--- Log opened Thu Mar 29 00:00:57 2012
00:17 < mcallan> conseo: my fix works.  i have to do some testing to make sure that it breaks nothing, but i think it's going to be okay :-)
02:03 < conseo> cool
02:03 < conseo> mcallan: finally i have it documented: http://whiletaker.homeip.net/votorola/harvester/javadoc/votorola/a/diff/harvest/pipermail/PipermailHarvester.html
02:25 < conseo> i am off for today, n8
02:29 < mcallan> n8 c
02:31 < mcallan> (i will try to look later tonight, after i finish some chores... or tomorrow)
08:29 < conseo> ok, back up :-)
15:27 < conseo> mcallan: i might do dfs (depth frist search) instead of (breath first search) in PipermailHarvester for the example as it is more about getting the newest messages first and not recompiling the list all the time (which still is covered by saved state, but anyway don't see this implementation detail set in stone yet)
15:29 < conseo> vector theme is sexy btw.
20:00 < conseo> depth first search works, now bursting is the last concept which i need to get right
20:42 < conseo> javadocs updated
23:43 < mcallan> conseo: i think i see the rough strategy now, but not how it actually works.  two things are confusing: 1) the use of numerals for tree nodes and levels.  maybe better would be named levels root/index/message, and then mark the index nodes i1,i2... and the message nodes m1,m2.., (don't use the word month, except in examples, because not all pipermail archives can also be indexed by year or week, for instance)
23:47 < mcallan> 2) there is no alogoritm for the jobs.  the prose description would be better replaced byt pseudo code or a 1,2,3 sequence of steps.  don't add detail though (there is already too much), but try to remove all detail non-esssential to a rough understanding of the sequence of events
23:56 < mcallan> (what i mean here is you need an algol for each of root job, index job, and message job, because that's where all the work is done.  impossible to understand without knowing what those jobs do - just *roughly* - and how they relate to/depend on each other.  the algols will tell that story)
--- Log closed Fri Mar 30 00:00:12 2012