Tuesday, 2014-05-06

*** mlncn has quit IRC01:15
*** bkuhn is now known as bkuhnIdle01:21
*** bkuhnIdle has quit IRC02:18
*** tgnitidle has quit IRC03:00
*** tgnit has joined #npoacct03:01
*** tgnit has quit IRC04:27
*** tgnit has joined #npoacct04:27
*** mlncn has joined #npoacct04:31
*** nesciens has quit IRC05:02
*** mlncn has quit IRC08:27
*** bkuhn has joined #npoacct12:27
*** bkuhn is now known as bkuhnIdle12:29
*** nesciens has joined #npoacct13:43
*** bkuhnIdle is now known as bkuhn13:57
*** tgnit has quit IRC13:59
*** tgnit has joined #npoacct14:00
*** mlncn has joined #npoacct14:20
*** bkuhn is now known as bkuhnIdle15:20
*** tgnit has quit IRC15:41
*** tgnit has joined #npoacct15:44
*** bkuhnIdle is now known as bkuhn17:21
*** bkuhn has quit IRC18:22
*** bkuhn has joined #npoacct18:24
*** tgnit_ has joined #npoacct18:27
bkuhntgnit_: So, you saw my email about the idea of writing up a short paper on this issue of fixed point/ floating point w/ precision?  Thoughts on that?18:34
tgnit_@bkuhn  Yes, I did. I will start the thread on the project mailing list soon18:36
tgnit_bkuhn: I have two questions on that?18:36
bkuhnOk, that sounds like a good idea.18:36
bkuhn(re questions) go ahead!18:36
tgnit_bkhuhn: How do you suggest that we can be more compelling on MPFR vs fixed point issue ?18:38
bkuhnMy main suggestion is to find a few more references of people who have studied this question... the ACCU article you found is good; but I suspect there are other places this is written about.18:38
bkuhnI'm particularly interested if you can find a source in some accounting journal or other academic publication specific to accounting that covers this issue.18:39
tgnit_Should we also start this discussion on the ledger cli google group?18:39
bkuhnBut, if we can't find that, scientific journals that talk about how MPFR library specifically (or the process it uses in general) are better than fixed point would be good.18:39
bkuhntgnit_: I was thinking that; it might be.18:39
bkuhnActually, maybe it's best to cross post to both that and the NPO accounting group?18:40
tgnit_Yes , I will do that18:40
bkuhnSomeone on the npoacct list a while back suggested fixed point arithmetic.  I'd have to look in the archives as to who.18:40
bkuhnbut, it's good to post it both places.18:40
bkuhntgnit_: do you want me to introduce you on the ledger list first?  I don't mind if it's useful to you.18:41
tgnit_I have access to academic papers and I will search to quote some references  of studies in this regard18:41
bkuhnI've told the folks on #ledger IRC channel already that you're doing GSoC, etc. but perhaps it'd be useful to say the same on the list?  It's really up to you -- whatever you're more comfortable with.18:41
bkuhntgnit_: Great, I thought you might have such access at your University.  See what you can find.  If you feel like writing a paper about this, that might be useful to you anyway.18:42
tgnit_Also I am thinking to ask this question on the MPFR mailing list, what do you think?18:42
bkuhntgnit_: that's a good  idea -- but, you should check first to see if it's in any FAQs.18:42
bkuhntgnit_: BTW, I wrote to a colleague I know (he's on Conservancy's board, Mark Galassi), who has contributed to MPFR before (he's an astrophysicist by training and now a developer of scientific software)...18:43
bkuhn.... I'll let you know what he says.18:43
bkuhntgnit_: did you see my question above about "introducing" you to the ledger mailing list?18:44
tgnit_I read their FAQs they've mentioned that you can get exact number of precision bits(eqaul to setting fixed precision)  and there is a function to do so .18:44
tgnit_I will do it myself  :)18:45
bkuhntgnit_: (re: MPFR FAQ)  my suggestion would be to post something to the MPFR mailing list quoting that part of the FAQ, and ask: "Given this, is there any reason to used fixed point arithmetic at all?  I'm working on an accounting application and it seems using a fixed number of precision bits might be the right answer.  Does anyone have general thoughts about using MPFR in an accounting application?"18:46
tgnit_Good idea !18:46
bkuhnThat should leave it open-ended enough to solicit lots of opinions. ;)18:46
tgnit_See The calc calculator by Landon Curt Nol on http://www.mpfr.org/18:48
tgnit_under software using MPFR18:49
* bkuhn looks.18:49
tgnit_I think I should mention that while mailing to MPFR as a suggested example.18:53
bkuhntgnit_: that's a good idea.18:53
*** bkuhn has quit IRC19:24
*** bkuhn has joined #npoacct19:27
*** tgnit_ has quit IRC19:31
*** mlncn has quit IRC23:09

Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!