[14:01] <mitchell> It's just about 2pm, so that means it's Office Hour time
[14:01] <mitchell> anyone with questions should ask away
[14:02] jonathan_busy (~jprusky@12-236-204-253.client.attbi.com) joined #chandler.
[14:02] <mitchell> there's no particular topic, we here to talk about what people are interested in
[14:02] Nick change: jonathan_busy -> jonathan
[14:04] <jonathan> anybody want to (and ready) to talk about a Chandler vs. Haystack comparison?
[14:06] <ducky> um.  I hear it's slow.
[14:06] <mitchell> hmm, I know Mitch looked at it, but he's traveling today
[14:07] <ducky> katie also looked at it a bit, but she and Morgen are having a conversation with sound waves...
[14:08] <ducky> Have any of you (non-OSAF people) looked at Haystack and have useful feedback?  Aside from "slow"?
[14:10] <jonathan> I just heard about it the other day when the link to the presentation was posted
[14:10] <jonathan> does anyone have an opinion as to *why* it is slow?  do we have reason to believe that Chandler will be faster? why?
[14:11] <atotic> it looks like they've loaded it with everything they could think of
[14:11] <jonathan> (not doubting that it would be, just interested in any differences in the focus / mission / architectural goals
[14:12] <mitchell> As i understand it, Haystack is written in Java, and little if anything has been done to optimize for performance
[14:12] <mitchell> again as i understand it, there are conceptual similarities but the technology underpinnings are different
[14:13] <john> I don't think it's slow because of Java
[14:13] <mitchell> what do you think the causes are?
[14:14] <john> I don't know for sure, but I suspect it might be the following:
[14:14] <john> not enough indexes
[14:14] <john> memory based vs. disk based
[14:15] <john> RDF tripple storage model
[14:15] <atotic> with RDF, it is easy to lose performance if you are not careful 
[14:15] <john> having to access lots of data to do anything
[14:16] <atotic> i do not think they have concentrated on performance at all, the first time i ran it, it took between 5 and 10 minutes to start
[14:17] <katie> john: you might want to check out haystack's "view architecture"
[14:17] <john> ouch
[14:17] <atotic> it is better now, around 1 min
[14:17] <katie> haystack performance: yes, looks like they just didn't make it a priority at all for the first release
[14:17] <atotic>  they did some db initialization on the 1st run
[14:17] <katie> they seem to think there's no reason they can't optimize to fix problems
[14:18] <katie> but it was so slow as to discourage one from playing with it
[14:44] amblin (~youwish@cosmo.itlab.musc.edu) left irc: Remote host closed the connection
[14:56] jurgen_ (~jbotz@63.90.156.153) left irc: Remote host closed the connection
[14:57] Nick change: jonathan -> jonathan_away
[14:58] jurgen_ (~jbotz@63.90.156.153) joined #chandler.
[14:59] <mitchell> Well, there don't appear to be a lot of buring questins for today, as i'll formally end our Office Hour for today.  Catch you all later