[Mon May 29 18:56:14 CDT 2006]

Well, well, well. Isn't it amazing how things change in the business world when one needs to resort to the sort of tactic that was widely criticized and derided just a few months ago? The more I see how companier truly work, the more convinced I am that there's little difference between business and politics, and that includes all those on-the-fly change of course operatoins too. Take Jonathan Schwartz, for example. Just a few months ago, he was scolding Linux and the opensource model as severely flawed because it always ended in forks. Of course, now that Sun has decided to opensource Java things are completely different. If you don't believe, just check out the interview with Jonathan Schwartz published by Information Week:

— My company develops in Java. How does this change our lives?
— The big risk is if Java forks —a group takes a version of the language in an incompatible direction. This had been Sun's main argument against open source Java. In reality, few open source projects fork. Linux hasn't. Java's large installed base would be hostile to an incompatible version. More likely, the big change would be positive: rapid advancement in the language from many outside contributions, still subjected to Sun's battery of compatibility tests.

This is the same guy who didn't think twice before calling Red Hat "a fork" just a few months ago. Things definitely change, huh? {link to this story}

[Thu May 25 14:49:13 CDT 2006]

Ever found yourself in a situation where you can SSH out of a system to the outside network but not the other way around? Well, I had read about OpenSSH's capabilities to perform a reverse login tunneled through an existing connection, but until yesterday I had never had to use it. It's pretty easy. On the first system, you enter:

ssh -R 8022:localhost:22 user@remotehost

Enter the password for user on the remote host, then when you are sitting at the remote host simply enter the following:

ssh -p 8022 user@localhost

In this case, the user is the account you previously logged from on the first system. Simply enter its password, and you're all set. Easy, huh? Of course, be aware of all the potential security issues. {link to this story}

[Thu May 25 14:23:41 CDT 2006]

Taking into account the pessimism that transpired most analysis about e-commerce just a few years ago (I'm talking about the dot-com bust here, when suddenly everybody who two years before thought the Web was the best thing since sliced bread now thought it'd doom us all to a worldwide crash), it's interesting to read that online sales for the year 2006 are well under way to top US $200 billion.

Online sales, including travel-related sales, are expected to rise to $211 billion this year, up 20% from the $176.4 billion recorded in 2005, according to The 2006 State of Retailing Online, the ninght annual Shop.org study of 174 retailers conducted by Forrester Research Inc. Sales excluding travel are expected to reach $138 billion, according to the study.

The largest nontravel categories this year will include computer hardware and software, valued at $16.8 billion; automobiles and auto parts, worth $15.9 billion; and apparel, accesories and footwear, pegged at $13.8 billion, according to Shop.org. The fastest growing categories are expected to be pet supplies and cosmetics and fragrances, which are forecast to see growth rates of more than 30%, according to the study.

So much for the old Pet.com guys, huh? Of course, this does nothing but to confirm my suspicions that the Web has always been a great place to do business and make money, and that didn't change because of the dot-com bust at all. What happened is that way too many people who set up these companies way too early wasted money as if the end of the world was around the corner, instead of tightening the belt and working hard. Other than that, the business plans were sound, I think, and the facts are coming to confirm that they were right. {link to this story}

[Mon May 22 16:05:28 CDT 2006]

Now, I'm sick of reading stories in the mainstream media that start just like this article by Linux Insider on the Linux desktop:

As the speakers at the Desktop Linux Summit 2006 in San Diego last monght looked out at their audiences, they couldn't help but notice that the number of attendees sporting T-shirts, sandals and bushy facial hair —the stereotypical look of Linux movement diehards— was much reduced from previous conferences.

Exactly! It's stereotypical. Perhaps one of these days journalists will find a way to give up their tired cliché about Linux users "sporting T-shirts, sandals and bushy facial hair", but then... I dream too much. {link to this story}

[Mon May 22 15:23:27 CDT 2006]

Tectonic published a short piece about Nat Friedman, the programmer who set up Ximian together with Miguel de Icaza, that includes a funny anecdote:

"Starting the company was a hack", admits Friedman. "How to get customers, how to get money —that was all kind of fun. It was a game that we were playing; I found that pretty easy". Running a company as a hack worked —HelixCode secured $2 million in 2000, despite disillusionment in the IT sector with start-ups— and secured a total of $15 million venture capital by the end of the year. Friedman recalls going to the ATM after the $2 million was deposited and getting a statement. "It ran off the edge of the page!" He then tried to transfer the $2 million from one account to another through the ATM, which the system couldn't handle. Eventually —with a good queue growing behind him— Friedman moved the money in $500,000 increments. A week later they hired a CFO, who banned them from moving millions of dollars around through the ATM.

That is some picture! {link to this story}

[Mon May 22 08:51:52 CDT 2006]

bblazer asks for advice in Slashdot on whether students should be taught with or without an IDE:

Beginning this next school year, there is a strong possibility I will be teachign an intro to Java and an intro to Python course at the local community college. I was wondering what the prevailing wisdom is when it comes to teaching languages —should students be taught with or without an IDE? I am a bit old school and wouldn't mind having them all use vi or emacs, but using a good IDE does have some advantages as well. I should note that the students I will be teaching will have had at least 1 semester of programming in VB or C++.

Well, for starters, it does not have to be vi or emacs. Why not use gedit, for example? I mean, there is such a thing as GUI text editors out there. The alternative to a fully fledged IDE does not have to be an old-style cryptic line editor (mind you, I am a vim fan, so spare the flames, please). In any case, and more to the point, I have taken a few programming classes in several community colleges here in Minnesota (the usual suspects: C, C++, Java, etc.), and there is little doubt in my mind that people who are taking an intro to any language benefit the most by concentrating on the language itself, and not on the quirks of this or that particular IDE. I distinctly remember several occasions where the teacher had to spend a good amount of time troubleshooting bugs and/or issues with whichever IDE people were using in class while I moved on with my Linux laptop, vim and GCC. So, I would definitely recommend to stay away from any particular IDE, at least during introductory classes. {link to this story}

[Fri May 19 15:10:33 CDT 2006]

Mark Hall writes a piece in ComputerWorld where he wonders if UNIX is dead, not so much because Microsoft killed it but rather because Linux is undermining it day after day. Some of his thoughts are quite interesting:

A look at the Top500 list of supercomputers tells the tale best. In 1998, Unix machines from Sun and SGI combined for 46% of the 500 fastest computers in the world. Linux accounted for one (0.2%). In 2005, Sun had 0.8% —or four systems— and SGI had 3.6%, while 72% of the Top500 ran Linux. IBM saw its prominence rise to the point where its systems represented 44% of that prestigious list in 2005, ip from 21% in 1998, all because of its investment in high-end Linux systems.

Linux's success in high-end, scientific and technical computing, like Unix's before it, preceded its success in your data center. Once Linux proved itself by executing the most complex calculations possible, IT managers quickly grasped that it could easily serve Web pages and run payroll. Naturally, it helps to be lucky: Free, downloadable Linux's star began to rise during one of the longest downturns in IT history. With companies doing more with less, one thing they could dump was Unix.

There is what many people from the old guard still refer to as "a toy OS". To me, it looks more like "the UNIX killer". {link to this story}

[Fri May 19 10:55:32 CDT 2006]

eWeek published today that Dell is about to launch a four-processor server product based on the Opteron chip, which would definitely be big news taking into account that Dell has always been solidly on the Intel side since its very inception. But the money is the money, of course, and nobody can blame Dell for hedging its bets, especially when AMD has been gaining on Intel for the past couple of years or so. Customers demand it. Period. {link to this story}

[Wed May 17 10:08:42 CDT 2006]

Well, here is another interesting tidbit of information. Sun has decided to open source Java and release it under the Operating System Distributor's License for Java (als known as the Distro License for Java or DLJ), which will now make it possible for all Linux distributions to include Sun's implementation of Java with their software. Sure, nothing ever stopped people from downloading a free legal copy of the Java SDK and installing it on a Linux system, but while it was necessary to jump over some hurdles before that will not be needed any more in the very near future. This is good news, I think. {link to this story}

[Wed May 17 09:48:15 CDT 2006]

Hmmm, this is interesting. CNet News publishes a piece today titled Sun flits with Ubuntu where they tell us about both Jonathan Schwartz and Mark Shuttleworth appearing on stage during the recent JavaOne keynote session and exchanging a few niceties:

"Ubuntu is gaining a ton of momentum", Schwartz said in a meeting with reporters after his keynote presentation. "It is arguably one of the most important —if not the most important— Linux distro out there". That's a poke in the eye for Red Hat and Novell, the other two major Linux distributors.

So, what is going on? It is far from clear, but there are rumors of an Ubuntu-Sun alliance already.

Ubuntu hopes to make a splash in the enterprise space with the Dapper Drake release of its operating system, slated for June 1. The new version will run on 32-bit and 64-bit Intel-compatible systems, PowerPC systems and a fourth, yet-to-be-announced architecture, Shuttleworth said. Could this be Sun's Sparc?

{link to this story}

[Thu May 11 15:44:49 CDT 2006]

It is amazing how times change. Not so long ago, BitTorrent was considered just a piracy tool, a peer-to-peer technology used to share files on the Net and download movies illegally. Now CNet News tells us BitTorrent and Warner Bros. Entertainment have reached an agreement to distribute films and TV shows starting sometime this coming summer. I suppose there is a lesson to be learned here by all those who are too prone to attack a given technology as anti-American and Communist, obviously ignoring that a given technology can have multiple purposes.

Incidentally, this week I also came across another interesting project I have to check out one of these days. It is called Democracy TV, an application that makes it easy for users to tune into TV channels over the Internet. You just subscribe to channels using its built-in Channel Guide feature, and view away. Obviousl, I would not expect to see many large broadcast networks there but it sounds like an awesome tool to aggregate indepedent channels. {link to this story}

[Thu May 11 15:24:46 CDT 2006]

Since I wrote about SGI's troubles yesterday, it is just fitting that we spend some time now on the Itanium fiasco. Steven Vaughan-Nichols wonders whether it is Linux that needs Itanium or rather the other way around, and his musings do make a lot of sense, to tell the truth.

SGI bet the farm that Linux and Itanium on systems like the SGI Prism would become the darling of high-end Linux users. Silicon Graphics just went bankrupt.

"SGI was the most aggressive vendor in term of enabling the Itanium chip for highest performance and they were sort of the poster boy for HPC (high performance computing) for a long time", noted a buddy of mine who works for a competitor.

He thinks "SGI's strategy to support the Intel Itanium architecture, as opposed to AMD or a combination, contributed heavily to their demise. HP might very well have been in the same boat if they hadn't started selling AMD Opteron servers two years ago".

(...)

It's not that the latest Itanium chips haven't been good. They are. The Itanium family really hit its stride with the Itanium II line —but it has been too little, too late, and, what I hear over and over again: it's too expensive.

I keep hearing this not just from AMD fans —you know who you are— but even from people who can't wait to get their hands on Intel's next desktop chip, the Conroe, which is due out later this year.

It's not like Linux is going to leave the Itanium behind. Both Red Hat and SUSE support it with their latest server offerings. I'm sure they'll do so with their next versions due out this year. But, Linux really doesn't need Itanium to be a success on servers. It's the other way around.

If the makers of Itanium-based servers can manage to put together Linux server packages that aren't just cheaper than Solaris/SPARC systems, but are at least price competitive with Linux/Opteron and, for that matter, Linux/Xeon systems, then they can finally make some serious money from Itanium.

Yes, you got that right. That will not be easy. By the way, if anyone needed a good example to figure out how long Linux has come, this is an excellent one, for Vaughan-Nichols is absolutely right: it is Intel that needs Linux, and not the other way around. Who would have said just a few years ago, right? {link to this story}

[Wed May 10 09:12:54 CDT 2006]

Well, it had to happen sooner or later. We all knew that. A couple of days ago, SGI filed for Chapter 11 protection. Most people thought it would still take 6 more months or so, and in that sense it caught us by surprise. Other than that, it was fully expected. Revenues have been dwindling for quite some time now, and the company strategy has just been limited to laying people off every few quarters for years and years. Few people doubt that Bob Bishop, the previous CEO, and his team are to blame for the company's problems today. For more than seven years, they did little but repeat the same old mantra that their strategy (i.e., limiting SGI's products to the high-performance scientific and research market, and only the very high-end at that) would save us if we just had enough faith. Heck, you do not need a PhD to realize that it was not working when the company failed to make any money whatsoever for a single quarter in more than seven years. Anybody could tell there was something wrong with the strategy, without any need to attend business school. Yet, Bishop continued down his path to nowhere.

So, what is going to happen now? Who knows? All I can say is that, for the time being, we do appear to have an executive team in charge of the company that has an idea where they want to go. It may not sound like much, but it is far better than what we have had for years at SGI. People are feeling OK about it all, and we all think it is definitely not bad that our debt has been wiped out. Yes, so were SGI stocks, and employees lost a lot of money there. Still, the new team appears to be honest and it has a sense of direction. The way things are, and taking into account what we had before, this is great news. {link to this story}