Development note for February 7, 2015
I managed to get scion-browser to compile last night with some help from various people on the freenode Haskell IRC channel last night. I then spent several hours trying to get SourceGraph to compile, but I have given up on after a couple dozen iterations with different dependency constraints. Fortunately, SourceGraph is not an essential component and I can probably get by without it. I can't say the experience was a complete loss; I learned quite a bit about how cabal works, at least.
So, I'm done tweaking my development environment for now. I'm going to put "getting a better development environment" on my to do list at a lower priority. I still need to learn a lot of cabal-fu and maybe figure out how to get a working cygwin environment that includes GHC (something I failed at miserably before).
Anyway, on to actual projects. I have a few open ideas I'm pursuing:
PHP parser
Some time ago, I wrote most of a Happy/Alex parser for PHP. However, I never developed a meaningful IR for the parse tree, but it needs updating to reflect changes to PHP since I wrote it, and there are a few constructs I don't think it handles correctly. Source code is at github: https://github.com/ab9rf/phaestus
ham radio repeater database
I used to have a database that was a curated combination of several different public repeater databases. I had it hosted on one of my own computers and it went away when I had to shut down the hardware. I want to resurrect it and rewrite it (it was a bunch of grotty ugly perl code) in Haskell. It seems like a good opportunity to learn about the various Haskell web engines (snap, yesod) and my previous experience tells me that Haskell is especially good at ETL, which is a lot of what the prior app was doing.
color space explorer
I've had a long interest in color theory and I've had a bunch of silly toys, usually web-based, to further this. One of the things I'm interested in is coming with parametric ways to come up with "harmonious" color schemes from one or two seed colors. Also, color names.
native haskell implementation of curses
There's an existing Haskell binding to the curses library, but it sucks. I kinda want to write a fork to Dwarf Fortress, and a curses library would help with this.
minecraft MCA file reader
Yes, I play Minecraft. And I actually already have Haskell code that does a pretty decent job of reading Minecraft .mca (Anvil) files. Mostly this needs to be cleaned up and extended to cover more cases. It's a shame that Minecraft uses the JVM, since that means I can't actually write Minecraft mods in Haskell (no, don't tell me about Scala, I already know about Scala). The "core" of this nonsense is on github in a gist: https://gist.github.com/ab9rf/6244b832f009a485ba85
Haskell implementation of longley-rice
Longley-Rice is a numerical model for estimating radio propagation over terrain. I'd like to write a Haskell implementation, partially to use in conjunction with the repeater database mentioned above. Might just solve this one by compiling the original FORTRAN code and using the FFI.