[switch to blog order view]

I finally did one of these myself

December 16, 2009

I've been a bit busy lately. Amy and the kids moved out to Los Alamos to be with me in September of '08, and suddenly I had no time. Well, I've just taken a job back in Bloomington, at FlyBase. I started a week ago, but Amy and the boys are still in Los Alamos, until Nattie is done with the fall semester of school. Adrian, as many of you know, is being home‑schooled.

Which means I finally had time to do it. I have heard that everybody who calls him or herself a programmer does this at some point, so I must be having delusions along those lines.

Been thinking for some time about putting up a web app that draws views of the Mandelbrot Set. I wrote a perl script to draw it a few months ago, and did some nice stuff with spectra to color it. But that was just for me to play with. I wanted to turn it into something that was a little different from the other Mandelbrot set explorers out there. (There are a lot, since this is a common obsession among programmers.)

So I finally had a few days to think it over. I wanted to have a point‑and‑click kind of update, which is normally a java thing. I also wanted to make use of the fact that the time‑consuming part is classifying pixels, not coloring them - so mine draws pictures in every color set I have available, every time. That meant that I also needed a way for users to switch easily between spectra. Finally, I wanted users to be able to make grand high‑resolution images of parts of the set that appealed to them.

So see what you think. I kept the perl script as the engine, and turned it into a cgi. But I wanted also to keep the php template around it. I toyed with the idea of grabbing the template php in perl and basically interpreting it, but finally decided instead to embed the perl cgi in an iframe. I understand that there may be a couple of browsers out there that can't handle iframes, so I'm sorry if that's you. I achieved point‑and‑click‑ness by using an image map, with the coordinates for the zoomed part embedded there dynamically. Javascript, of course, handles switching between color schemes.

For GURPS fans

May, 2015

I wrote a character sheet manager for the GURPS role-playing game system. Follow the 'GURPS JCSP' sidebar link, or click here.

Just in case you were wondering, this took a lot more time than the Mandelbrot set generator I wrote a few years ago.

A lot more time. Srsly.

 
 

Not much content yet

Oct. 27, 2007

I'm learning a bunch about CSS and stuff, so I'm concentrating on the site layout and structure. I don't want to be stuck with something I don't like maintaining, so I'm not adding lots of material until I get things the way I want them.

Stuff may well change while I work through this part. I've already gone through a couple of versions. This was, in fact, an experimental layout, to see if I could make it work - so far, it has only one problem I haven't been able to lick. 'Course, I haven't tested it in any browsers except Firefox and IE.

One remote test with Safari on a Mac...which is to say that my dad has now seen the site.

Code pulling together

Oct. 29, 2007

OK - fairly major update to the site structure...

I decided to move to a php-based template system. This will make it easier to expand the site while keeping a consistent look, and of course easier to change that look across the site if I like. Slightly more complicated to work with, of course - but I've had some experience with this sort of thing, so it's not bad. Besides, it also opens lots of possibilities for interesting dynamic content. Now that I've said that, I'll probably feel obligated to do some of that, though...

Nov. 3, 2007

I dug into some alternative ways to get my column heights to work the way I wanted - that 'one problem' I mentioned earlier. Found a pretty comprehensive list of various ways to solve the problem on the Front-Page CSS Wiki.

I'm going with 'Solution A' (padding excess / negative margin compensation / overflow:hidden cut) and crossing my fingers, because I don't want to use faux columns - too many things would have to be changed. Sounds like the only disadvantage is that I can no longer safely use anchor links in my text - something I probably wouldn't do anyway.

So now the attentive site viewer (so far that will basically be me) will notice that the text on Adrian's page no longer has a scroll bar - the page now extends to accomodate all of the text.
As does this page, now that I've got more stuff on it.

I've just noticed that I'm updating this page in old-school anti-blog order, that is, chronologically.
*Ahem*

The Gömböc

Dec. 9, 2007

I have discovered the mysterious enlightenment of the Gömböc. And I want one.

You know you want it...

I stumbled across a reference to the Gömböc in a NYT article; part of a page on noteworthy scientific advancements of 2007. This is a beautiful example of mathematics given form, and that form then transcending its arid origins to become art.

This is a mono-monostatic object, i.e. one with but a single stable equilibrium orientation. Actually, it's pretty easy to make an object like this (think Weebles), but this is the first known example of one that is convex and homogeneous (no ballast). Or it was thought to be, until someone noticed that certain types of turtles have shells that are quite close to this shape:

Coincidence?  We think not.

It seems that the shape makes it pretty easy for these turtles to turn themselves back upright if they are turned over. Other kinds use the ancient art of turtle-fu:

I would not mess with this turtle.

This, and other arcana related to the mysterious Gömböc, can be found here.

A Little of That Ol' Dynamic Content

Jan. 21, 2008

Well, it continued to bother me that I was 'blogging', but I was doing it backward. I just can't seem to get with the anti-chronological-order thing. So I decided I'd have it both ways.

Click on the '[switch to (other) order view]' link above to reverse the order of my posts. Those of us from the (really) old school who prefer to read things from past to future, older to newer, can still do so. You youngsters can read things the new way.

OK, OK. I realize that the entire point of blog ordering is so that you impatient folks can read the very latest stuff without having to do anything - no clicking, no scrolling. So I suppose it would be pointless to leave my text in chronological order, even though I like it better that way. Us more patient oldsters don't mind a click to get things the way we want them, so I'll make the upside-down order the default. Happy?

The Protein Structure Viewer

Feb. 3, 2008

I've put here a copy of a tool I'm building for the team I work with at the lab. There's also a link in the sidebar at the right.

I'll keep my copy as similar as I can, but by the time I'm done with this tool, it will be hooked in to the infrastructure of the HIV website in ways I can't emulate here. So my local version will be more of a demo, and won't do all the cool stuff that the real thing will do. When the real one goes live, I'll probably put a link to it here too.

In the meantime, if you know anything about biochemistry, the demo here is already fairly interesting. It uses Jmol, a nice java applet for viewing molecules. But I've added a little magic of my own, to make the tool useful for AIDS/HIV researchers. Users will be able to bring a set of sequences, perhaps newly sequenced from a patient of interest, and align it to a reference sequence we have and an HIV protein sequence for which we have solved the structure. The tool shows a 3-D graphic of our protein, and shows them how and where the sequence(s) they bring are different from it. The part that will be of interest, we think, is that users can then select parts of this sequence and the tool will 'highlight' the corresponding amino acids in the protein in the graphic, immediately. They can see which physical parts of the protein are affected by mutations in their sequences, compared to the reference and protein sequences.

Try it out, if you are interested in that kind of thing.

The Real Protein Structure Viewer

March 20, 2008

One of my bosses (Bette Korber) has a workshop coming up (in Banff - but she didn't invite me along :( ) where she wanted to be able to show off some new stuff. As it happens, much of her new stuff is projects I have been doing. So she wanted the mosaic vaccine design tool suite and my protein sequence structure feature viewer to be put on the public HIV server right away. So, there they are, even though the protein viz thing isn't really finished. When Bette says 'jump', everybody tries to see who can jump the highest.

We really need a better name for that protvis thing.

The mosaic vaccine thing is also pretty cool, although I only wrote part of that set of tools. The idea is that you can design vaccines for badass hyper-variable viruses like HIV and Ebola by making artificial proteins that are a kind of 'super-combination' of pieces of the real virus's proteins. The trick is in how you distill the real proteins - it's a pretty sweet idea, and it just might work, too. Did I mention before that It Is Cool to be working on a cure for AIDS?