Question for any LJ hackers or mathematicians out there

[robinturner]

Well this is puzzling. I'm working on a quick dirty Perl hack for downloading my journal (comments and all). It goes like this:

#!/usr/bin/perl
use LWP::Simple;
for ($count=579; $count<600; $count++) {
$head="http://www.livejournal.com/talkread.bml?journal=solri&itemid=";
$url=$head . $count;
$content = get($url);
print "$content \n";
}

Of course the last bit will be changed to append to a file, rather than fill the terminal with HTML. The problem with this method is that most itemids aren't used (so you download zillions of error pages), and I can't see a pattern for the one's which are used. I mean, can anyone see anything meaningful in this sequence?

76946
77116
77555
77741

OK, the numbers get bigger, but that's not much help. Of course I could include a search string for "No such entry" and not print that to the file, but I'd still waste time downloading a few hundred error messages for each journal entry.

Comments

[solri.livejournal.com]

Oh yes, so it does. This is worth studying, since I'm largely doing this exercise to improve my Perl skills, which were never up to much and are now rustier than my car.

If I can get this written, then I can progress to the fun bit, which is to convert the downloaded entries to LaTeX. Of course I could just use latex2html, but then I'd still have to write something to strip out the stuff I don't want, and besides, the HTML involved in so basic it shouldn't be too hard to convert.

Nice userpic, BTW.