stupid day: displacing blame

Stupid Day: Displacing Blame – Oracle’s story was one of the more amazing of the past week. Oracle announced terrible earnings, when they had been predicting hearty earnings to the analysts just a week before. Larry Ellison cashed out nearly a billion dollars of Oracle stock before realizing that their earnings forecasts were fantasy. Maybe this seems a bit suspicious, but he did cash that stock nearly two months before the earnings surprise. More suspicious is the idea that Oracle thought everything was just fine until a few short days before earnings reports, when all of the company’s customers (allegedly) suddenly cancelled all of their orders. Is it really possible to be off by that much money and be blissfully ignorant just a week before earnings reports are due? Even that is possible, considering that Ellison has been distracted by efforts to diversify the company’s competetive strategies. The strangest thing here is that a disappointed gambler is suing the horse that they bet on. Since when has “Caveat Emptor” not applied to the stock markets?

stupid day: gambling with borrowed cash

Stupid Day: Gambling with Borrowed Cash – Many day-traders admit that daily stock-trading is similar to gambling. There is nothing illegal or even distasteful about speculating with your own money. But when the CEO of an otherwise stellar company gambles away his shareholder’s cash on the Internet bubble, “stupid” seems appropriate. Compaq has unquestionably been successful. They have Sun and IBM running scared, and are competing ferociously in both the Unix and WIndows hardware markets. They make great hardware. In fact, they make the most popular PocketPC on the market. Too bad they still haven’t been able to come close to meeting demand for the devices, even after having a year to get their manufacturing act together. Most companies would put top priority on fixing their production problems if they were turning away sales and delaying orders for many months at a time. But it’s hard to focus on just hardware when you have visions of Internet riches dancing in your head. A few months ago, Compaq announced that they had lost $1.8 billion due to losses on Internet stock investments!!! Without that “expense”, their profits would have been 500% higher. Or they might have been able to spend some more money fixing their iPAQ factories… I’m just some programmer guy, so I might have no idea what I am talking about; but doesn’t it seem a bit strange for a hardware company to be speculating in Internet stocks?

stupid day: gratuitous linkage

Stupid Day: Gratuitous Linkage – Do online newspapers exist to inform or to attract eyeballs? If you want to attract eyeballs, you can try to associate something slightly newsworthy with something that catches people’s fancy. For example, reports that the media-darling Miguel de Cervantes yi Icaza (or is it Saavedra?) and his equally hype-worthy Simian project for Linux will be working to make SOAP available for Linux. The article pitches this as yet another revolutionary thing; implying that SOAP was a Microsoft-only thing that requires some sort of populist Zapatistan to bring it “to the people”. They miss the fact that the very first implementation of SOAP was for Linux, as part of the Apache project! Not only was Apache the first SOAP implementation; it was the first production-quality implementation, and is still one of the best. This single, simple fact makes the whole article pointless. And explain to me again how Gnome and SOAP are tightly-related technologies? And what is Icaza doing for SOAP again? Is this an example of an open-source media-moth whoring for attention by aligning with the popularity of SOAP, or is it a cynical progeny of Jai Singh at counting on the click-throughs? You clicked on it, didn’t you? Shameless!

stupid day: software declarations

Stupid Day: Software Declarations – reports that IBM makes declaration of independence for chips, software. This sure sounds amazing, let’s check it out! Hmm, IBM will continue to compile their software on a chip so they don’t have to GPL it. IBM is has software and hardware that tunes itself (who doesn’t?). There may be something stunningly innovative going on at IBM, but this article sure doesn’t cover that. The article says absolutely nothing! Of course, CNET has announced their earnings estimates are reduced, and we all know that things like “declaration of independence”, “revolutionary technology”, and “manifestos” tend to generate clicks. You don’t actually have to say anything meaningful, so long as you quote Garibaldi or something, people will want to check it out.

stupid day: internet bubble

Stupid Day: Internet Bubble – Today we report on stupid things. The news recently has been talking about the tech market quake, with examples being Intel cutting 5,000 jobs and Cisco cutting 8,000. The latest reporting trend is to talk about who’s to blame. Satirewire seems to think that the problems are either due to stupid management or stupid people fooled by stupid reporters. As the investigation into the cause of these past two years of madness continues, I am reminded of the story The Master and Margarita, by Mikhail Bulgakov. I will not spoil the story for you, save to say that madness descends on Moscow, and in the books’s epilogue, we are given a glimpse of how the Moscow police attempt to explain the events: “Educated and cultured people, of course, took no part in all this gossip about evil spirits descending on Moscow, and even laughed at those who did, and tried to bring them to reason. But facts, as they say, are
facts and they could not be brushed aside without some explanation : someone had come to Moscow. The few charred cinders which were all that was left of Griboyedov, and much more besides, were eloquent proof of it. Cultured people took the viewpoint of the police : a gang of brilliantly skilful hypnotists and ventriloquists had been at work.”

mark mceahern

Mark McEahern – wrote a brute-force program to solve the ‘base n’ problem. The answers tantalize with the thought that there must be something useful there. 7 in base(3) is 21; and 7 * 3 is 21. Funny how most libraries convert only up to base 36, since that allows alphanumeric to be used as digits. Obviously, it would be rather hard to come up with character representations for base(257) without going to multibyte character sets. But the nice thing about this problem is that the only numbers you care about are the ones that use only the first ten digits of whatever base system they are represented in. So base 10,000 is perfectly valid. This also suggests a way to prove that more factors could not exist, or at least that they become increasingly far separated. The nested loop that Mark is doing gives an O(n2) cost. Well, it got 7 and 3 quickly enough! (Update: this is part of my Smart People category, and is not a part of “Stupid day”. Also, if you like brainteasers like the “base n” puzzle, you will really enjoy the puzzles at TechInterview. These are hard puzzles to come up with, so if you have some really cool ones, donate some to these guys.)

radio internals

Radio Internals – If you are wondering how I fixed the problem in my Radio blog, dialog.alert() is all you need to know. I poked around in the (in myUserland.root) outline inserting such lines of code until I found the spot it was hanging. When I realized it was in the piece of code that draws calendar for the archive, I just went into the myUserlandData.root and flipped the variable saying that I wanted outlines to “false”. voila! Back alive! This was actually pretty fun, because the scripting language, Usertalk, was new. And it was also fun to figure out how things are working underneath. Another thing I found is that the script editor in RU allows breakpoints. I never use breakpoints, though. A common joke about ASP’s intrinsic Response object is that it should have been called Debug, because everyone debugs ASP by doing a bunch of Response.Write() statement. JavaScript tends to be the same; people just use alert() all over the place. Of course, there are debuggers available for most every language, but they are a pain to set up. And there is a small part of me that feels some brain muscles must get atrophied if you never have to use your imagination and intuition to find where something is breaking. Breakpoints and code-stepping are just too … boring.


Just finished my 14-page paper on setting up web services that span multiple sites in different geographic locations. After outlining everything, it was simply a matter of a many hour typing marathon. It is all original material; not a single line of cut-and-paste. It encapsulates lessons learned from a number of practical deployments, so hopefully it will be published soon. Funny how the paper has nothing at all to do with my current job responsibilities, and I have let it encroach my personal time normally spent writing XSLTs. I will be glad to be done with this obsession. Later today I will post an interesting reply to my flame on Bruce Perens.


ALIDL – Dave shows ALIDL side-by-side with WSDL; how can I resist the urge to write XSLT? 🙂 He says “I’m a big believer in A-B comparisons. So here’s the WSDL version of the Manila interface, graciously authored by Simon Fell. It’s not exactly an apples-apples comparison, Simon’s spec describes the interface for all copies of Manila, the ALIDL version shows all the interfaces supported by a single machine. You’ll find Manila in there alongside a bunch of other pieces of software running on the machine.”