A friend of mine at work gave me a 512 meg imation usb stick... Being unsure what else to do with it, I chose to format it with luks (AES sha256, etc). It was something of a pain to do by hand:

new:
    cryptsetup --verbose --verify-passphrase luksFormat \
        /dev/sdd1
    cryptsetup luksOpen /dev/sdd1 imation
    mkfs.ext3 -j -m 1 -O dir_index,filetype,sparse_super \
        /dev/mapper/imation
    mount /dev/mapper/imation /someplace/

mount:
    cryptsetup luksOpen /dev/sdd1 imation
    mount /dev/mapper/imation /someplace/

umount:
    umount /someplace/
    cryptsetup luksClose /dev/mapper/imation

I'm happy to report that (as long as you have cryptsetup installed) ubuntu can read the stick out of the box. It even has a nice gnometerface for the passwd (part of HAL I gather). Neato.

-Paul

I started messing around with jquery. My goal was a simple comment system, but jquery is so easy and fun that it turned into an overly elaborate ajax chatterface. Woops.

(I'm a big fan of OpenID, so I chose that for identification purposes.)

-Paul

I'm learning abstract algebra for fun and I was pleased to discover various blackboard bold symbols are available in unicode. They are (naturally) not supported in IE — presumably ∃ a font package for them —...

Anyway, here are the ones I care about so far: ℤ and ℕ.

There are more, but firefox/gentoo doesn't seem to support the majority of them. Curiously, firefox/XP supports nearly all of them. (disclaimer: supports is defined to be on my system). The complete list is on the wiki.

-Paul

When I decided to use xslt to generate the html for my page, I did it in part to evaluate the feasibility of using it on client sites.

At this point, I'd say it's rather unadvisable to use xslt for your website the way I have done. For one thing, google doesn't cache it very well. I can live with that and I made an html path to deal with it. I'm pretty sure xslt isn't really intended for the things I have done with it here, but it should at least work right...

The really irritating thing is the last-modified time. IE[67] seem to handle it quite a bit better than firefox — which is not something I say often. Go ahead and check right now. Tools->page info in FF and right click for properties in IE.

Elinks doesn't seem to do xslt at all yet, but I'd say few people use that for real websurfing anyway. And the ones that would use elinks that way could read the source of this document with little difficulty I imagine. I wonder about Safari and Opera though. Yes I do.

Anyway, ∃ a bug ∈ Bugzilla for this little problem. It seems to be less of a bug and more of a “well, how do you handle multiple modified dates?” type of a problem. Not knowing which modified date to use, they just chose the epoch: 1/1/1970.

-Paul

I redid the PerlModules page somewhat. It was pretty bad. Now it at least sorts them by modification date... though their formatting is still pretty weak. It probably doesn't really matter. Merlyn was kind enough to point out that nobody uses my stuff anyway. I assume that's true, but a mild exaggeration.

I think it's his prerogative to be a little cruel. That's why he has a disclaimer in his sig.

-Paul

OK, I figured out how to split my links into a separate file, which is a lot easier for me to read. It occurs to me that google probably won't parse it though, which is a real problem for sites that have content people might want to find. I have a solution I think...

I used mod_rewrite, some perl, and libxslt to build a dynamically generated html path for the site. The downside is (of course) all the same things that made xslt cool in the first place imo. I also don't have a way to deal with the dynamically generated xml pages (e.g. PerlModules).

Update: I solved the dynxml problem. The only thing left is some magic robots.txt setting that will bump spiders into the html version -- iff necessary. I really wonder if google will index an xslt generated page.

-Paul

Today, I added another custom tag. I was surprised that XSLT automatically URL encodes when you cram text into an attribute. I did not know that. I was also able to make my <w> tags more powerful. XSLT is getting to be more and more fun, the more I wrap my mind around it.

Just to make sure I have taken everything to ridiculous new heights, I have added a perlmonks take so I can go to a specific node (jettero) or to a node by id (jet) — or er, like this.

And speaking of perl monks, I just got into the holders of unholy power group!

-Paul

Doesn't make sense to maintain a blog without blogging I guess. It hand't occured to me that I had a blog until I ... realized it. This all came up on the chatterbox and I was deeply amused by it. I am 12.25 pAU tall — which is about 59.4 attoParsecs. Although, I question whether google's units are really precise enough for that second number to be quite right.

In other news, I figured out how to build custom tags so I can link to Solanaceae (aka Nightshades) easily. You'd have to view the source to see why it's cool.

-Paul

I have put this off for far too long.

OK, I'm really getting a handle on things. I have made real headway. There's still a lot to do, but I think I got most things to a less embarassing place. The page sources went from 1995 stuff to um, ... well, the xsl transform namespace I'm using is from 1999. Heh. And I wanted to pretend I'd come so far...

-Paul

I have put this off for far too long.

I'm doing some serious house cleaning. My webpage has been a terrible mess for years. Worse, it's several gigs of stuff people never see.

If you came here looking for some ancient abandoned project I inflicted on the public, email me about it and I'll send you the sources.

Also, I just started, so if things look a little peice meal at first, it's because they are.

-Paul