Author archives: Mike
It's been way too long since I listened to Brave New Waves (Ogg stream reminds me that I've been neglecting experimentalism...)
Written by
on
in
Vindaloo.
The CBC, Canada's public broadcaster has had a long-running experimental-music show, Brave New Waves, which I used to listen to back in University (using a plain old stereo). I would actually listen to the CBC most of the day while doing my thesis, switching back and forth between CBC1 and CBC2 as various ...
Is my only muse insomnia (Little bit of poetry refuses to leave my mind...)
Written by
on
in
Vindaloo.
Dashing the poor poetry against the rock of self-publishing should teach it better than to keep me up at night.
This is largely a reaction to something a friend once told me, namely that she just wanted a relationship where she could feel something when it ended. It's somewhat foreign to me, because I tend ...!-->!-->
Gee these days are getting a bit too long... (Oh well, SIP project is basically finished...)
Written by
on
in
Snaking.
Moved the SIP project to about 90% finished. Everything is there technically (save actually testing with the hardware, which I can't do myself), but needs some UI polishing to be pretty enough for a real demo.
Went to the restaurant Jerusalem this evening for the management meeting... the food was far better than the usual ...!-->!-->
Apparatus for allowing transfer of electrical equipment without service interruption (An idea for those inventors out there who like such things...)
Written by
on
in
Vindaloo.
The apparatus described here is intended to allow you to transfer, for instance, a server computer from one electrical connection to another without interrupting power to the machine. The apparatus consists of a flexible set of prongs arranged in such a way that they can slide between a North American-style power plug and socket (from ...
Should have looked at Shtoom (He realises too late in the evening to do it...)
Written by
on
in
Snaking.
Just realised that Shtoom will have the authentication mechanism already written (somewhere), since it acts as a client to normal (authenticating) servers. Should have just wandered over there and pulled out the authentication-generation code.
I've developed a bit of a blind spot there I guess, wanting to avoid learning a large package while on a ...!-->!-->
New server is in (Looks much faster than the old one...)
Written by
on
in
Snaking.
Not able to do much on it yet, but setting up an SSH session was so much faster it's almost scary. Hopefully with just the demo server running on it we'll retain some of that snappiness.
Bog, or tar-pit? (Authentication detour goes a bit haywire time-wise)
Written by
on
in
Snaking.
This little detour trying to get authentication working has taken way more time than it should have. Thinking I may just back the changes out and ignore the question for the demo. Blah.
Sigh, looks like the collapse is coming... (Myopic, spendthrift policies to continue south of the border)
Written by
on
in
Vindaloo.
It's interesting talking with people who are watching the US slide deeper and deeper into unsustainable debt while ticking off just about everyone else in the world. There's talk of massive recession, complete devaluation of the dollar, and basically a collapse of what passes for society down there.
The problem, of course, is that as ...!-->!-->
SIP the cool-aide (Little things are so irritating...)
Written by
on
in
Snaking.
"So," figures I, "really should have at least some user authentication in the SIP server." Little did I realise that the Twisted code, while it supports authentication quite simply on the server, doesn't seem to have any code on the client side.
That is, the server can handle a client that knows how to authenticate, ...!-->!-->
Should have spent a day or two just familiarising myself with SIP (But oh well, that's life...)
Written by
on
in
Snaking.
AFAICT, the SIP protocol has no agreed upon conventions for registering display-names or other meta-data. That is, while you can attach display-names to just about every URI, there's nothing stating which URI's display name should be considered canonical when parsing a registration request. For now I'm assuming I should display the contact field's display name. ...