The ICFP Programming Contest has kind of started. They've released the data, but haven't told us what we're supposed to do. That hasn't stopped us from making a number of interesting discoveries about the data.
However, the first program I've written for the contest so far was in C++. It is a pretty small and stupid program, but it was actually faster to write it in C++ than figure out the Haskell bindings for the appropriate library.
For those interested in such things, I've migrated a few more of my projects from the CIS Subversion server to Free the Mallocs. In particular, I've moved
otftofd, and my XSLT library for manipulating trees. So you can now grab the latest versions by
darcs get http://free-the-mallocs.com/repos/env-tool
darcs get http://free-the-mallocs.com/repos/otftofd
darcs get http://free-the-mallocs.com/repos/tree-xslt
When I have time, I will probably also put up my repository for the very preliminary attempt at a play-by-email version of Settlers of Catan.
So I just declared the current state of OCamlTeX to be good enough for a new release. At this point I am debating whether I think it would be more valuable to me to focus more on a Haskell port at this time. I guess I will see what the response to this new release is like.
I've also got PmWiki up and running on Free the Mallocs now, in particular an OCamlTeX page. It still needs a ton of work, but I'll attack it as time allows.
Despite my best efforts, I didn't quite complete the revisions to the JFP version of "Boxes Go Bananas" by today. And do to the Engineering Alumni BBQ and unforeseen technical difficulties I didn't present on it for PLClub today either. So I'm scheduled for next Thursday.
I'm probably going to give BGB a rest while Stephanie is away and get back to hacking on AspectML and starting on InforML. I had some good ideas the other day on how to write type-directed functions without having to use declassification all the time.
In short, Robert Bringhurst says
- He does not object to starting a sentence with a word in small caps. He doesn't object to capitalizing a word in small caps, as long it is a word and not an acronym.
- He says that if I like guillemets, I should use them.
- He is not enthusiastic about concocted ligatures, and seems to indicate that other than for exact transcriptions, ligatures should only be used when the design of the typeface requires them.
- As for using typefaces to emphasize or pick out semantic distinctions he says that it would be best to consider the problem on a case by case basis.
- He did not seem to have much specific advice when answering Chris's question about typography in other mediums (especially for presentations or on the web).
- He thinks AMS Euler is a wonderful typeface design, but that it is incomplete in many ways (lack of proper kern tables, missing glyphs, etc.) and that the AMS should fund someone to produce a proper OpenType version.
I'll probably verify with him before too long that it he is fine with using more detailed excerpts of his response in a public setting.
I just found an e-mail from Bringhurst in my spam folder. The fact that it was 99% PDF probably confused SpamAssassin. Thankfully, it wasn't 99% certain or it would have sent it to /dev/null. I don't think I'll read it until tomorrow; getting my brain too worked up this late in the day is a recipe for insomnia.
This isn't really significant to talk about at the moment. Yesterday I finally got my questions for Robert Bringhurst out in an e-mail, hopefully I will hear back from him eventually.
It looks like a group of us here at Penn will be trying this year's ICFP Programming Contest. The plan is to use Haskell. As the hint is that it will involve computational archaeolinguistics, I think it is fair to guess that Haskell will be well suited to the task.
Otherwise, I've been making progress on revisions to the journal version of "Boxes Go Bananas". I'm hoping to be close to finished tomorrow or Thursday. We'll see.
I've thrown up a PmWiki installation on Free the Mallocs, but I'm still trying to figure out how to fit all the pieces together in my new web strategy. Some content from my existing web pages should probably go there, but how much? I have yet to get any feedback on Existential Type, and I've purposely been avoiding the trap of reading the logs. So if you have suggestions on the current or future design of either it or Free the Mallocs, be sure to let me know.
I'm getting pretty close to a new release of OCamlTeX. Yesterday and today I've been working on tracking down a bug with
\ocamlnewenvironment. Part of the problem, was something strange about defining the
\endmacro that seems like it should have also been broken in PerlTeX. The other problem was a result of my not thinking about the case of zero argument macros. In PerlTeX this isn't a problem because if you are calling a zero argument subroutine, you can simply give it no arguments. So I had to hack in some code to handle putting in a unit argument for zero argument macros.
Additionally, I've switched to using darcs for OCamlTeX version control. This is partly because DreamHost only provides an ancient version of Subversion. And I am admittedly rather keen on the offline commit feature. So far I've been fairly pleased, but I've been having a hard time choosing "good" patch names. Anyway, if you want to grab the latest OCamlTeX, install darcs and run:
darcs get http://free-the-mallocs.com/repos/ocamltex/
I just spent a couple minutes searching around on whether anyone else has debated the merits of a "qu" ligature. Surprisingly, I found that the typeface Andralis, designed by Rubén Fontana of Argentina, does come with ligatures for "qu".
I finally look a took at Bringhurst's "glossary of characters" and he states that the ligatures æ and œ are required in English, but only for "Deliberate archaism and pedantically correct quotation". I would guess that Bringhurst would probably not endorse my proposed "qu" ligature, but I'll perhaps ask anyway.
And for the curious, æ is called an "aesc" and œ is called an "ethel".