Older blog entries for haruspex (starting at number 56)

19 Feb 2004 (updated 19 Feb 2004 at 12:31 UTC) »
cactus: I so wish I could agree with you about Enemy at the Gates - but it was crud. :-) The only bright spots were:
  • Bob Hoskins IS Khrushchev - the role he was born to play!
  • and Rachel Weisz. 'nuff said.

My biggest problem is I didn't know which side to root for. And neither did the screenwriter. Let's see... do I cheer THESE fascists, or THESE fascists? In Hollywood tradition, the suspense depended upon manipulative, circumstantial, episodic peaks and troughs, such as the hanging of the child-traitor. My favourite scene from the whole film, the only one I felt had a ring of reality about it - the grandfather and child shooting the wolf in the snow - was completely fictional, according to Russian friends familiar with Vassili's life story. Ed Harris' character was a two-dimensional caricature of a Nazi, who belonged to the Indiana Jones genre instead of a pretentious war film. I suggest anyone in the mood for a bleak WWII period piece should see The Pianist and avoid Enemy like the plague. Sorry, had to get that off my chest.

This [Helma Object Publisher] seems interesting:

Helma is a highly integrated open source web application platform that combines the strength of the Java platform with the power and simplicity of scripting languages. Server side scripting languages such as Perl and PHP have been a mainstay of the Web since the early days. But scripted Web apps have often proved to be less than well-designed and maintainable. Helma, which is scripted with standard JavaScript, was built to encourage clean and simple application design.

Why another scripting framework/application server?

Because, quite honestly, the one we need hasn't been built yet. What we need is a platform that offers a high level of abstraction on the prevalent web application concepts, but is not dumbed down and stripped of power and flexibility in order to be "idiot proof". The problem of so many J2EE solutions is that they are targeted to Java developers. But how many web developers you know sleep with the EJB spec under their pillow? That's what we thought. On the other hand, high level platforms often sacrifice the smart concepts in the underlying layers because their architects don't trust their users to be able to grasp those concepts. That's why JSP was patterened after Microsoft ASP when everybody doing serious web work already knew how broken that approach was.

Our credo is: Building websites should not be a task of system level programming. But when you bring web site building to the "ordinary people", they should be able to use the serious tools to get the job done. That's what we're trying to provide with Helma.

"Great moments at work brought to you by the Microsoft Office System."

Are you kidding? I could name some Not-So-Great moments it's brought me.

"The new Microsoft Office System helps you focus on what's important" -- presumably, they mean, "such as sending us more money."

As a great friend reminded me yesterday, vector products are alien technology. Then he mentioned Clifford algebras.

I know I'm never going to know what he's talking about in this lifetime. I can't even get an lburg code generator working...

"This code is in the public domain. You may not use it for commerical purposes"
IANAL but I think that's a contradiction. Either it's in the public domain - fair game for all - or it's a license with strings (see GPL). If that combination of conditions is indeed ambiguous, the original author needs to make it sensible.
13 Feb 2004 (updated 13 Feb 2004 at 14:46 UTC) »

Microsoft spokesman Tom Pilla today told The Associated Press that some incomplete portions of the Windows 2000 and Windows NT4.0 source code had been "illegally made available on the Internet".

... access could also provide a competitive edge to Microsoft rivals, who would gain a much better understanding of the inner workings of Microsoft's technology.

It is but to laugh. News flash, M$: your rivals already have the competitive edge. Looking at M$' code can only make things worse...

Sounds more like this "leak" is a straw-man invention to allow M$ to deflect the blame for the accelerating virus plague.

Use this one

Ack... nothing against johnnyb or his new book - and not to start another flamewar like the great man vs info cataclysm of 2003 - but I get the shudders thinking of x86 assembly being used as a first contact to machine language. I have never come across anything so baroque as the Intel instruction set... it's a shame that its pointlessly painful perversities are now by default a rite of passage for student programmers... There are many more fun machines to learn about :-)

But I'm not the first to say it... I'll let the always erudite /. crowd speak:

Calling X86 assembly "baroque" is like calling the Grand Canyon a "ditch".
(From a thread reviewing a different book on x86 assembly.)
31 Jan 2004 (updated 31 Jan 2004 at 11:10 UTC) »

Don't you hate it when your M9312 bootstrap/terminator card won't work and your PDP-11/34A just sits there with blank console and no RUN light. Replace it with an M9301-YF and everything's sweet including Console Emulator running via the SLU. I wonder if it's just a ROM problem on the M9312.

The other 11/34A doesn't even like the M9301. It's blank console and no RUN with either card in. Hope it's not a dead CPU.

Going to try exchanging the CPU card sets between the two machines.

Update: it's a dead KY11-LB console card.

lloydwood, when you wrote Why, Win2000+Cygwin gives you all the functionality of MacOS X! I'll have to assume you had tongue firmly in cheek.
19 Jan 2004 (updated 19 Jan 2004 at 08:08 UTC) »
I beg your pardon. I failed to read man printf before posting, and after also consulting the documentation for glibc, I agree, I would have expected it to print the same value twice. It looks like your snippet may show a bug, unless we are missing something...

As you say, using hhu the output is in line with the wording of the manual (quoted in my last diary entry):

        signed char x = (signed char)-1000000;
        printf ("%hhu\n", x);
        printf ("%hhu\n", -1000000);

Of course the solution to the whole IP Patent Opensource Infringment Thing is really quite simple: Release all code anonymously via identity-protecting file-trading networks ...
I read this, and I first thought what a great idea. Then, thinking further, I realised it would often backfire, because despite the anonymity of publishing, specific bits of code are often "owned" or accessible to very identifiable people, who would be traced and punished mercilessly for releasing it. Hell hath no fury like a corporation scorned (ask the...oh never mind).
17 Jan 2004 (updated 25 Jan 2004 at 00:11 UTC) »
mathieu: You ran,
        signed char x = (signed char)-1000000;
        printf ("%hhd\n", x);
        printf ("%hhd\n", -1000000);
and presumably you saw, like I do,
man printf on this Debian system does not specify whether the argument is really supposed to be narrowed to char, but the glibc manual interprets C99 to mean:
Specifies that the argument is a signed char or unsigned char, as appropriate. A char argument is converted to an int or unsigned int by the default argument promotions anyway, but the h modifier says to convert it back to a char again.
...which would support your suspicion that the runtime behaviour is wrong (and the argument should be narrowed before printing).

47 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!