another weblog

AWA 13, ACen 2007 galleries back online:

http://photon.ninjawedding.org/awa13/
http://photon.ninjawedding.org/acen2007-chicago/

Web Pages That Suck covered the mystifying navigation problem almost a decade ago. But it looks like people still haven’t gotten the message.

Take a look at the Acer Aspire One marketing website.

It’s like we’re back in 1999!

• Blob interfaces: The interface is one big Flash blob. If you do not have a Flash player, you have no way to get to any successor nodes in the web. (By the way, all Acer is using Flash for on this website is pointless animations, and you can easily do those outside of Flash. Look up script.aculo.us or jQuery’s effect library.)
• Shiny balls: Can you figure out what those mean? If you can, then you’re either sharper than me or know that Acer produces the Aspire One in those four colors. I thought it was meant to be some sort of range: left is the low-end, right is the high-end (called the “Aspire One” for some reason). No, that’s not the case at all: they’re all the same model. But there is no feedback to indicate this. If you hover over the balls then you get nothing to indicate just what they represent. The problem is compounded by the replacement of the mouse cursor with text reading “Select your one”, which is ambiguous — you could be selecting a color scheme or an entirely different model. The only way is to click and guess!

If you’re ever asked to design a website, please don’t put design ahead of functionality. Aside from being a hinderance, it’s no longer trendy to be obtuse.

Exporting wireframe data and light intensity as render passes and feeding those into the compositing engine. Quick test for some AMV opening thing I got asked to volunteered to do.

I manage ACen’s public website with git. Recently, I made a fairly large series of commits to add information about m.o.v.e and their presence at ACen 2009.

Now that all that information and content is public, I thought I’d share this little tidbit:

nevrast:~/prj/acen trythil$ git diff --shortstat before-m.o.v.e..m.o.v.e
 225 files changed, 6508 insertions(+), 4715 deletions(-)

before-m.o.v.e and m.o.v.e are tags pointing to before and after the m.o.v.e additions, respectively.

Lots of changes.

It looks like I’ll be going to Erlang Factory. Anyone in the San Francisco Bay Area interested in meeting up?

EDIT: I am dumb. See the panel in the lower-right corner?

You click it, and the mode changes. Obviously intuitive, as are the command descriptions. (TorF? PkAt? MvSZ? Yeah, I know those.) And, as you can tell, I’ve found a model of cholesterol, so I think we’re good to go. Reset of post left for historical purposes.

Is 3D molecular modelling a really complex domain or something?

I think I have a fairly good grasp of the basics of chemistry, but I’m having a hell of a time trying to figure out how to do simple stuff like adding atoms and bonds in PyMOL. Supposedly, there’s a better modelling program out there called “O”, but good luck searching for that. (And it’s probably high-priced proprietary software anyway, which means I wouldn’t be able to use it even if I found it.)

Anyone have any suggestions for molecular modeling software? I need it to build 3D structures of organic molecules having roughly the same complexity of cholesterol. It should be free/open-source software.

Links to friendly PDB databases would be helpful, too; I do have tools to convert PDBs to my required formats. I’m aware of services like RCSB’s data bank, but there’s a bit too much data for me to sort through there, and my knowledge base of organic chemistry isn’t rich enough to do pattern-matching on the models they do have. (Do a search on “cholesterol” there and you’ll see what I mean.)

Being a proponent of “open source” isn’t an excuse for not buying anything, and vice versa. Not buying anything just means you’re frugal, a cheapskate, and/or a copyright infringer, depending on context. (Some cheapskates do gravitate to free and open source software, but that doesn’t make them proponents of the respective philosophies. It just means they’re cheap.)

Usually I don’t care about random misconceptions on the Web, but a lot of people (including people I know) get this wrong. So there you go.

I file bug reports and — when possible — offer patches to Free/Open Source software I use in my work. It’s the Right Thing to do.

I recently (as in “past couple of days”) filed a bug report against nested model forms in Ruby on Rails 2.3.1, and attached a patch. The ticket has since received some discussion and the patch has been revised twice.

I also recently (as in “past couple of months”) filed two bugs on components in Adobe’s Flex SDK. The first one was a change request to change the commit behavior of the change event in the numeric stepper control, and had a workaround (easily convertable into a patch) attached. The second described an internal compiler error on invalid input, and had no patch attached since (as far as I can tell) the source code to Adobe’s ActionScript compiler is not publicly available. To date, there has been no public activity on the numeric stepper ticket, and the internal compiler error has been relabeled as a duplicate of an internal-Adobe-bug-made-public-but-with-lots-of-redacted-information, and has no additional followups.

I can’t make any comparison between these tickets because they’re on different technologies, different platforms, different degrees of difficulty to implement. But I still think the difference in response time is funny. (It probably also says something about organizational priorities, too, but there’s too much noise to say anything definitive there.)