Been a busy bee since I got back from the
xmpp interop event all those months back - hence the paucity of updates.
For one, we had a bunch of things to fix and for another we decided to redesign parts of the server in light of what we learned so that future changes become easier. This week, checked in a huge tonne of diffs and a whole bunch of new files - the total change around 1 Mb or so :-)
Does the server still work ? It does :-P We will see how much regressions when QA gets their hands on it ... *shiver*
In other news, started working on a new incremental attacktable based chess engine (ya, ya, god knows how many I have written by now !) : this time focus is on trying to find out how much of info can be leveraged from an attacktable in different facets of the engine, and surprisingly - it is extremely useful in a whole bunch of places : even places where I did not initially expect an impact.
And if you want a good laugh, take a look at the REST Dialogues
1 ,
2. Its a nine part series, but the first two have been really good :-)
I was reminded of something strange today btw : A long time back, I posted
this and
this. In response to my first post,
Kousik had actually commented about using
xmpp as the backed (damn Haloscan - all the comments gone , gone :-( ).
At that time I did not know about xmpp - and yet today I was reflecting about something similar of how things like this could get retrofitted into xmpp if we solve the issues of scale (eg:
3,
4) with the xmpp federation (at the extreme ends so to speak).
The addressing/routing, authentication, privacy, server to federation, etc considerations are already handled within the protocol itself, and there are extensions on how to publish/subscribe, multicast on top of this network : along with alternate forms of connectivity like HTTP - not just direct tcp/ip - damn, someone using this to make an online marketplace can make a killing ! VC anyone ?!
*Update* Take a look at this post - embodies most of what I totally dislike about SOAP, other than the fact that it is so un-HTTP like and is essentially fancy rpc tunnelling on top of HTTP.
2 Comments:
It's interesting from an intellectual standpoint but I don't think answer to the question or the search for an answer actually improves a person. Therefore, it doesn't seem worth it.
-D
The search for answers to questions like this could make a person better ... depending on how sincere he/she is.
Like, Gray Walter's tortoise model can be used to simulate some pretty complex ideas ... though I never got around to implementing it : the basic idea is so simple and yet so extensible.
Ofcourse, in Scott Adams case, I guess it is baiting readers : pure and simple !
Post a Comment
<< Home