of a stage in any computer fair without uttering a few outrageous criticisms against Microsoft. Now, after their sudden detente, they are picking on Red Hat. Today, Jonathan Schwartz, proud successor of Mr. McNealy in his clownish ambitions, accused Red Hat of being proprietary and having forked the Linux kernel. I am not even going to bother to discuss the comment. Suffice to say that every single piece of software released by Red Hat is out there, freely downloadable at least in source, and there are several companies and projects out there that simply compile Red Hat's source RPMs (or patched kernel) to include them in their own products. {link to this story}

[Tue Apr 27 13:05:52 CDT 2004]

In these days of freely accessible office suites (like OpenOffice) and IDEs, I must acknowledge I still stick to good old vi (well, vim, to be precise). Sure, it does not offer a barrage of features (well, that is subject to discussion) or a nice and user-friendly GUI but, on the other hand, it is small, fast and it usually is installed by default in all UNIX and UNIX-like systems. If you would like to learn more about this editor, check out a long article on vim published by Linux Productivity Magazine, which provides some nice pointers on the most important commands, features, maps, abbreviations, addons... to help you quickly become a power user. By the way, once you are there, check out the other monothematic articles Linux Productivity Magazine has published on other topics (Spamassassin, Tripwire, iptables, Perl, Gimp, LyX...). It is a very good but mainly unknown resource. {link to this story}

[Mon Apr 26 09:20:33 CDT 2004]

Miguel de Icaza publishes a good article on Longhorn's threat to the Linux desktop that is well worth reading, as most of what he writes. Among many other things, Miguel states:

... what makes Longhorn dangerous for the viability of Linux on the desktop is that the combination of Microsoft deployment power, XAML, Avalon and .NET is killer. It is what Java wanted to do with the Web, but with the channel to deploy it and the lessons learned from Java mistakes. The combination means that Longhorn apps get the web-like deployment benefits: develop centrally, deploy centrally, and safely access any content with your browser. The sandboxed execution in .NET means that you can visit any web site and run local rich applications as oppposed to web applications without fearing about your data security: spyware, trojans and what have you. Avalon means also that these new "Web" applications can visually integrate with your OS, that can use native dialogs can use the functionality in the OS (like the local contact picker). And building fat-clients is arguably easier than building good looking, integrated, secure web applications (notice: applications, not static web pages). And finally, Longhorn will get deployed, XAML/Avalon applications will be written, and people will consume them. The worst bit: people will expect their desktop to be able to access these "rich" sites. With 90% market share, it seems doable.

Also, Miguel reminds to all those who want to argue the merits of the GTK and QT toolkits that:

Although some of the discussion has centered around using a native toolkit like Gtk+/XUL to build a competitor that would have ISV sex-appeal, this is not a good foundation as it wont give us Web-like deployment (we need a stack that can be secured to run untrusted applications, and we need to be able to verify the code that is downloaded, which leaves us with Java or .NET). The time is short, Microsoft will ship Avalon in 2-3 years, and they got a preview of the technology out. I see two possible options:

• Implement Avalon/XAML and ship it with Linux (with Mono)
• Come up with our own, competitive stack

{link to this story}

[Sun Apr 25 17:15:06 CDT 2004]

Just like politics, our field is full of FUD, gossip and conspiracy theories. One of many crazy rumors I head of recently (and that caught my attention because I work for SGI and have friends at Sun) was that of Sun adopting Itanium through the purchase of SGI. The rationale is that sooner or later Sun will have to adopt Itanium or face death and, taking into account that they are coming a little bit late to the party, their only choice is to purchase a company that already has the expertise designing systems around the Itanium chip. As the author clarifies,

Sun has only one choice available; it can acquire the expertise needed for designing high performance Itanium based systems. Sun can't afford to pay too much money for an acquisition; the company's financial position does not allow that. This leaves just one attractive acquisition target available, and that acquisition target is SGI.

SGI has a history very much reminiscent of Sun's. SGI too was once a high flyer in the high performance computing scene, and was laid low by competition from cheaper Intel based systems. Like Sun, SGI too had its own custom OS, and its own line of custom processors. In a fairly painful reorganization SGI got rid of its non-profitable assets, and has now positioned itself as a high performance computing vendor centered around Itanium, and Linux. The company seems to have turned the corner, and management is claiming that SGI is back on track.

SGI has excellent technology, and the Itanium expertise Sun needs desperately. But, for Sun the most prized part of an SGI acquisition would be SGI's management. SGI has gone through what Sun is going through right now, and SGI's management understands the steps needed to deal with Sun's situation; CEO Robert Bishop has done an excellent job managing a company teetering on the edge of bankruptcy. On the other hand, Sun is still confused about the pieces it needs to ditch, and the ones it needs to keep. The correct move for Sun would be to acquire SGI, and appoint Robert Bishop as Sun's CEO. Sun has an atrocious record of acquisitions, and in case of any major acquisition Sun's current management has to go before there is any chance for the acquisition to be a success.

Now, there is little doubt in my mind that Sun truly is in deep trouble, and that their obsession with the SPARC line of products is what will ultimately lead them to a dead-end. Sun management has been slow to react to both the Intel threat and, more recently, Linux and the opensource movement. They have given us a few nice technologies and open standards, but failed to see those two threats building up and coming down on them. Will they purchase SGI? I doubt it. Would that save them? I also doubt this. Why? Sun still has to start taking its own problems seriously. It still has to start reinventing itself and laying down a clear vision of what they want to do in the future. I have always been convinced that Solaris will be the last UNIX elephant to die away in the face of the unstoppable Linux attack (yes, AIX and HP-UX are still there, but limited to such small markets that they barely count anymore, and only survive because they are part of much larger companies that make money with other products), but it will still die sooner or later. Well, unless they manage to reinvent themselves, this is the beginning of the end for Sun. As for SGI, I think a merger/purchase with Novell sounds much more exciting and also has more future. Not only do they make money, but after Novell bought Ximian and SuSE they are starting to offer a nice and comprehensive product line that definitely poses a serious threat to Sun and Microsoft both in the mid-range server market and the business desktop front. What is missing from Novell's new strategy? Precisely what SGI has: expertise in the high-end performance computing, Itanium and hardware. Also, unlike in the case of Sun, there is little or no overlap between the two. Now, that would be some compelling story! {link to this story}

[Sun Apr 25 16:57:16 CDT 2004]

SuSE's CTO, Jürgen Geck, started a controvery in the Linux world when he criticized Red Hat in a speech in Toronto for the practice of backporting features from the 2.6 kernel into the 2.4 kernel bundled with their Linux distribution. His main argument, which I consider far from simplistic or stupid, is that by promoting such practices, we interfere with the process of standardization of the operating system. There is very little doubt in my mind that Geck is right to some extent: it is not the same to run an application on a Red Hat kernel with NPTL compiled in than running it on a different kernel without that functionality backported into it. I can see companies taking all this into consideration when making purchasing decisions and, as Geck points out, I can also see the confusion it generates. However, as Linus Torvalds has argued,

I think it makes sense from a company standpoint to basically "cherry-pick" stuff from the development version that they feel is important to their customers. And in that sense I think the back-porting is actually a very good thing.

Why? Well, it is precisely this type of flexibility that makes Linux and opensource powerful: both the customer and the distributors have the power to customize the software to their heart's content. Let us not destroy the hen that lays the golden eggs in the name of some misunderstood egalitarianism. {link to this story}

[Sun Apr 25 16:13:57 CDT 2004]

If you did not hear talk about Python in the last few years, there is a good chance you are not working in this field (or are irreversibly out of touch, which is way worse). Open Enterprise Trends publishes an interview with Alex Martelli about Python in the enterprise. Martelli is the author of a couple of best selling O'Reilly books on the topic: Python in a Nutshell and Python Cookbook. So, what is so important about Python? I just dabbled in it for a bit, but it has been enough to realize that the language is powerful without at the same time being complex. The only two things stopping me from using it more often are: the fact that, although it already has a good amount of libraries, it is nowhere close to what the Perl community has built over the years in the form of its famous Perl modules; and the fact that I am still an old procedural type of guy (more on that below). Nevertheless, Python is consistent, clean and simple. How simple? Martelli illustrates the power of the language with the following example: say you have to find out whether a certain string (we call it needle here) is part of another larger string (we will refer to it as haystack). Here is how you do it in Python:

if needle in haystack:
print 'found'
else:
print 'not found'


And that is it! Simple, right? Of course, things get far more complex as one tries to accomplish something actually useful, but it usually is the case that other programming languages make it extremely difficult to do something as simple as what we did with the four lines of code listed above. On top of that, Python is quite clean and has a sound design. As Martelli explains,

Powerful tools such as regular expressions play a completely different role than language fundamentals such as assignment, argument passing, raising and handling exceptions, returning values from functions. Python includes in the language proper only those concepts and functionalities that really belong there (...). In Python (much like in C, C++, Java and .NET), everything that can most reasonably be supplied via a library module is kept in a library module, always available for import as and when needed, but not complicating the language itself needlessly.

So, why not use it most of the time? As I explained above, I started writing some simple programs back in the days of BASIC, and am still too used to the procedural way of programming. Yes, I do see the advantages of an object-oriented approach. Yes, I do understand it. Still, for whatever reason, I keep thinking of my programming challenges in a procedural way. I suppose it is not easy to change an old dog's ways. "So? Why not write in Python as if it were a procedural language?", you may ask. Well, it is not only sort of a waste to do that, but I have to disagree with Martelli when he states that "Pyton offers the programmer object-orientation not as a hurdle to be negotiated, but rather as one more useful tool to get the job done". Believe me, if you are not already familiar with the object-oriented approach to programming, you will find it difficult to write in Python, at least once you get past the simplest test programs.

In any case, the interview to Martelli is well worth reading as an introductory piece to the Python language. He also talks about the Twisted Matrix framework to write networked applications, and which includes implementations of several commonly used networking services such as a web server, an IRC chat server, a mail server, a relational database interface and an object broker. I had never heard of it before, but it sounds like a quick way to put together stable and scalable networked applications in Python. He also gives some advise on GUI toolkits for Python:

The GUI toolkit that comes with Python, Tkinter, which I cover in Python in a Nutshell, is widely used because it's widespread and quite easy to use, but I wouldn't recommend it for commercial-quality applications, because I think wxPython and PyQT are better.

{link to this story}

[Wed Apr 21 14:51:58 CDT 2004]

The Linux desktop front continues to move and it is making quite a bit of noise lately. ZDNet publishes that Novell plans to switch more than 5,000 employees to Linux and OpenOffice in the next 12 months, while a company called SpecOpS Labs is already hyping its Project David which aims at building a Windows compatibility layer that runs on top of Linux and enables "all major Microsoft Windows applications to run on the free and open source Linux OS". Needless to say, we know nothing about this latter product. We do not know how much it will cost, whether it actually works or if it performs like molasses. My only point is that there is some new interest and impetus in the Linux desktop as an alternative for large as well as medium-sized businesses. If you add to that the nice revies GNOME 2.6 is already getting and the Mozilla browser is gaining more and more adepts, we could be witnessing some fascinating competition in the field in the next two years or so. {link to this story}

[Tue Apr 20 10:59:36 CDT 2004]

O'Reilly's Linux DevCenter publishes a good introductory article to XMMS. The piece does not break any grounds, but it serves its purpose well: to provide a good introduction to the uninitiated on how to use the different controls the application provides, some of its most impoerant features, how to load different skins, save playlists, etc. {link to this story}

[Tue Apr 20 10:53:07 CDT 2004]

Well, it seems that perhaps Netscape is not as dead as we thought after all, at least formally speaking. AOL has just announced a new version of the Netscape browser, which will be released as version 7.1, together with a new product called the Netscape Desktop Navigator that so far only runs in Windows and provides localized access news, weather information, TV schedules, yellow and white pages, personals, etc. In other words, it is nothing ground-breaking, if it were not because everybody considered Netscape completely dead since AOL laid off most of its developers recently. So, does it matter anymore? I doubt it, and to make matters worse it does not even look as if AOL has any serious plans for the future in this field. If you ask me, it is quite of an enigma why they may have decided to release this product precisely now. {link to this story}

[Thu Apr 15 20:44:10 CDT 2004]

From time to time one hears about the promises of the Linux desktop. I already wrote about this some time ago, and I am afraid that things did not show so much during this time. Still, I read an interview with Nat Friedman on the risings expectations over the Linux desktop that contains a few insightful comments:

I'm always asked when is the Linux desktop going to be ready for everyone to use? The reality is, there is not going to be a single moment when a set of technology problems are solved. There's going to be gradual adoption like we've seen with servers.

(...)

A lot is happening outside the U.S. where there is not as big an installed Windows base.

(...)

Application availability is the No. 1 barrier. If an application does not exist, enterprises are not likely to move to Linux.

(...)

We're developing a Windows migration program to make it easier to move to the Linux desktop with training and documentation and migration tools that automate tasks.

(...)

The No. 1 misconception is that usability is a major barrier to adoption and that's not true. It used to be. There was a study done recently with a group of 20 users who had never used a computer before. Ten were put at a Windows PC, 10 at a Linux PC and they were given a list of simple tasks like sending an e-mail, surfing to a Web page and the usability results were pretty much the same. The real problem is getting your work done if the applications don't exist.

(...)

Another misconception is that no innovation takes place on the desktop and that we're just copying Apple and Microsoft. That's not true. People just don't notice it because it's a small market. For example, there's been a bug-reporting feature [in Linux] since GNOME 97. Microsoft is just doing it now.

It is certainly too early to tell, but perhaps the combined effort of Red Hat, IBM and the new Novell (after the purchase of SuSE and Ximian) will put Linux in a position where it will be able to compete against the Microsoft behemoth. If it happens, which is far from clear, there is little doubt in my mind that it will start only in a few limited areas (business desktop, point of sale, kiosks, developing nations...) and it will spread from there onto the home market. {link to this story}

[Mon Apr 12 14:36:42 CDT 2004]

The company may be in financial trouble, but something tells me that as long as Scott McNealy is at the driving wheel of Sun Microsystems the company will continue to make it to the headlines every other day. It just appears to be in his nature to show off and occupy the center stage. Yahoo! News has a commonsensical piece on the agreement Sun and Microsoft reached a couple of weeks ago, avoiding the anti-Linux nonsense we heard in other quarters (surprisingly enough, we even heard some of the so-called industry analysts come up with this sort of conspiracy theories). John Fowler, Sun's Chief Technology Officer for software, has clarified that

The agreement with Microsoft is about interoperability and choice for our customers, not about battling Linux or open source. (...) Sun recognizes that mixed environments are a reality and Sun is investing to make customers successful with Sun products in a mixed environment. (...) To portray Sun as anti-Linux and anti-open source is plain and simply wrong. In fact, outside of the University of California at Berkeley, Sun is the largest contributor of open-source code on the planet, and will continue to contribute to the community. Sun is as committed to the open-source community as ever and will continue to provide engineering support for Apache, Mozilla, Gnome, Grid, OpenOffice, JXTA, OSDL, etc.

It is not easy to measure who contributes more to opensource, but one has to acknowledge that Sun has usually played by the book and their products are based, for the most part, on impeccably open standards. In any case, The Economist does not appear to have a very positive view of the agreement and calls it "a desperate move" on behalf of Sun without thinking twice. It is too early to tell, but it certainly smacks of opportunism, especially coming from Microsoft's most entrenched enemy over the years. {link to this story}

[Wed Apr 7 15:14:48 CDT 2004]

OSNews published an article on Enterprise Linux distributions that, in my view, does nothing but to promote the idea that quite a few sysadmins out there should perhaps switch to Debian instead of ponying up the money that companies like Red Hat are starting to demand. The author of the article talks very briefly about the enterprise products from Red Hat, SuSE and Mandrake, to sum up with the following:

So basically, the concept of an enterprise linux offering is that you have a fairly long time between releases (12-24 months) and a product that remains supported for an even longer period of time than the release cycle. This allows for stable server deployments with guaranteed bug fixes and security updates for an extend period of time (up to 5 years).

Now, let us think about the price of these products. The cheapest prices mentioned in that very same article are: US $1499 for RHEL, US $999 for SLES, and US $799.90 for Mandrake Linux Corporate Server. If all we get in exchange for those prices is a more stable server deployment with longer release cycles, why not run Debian stable for free? The community-based Linux distribution offers precisely the very same advantages when compared to Fedora and other distributions. About the only drawback I can see is that not many enterprise software applications (especially the likes of Oracle or DB2) are certified on Debian, but that is something that may change (and could actually change if its market grows enough). In any case, for all those running Linux on infrastructure servers (web, email, DNS, file and printing...) Debian makes as much sense as any of these enterprise editions and saves them money. {link to this story}

[Tue Apr 6 16:10:18 CDT 2004]

It turns out the deal between Sun and Microsoft does not appear to be generating so much excitement among developers. For the time being, all I can say is that Scott McNealy and Steve Ballmer's statements appear to be far-fetched in the sense that there are little details to back up the assumption that the agreement between the two companies is about to herald a new beginning for the technology field as they want us to believe. I just fail to understand what the hoopla is all about, at least based on the information both companies have made public so far. For the time being, all we heard is talk about cooperation between Sun and Microsoft limited only to the Java Virtual Machine that used to be bundled with the Windows OS and, perhaps, an effort to guarantee interoperability in the field of web services. In other words, we have just heard hype, but nothing substantial. If anybody asks me, I would dare to say it nearly sounds as if this is nothing but an effort on behalf of Sun to still make it look as if they matter at all in the industry. {link to this story}

[Tue Apr 6 09:17:40 CDT 2004]

Forbes publishes an article about the growing caution in the industry regarding Linux deployments that is well worth reflecting about. Most people would automatically think of the SCO lawsuit as perhaps the main issue that could cause skepticism, but the article refers to another issue that I consider far more important the rising cost of deploying Linux. Red Hat's decision to discontinue its low-end distribution and center all future development on its enterprise products is already having an impact on the overall Linux market because, let us face it, Red Hat is not only the leading vendor in this field but also the symbolic banner of Linux in the corporate world. I will be the first one to admit that there are many reasons to deploy Linux that have absolutely nothing to do with the free as in beer concept, but at the same time it seems to me that the Linux and opensource communities have fooled themselves to some extent in their refusal to understand that money was truly one of the main reasons driving Linux adoption in the business world. Now that the cost advantage has pretty much disappeared, I do not think it is so clear that Linux will win hands down. Sure, it still makes much more sense to run Linux on home servers as well as small businesses, especially since using Fedora, Debian or Slackware are not an issue there. However, software certification and availability quickly becomes an issue when it comes to larger businesses, and most applications out there are usually certified only against Red Hat more due to strategic reasons than anything else. In this sense, I liked an article written by Jim Rapoza in eWeek where he stresses how the decision to certify applications such as Oracle against a single Linux distribution is purely arbitrary. One could always argue there is a cost associated to the tests needed for certification, but it is not less true that at least up until now the configuration differences between Linux distributions are minimal and cannot be used as an excuse. {link to this story}

[Sun Apr 4 17:10:02 CDT 2004]

A couple of weeks ago, I wrote about Eric Raymond's suggestion that Java should be opensourced, and mentioned that Scott McNealy was perhaps the biggest obstacle to the idea. Well, Scott McNealy has already announced that he opposes the idea. He repeated his well known position that Java is an open standard that software manufacturers can license in order to build products that compete with Sun's own. According to McNealy, the whole thing boils down to one issue: "trying to understand what problem does [opensourcing] solve that is not already solved". I just do not feel like coming back to the same issue again, but Eric Raymond already explains it very well in his article. For whatever reason McNealy is choosing to do what he does best: ignoring the arguments and repeating the same old mantra again. It is just fine, but he is definitely not addressing any of Raymond's points, which is precisely what he should be doing. Oh, well, I suppose that should only worry whoever has anything invested in Sun. To be fair, there is an argument used from the Sun camp that does make sense: an opensource Java runs the risk of being forked, thus breaking portability and compatibility, which happens to be a major feature of the platform. The problem is that we are currently witnessing something like this without the need to opensource Java: Eclipe and Swing are going different ways; BEA, IBM and other companies are building products that are not fully compatible with the competitor's, etc. In the meantime, we also heard last week that Sun reached an agreement to settle its legal disputes with Microsoft. In exchange for US $1.6 billion, Sun drops all its charges against Microsoft and reaches an agreement with them to pay royalties for each other technologies. It is hardly revolutionary, except that it points out the financial troubles Sun appears to be experiencing when they are willing to sit down and settle a legal dispute with their archnemesis. But the reason why I brought it up here is because with this new agreement Sun allows Microsoft to continue support for their Microsoft Java Virtual Machine product. I am afraid I need some clarification here: was it not McNealy himself who said over and over again that this was unacceptable because it broke Java's portability and compatibility? So, what is so different now, except for the considerable amount of money that is moving from one pocket to another? I am sorry, but it seems to me that McNealy, as usual, is speaking on both sides of his mouth: on the one hand, explains that outsourcing Java is impossible because it would break its portability and compatibility, but on the other hand he reaches an agreement with Microsoft to do just that. What is it, Mr. McNealy? {link to this story}