09 Feb 2003

projects that never finish


there are alot of things that i want to do. one can just look at my ideas/todo list and see how many stupid ideas have (ohh .. the python bindings for spirit is finally done! i should update that page :)

there are a couple of cool projects that i haven't taken all the way, but i hope that i will eventually. i'll sum up some of them here to let you know what they're about.

firstly, there's a library i wrote when i first came over to uk called libworldtime. i noticed that there isn't a real good way to extract timezone information in linux/*nix. evolution uses its own ical stuff, glibc has built in timezone description files that is used for their timezone stuff. but there hasn't been any easy api to access it. so i wrote this lib to try and write an applet to do very easy world time conversions, rather than dodgy hacks. i should finish this and write a world time applet for gkrellm.

second, i've been working on a pure python implementation of the icq v8 protocol. its very close to completion, although the code is a little messy. python, unfortunately is not a pretty language to implement a protocol easily. especially when the protocol is a piece of shit like ICQ. as an extension to this, i implemented inside a cross-protocol framework allowing the plugging in of different IM protocols. so eventually i would like to have a single interface to all thee protocols currently available, like MSN, yahoo, ICQ, AIM, Jabber, etc.

but even better, on top of that is a plugin based bot architecture that can allow python based bots to roam on IM networks. using my unified IM architecture, it means that you write one bot, and people can access them in anyway, thru IM, or even thru SMS or Web, as long as you implement a plugin to my unified IM interface. lots of fluff and hand waving right now. hopefully i'll have something to show for and demonstrate soon!

one of the biggest projects i'm working on right now, is a source-based linux distro for ipaqs. basically, i'm sick of the clunkiness of familiar, so i'm working on a replacement, my own handrolled custom distro where i can build everything from scratch (like my cross compiling toolchain, to high level stuff like X and GPE, etc.) the distro is gonna be based on uclibc and busybox, which are tiny replacements for glibc and gnu*utils. then once the basic stuff is there to boot and login to my ipaq, i'll start with another project i've got on the queue.

those are the active projects on my plate right now. the libworldtime is in a nearly complete state, i've just got to do some testing and write a demo app with it, and hopefully release it. thats the one closest to completion. my py-icq interface is preliminarily working. i can run a demo bot on the network that does cool queries, but i need to read up on natural language processing to make it neat and fun to use. the distro for ipaq is currently receiving most of my attention right now. its got alot of the gentoo build characteristics. uses files call ibuilds, and similar concepts like merge and unmerge, use flags, etc. it'll nearly be like gentoo cross-compiling build system for ipaq.

i can't wait to get these projects released out of their 0.0.1 stages.

other projects i've got on the back of my mind, is using a combination of ecore and evas to write a window(screen)-manager for ipaq and possibly write a gtk2 over evas binding, so gtk apps can directly run on top of it without modification. i think raster (author of enlightenment) has some similar ideas as me. also, i'm thinking of writing a gtk2 mp3 player that is lightweight just like xmms but without skins! i hate xmms because of the skins and insisting on gtk1.2. but i also hate rhythmbox because they hate gentoo and its a bulky imitation of itunes.

so i guess its a lot of coding for me for the next couple of months!

al out.


You can reply to me about this on Twitter: