--- Log opened Sat Oct 27 00:00:26 2012
04:38 < mcallan> conseo: let's discuss the problem with count engine first, before looking at solutions.  that's my suggest
05:03 < conseo> mcallan: sure, i can layout the requirements which make that a good design choice. lock-free stateless multithreading alone is reason enough imho, which is requirement for (near) realtime counting. but i have to think about it a bit further first since ovn has to belinked in as well.
05:06 < conseo> mcallan: we have the history problem which means we need an overhaul of the counting concept anyway. it will also make our engine a lot more flexible than that of liquid-feedback for instance, since we in fact only define algorithms for the tree counting structure, which can be reused in any context (pure functions are perfect reusable components)
05:09 < conseo> anyway, let me play with it and then we will see. i have only told you so, because i don't want to code something you don't like and understand. i would also write the next harvester functionally as far as i can, because i have realized that due to the multithreading book you gave me, i have unknowingly already put it in closures (MonthFetcher...) (runnables with immutable state)
05:14 < conseo> ok, let's be blunt. i would like you to agree on a functional language and on the techniques, so we can apply it in places where it gives us flexibility (my main attractor). you cannot approach that pragmatically by problem, because you can solve each problem with OO/imperative programing as well
05:16 < mcallan> so you suggest we move to a func lang?  again, i would want to begin discussing problem, not solution.  it's wrong to assume func lang is the sol'n, when problem not yet discussed
05:24 < conseo> well, i can do it in java or better javascript as well. it is rather a programming/design pattern. i'll take everything back and ask you again, what do you think about functional programming? (also note: we don't do that at university, i am really convinced that it has many benefits)
05:24 < conseo> (i can also draft the reasoning for the count-engine, but i will do it on the list, if you are open to functional programming in general)
05:26 < mcallan> am open to any solution that best solves a real problem.  otherwise (if there's no problem), not much interested in solutions.  lack of time
05:27 < conseo> agreed
05:27 < mcallan> ok
05:28 < conseo> i will post to the list then
05:33 < mcallan> ok, and i am also free to skype (or mumble etc).  whatever you think is best
05:33 < mcallan> (irc no good for this)
05:42 < conseo> mcallan: do you have mumble running?
05:42 < conseo> i need to go to my singing lessons now, will be back in 2-3 hours
05:44 < conseo> cu later
05:45 < conseo> btw. they made it although on thursday they only had 400k. i have pledged for two boards, i am curious about the potential and impressed by the efforts :-)
05:45 < conseo> http://www.kickstarter.com/projects/adapteva/parallella-a-supercomputer-for-everyone/posts
05:59 < mcallan> interesting... i only have mumble client, true.  so i guess that's not enough.  skype then
09:25 < conseo> ok
09:44 < mcallan> ok, your lead.  let me know where/when
09:46 < conseo> mcallan: about the functional stuff, its premature. i have to model it first. i can give the reasons and advantages, but i am still new to it. in general we can talk about progress and needs for the talk track, but besides enjoying talking with you i have no concrete issue
09:46 < conseo> i still need to setup the list and wire it all together
10:00 < mcallan> conseo: i think you simply like func stuff. :-)  maybe best apply it first in school context.  because i fear (while it's good for votorola) u take too much time away from school
10:01 < mcallan> i think we will talk soon about concrete stuff, anyway.  i doubt agm and oddi will take up too much more of my time.  then back to coding for me
10:08 < mcallan> our talks are never short, and i want to ensure we get at least *something* concrete for our time :-)
10:09 < mcallan> thomasvonderelbe: methinks u are here only in spirit
10:19 < conseo> mcallan: yes, i agree
10:20 < conseo> i think thomasvonderelbe is only here, because i have not deactivated the account. it would be nice if he would rejoin us from time to time for a chat, but maybe irc is too boring
10:26 < mcallan> no, i think thomas has other reasons for not being as active as he'd like
10:28 < mcallan> but i was only joking about his irc account.  i think he has no client running.  u know his setup better.  maybe email him, and ask if there are problems
--- Log closed Sun Oct 28 00:00:44 2012