Thursday, March 30, 2006
I have not done anything with Lisp for about the past month, except some evangelization. Very frustrating. I just saw that the next rev of SBCL hit the streets, and I'm still behind on getting RPMs built for the previous rev. In my defense, I think I need to vacuum/blow out the insides of my computer. I think there's a whole heap of dust in there, causing the processor thermal protection to shut down the system when it's under heavy use. I fired off an SBCL build a couple weeks ago to try to build the RPMs and the workload caused the box to shutdown mid-way through.
On the evangelism front, the folks at work are getting almost sick of me mentioning Lisp (note to self to back off once in a while, lest I devolve fully into an over-the-top Smug Lisp Weenie). I have to say that Lisp would be very cool for a network appliance CLI, though. There is a huge amount of Greenspunning in the XORP codebase with which we spend a lot of time (oodles of template files that drive the CLI, along with custom yacc/lex parsers for those, etc.). All this could be replaced by Lisp, decreasing the total amount of code maintained while increasing functionality.
It's sad to hear your coworkers are almost getting sick of it. Your company seems to be working on opensource; yet a few years ago, I think most people would've been sick of the opensource advocates who paved the way for opensource acceptance.
Its interesting to hear that you work on Xorp. I could see Common Lisp being quite useful in a long running daemon like situation. I've been working on my own p2p-like applications written in Common Lisp. Not much has come of them yet but they're growing, slowly but surely. Did you think about possibly rewriting pieces of XORP outside of work? Something that would give your coworkers a concrete example of Lisp's power? I think I'll pull down a copy of it and take a look for myself.
and again, appreciate the work your business is doing. If we can only get a replacement for the traffic load balanacing (application level/Packeteer) like technologies, something that was distributed, open source, and easily scalable; boy would that be amazing.
I've been enjoying learning Common Lisp from PCL, but I've yet to think of an interesting project to do, aside from some Emacs stuff. Any ideas?
Post a Comment
Links to this post: