Tuesday, 2013-12-17

*** bkuhn is now known as bkuhnIdle01:57
*** bkuhnIdle has quit IRC04:22
*** pokoli has joined #npoacct09:28
*** tbm_ is now known as tbm11:26
*** bkuhn has joined #npoacct11:29
joarbkuhn: I've been working on the API stuff lately.13:11
bkuhnjoar: great, how's it going?13:11
joarmade a small P-o-C GTK client, started on a flexible storage system that could let you use ledger, SQL or anything else.13:11
joarrealized that editing transactions will be hard with ledger.13:12
joarI will try to solve that issue.13:12
bkuhnYeah, I knew that would be coming.  Ledger is bad at that in some ways, but it would be ok if you made extra rules.13:13
joarIt will probably end up with a regular expression looking for the transaction ID, which will be a tag in the ledger file and then finding out the bounds of the transaction in the text file, possibly changing it in place or placing it in a temporary file, deleting it from the main file and let ledger parse it then edit it and write it to the end of the file again.13:13
bkuhnFor example: you could demand that the entries be valid for Ledger but not as free-form.13:14
bkuhnjoar: another idea: you could insist that every transaction is in its own file.13:14
joarI also contemplated that you could provide a way to simply reverse the transaction.13:14
bkuhnThis would be annoying to command line users, but solvable.13:14
joarI think that would make it a hurdle for people migrating from ledger-cli.13:15
tbmjoar: ofx has the concept of a unique id for each transaction; it might be worth looking into that and adopting that to identify transactions13:15
joartbm: I use uuids at the moment :)13:15
tbmif you can import ofx from your bank, you'd get a unique id for many transactions automatically13:15
tbmand it would wave the way towards exporting ledger as ofx, which might be useful to import it into other software13:15
joarthat's good. The IDs do not have a restriction on them, I use UUIDs to programmatically generate them, but a few are set to nonsense strings by me.13:16
bkuhntbm: do you have any other thoughts on how to handle the issue of editing the ledger file in place by the API?13:17
tbmbkuhn: I don't13:18
joarhere is an ofx python module: https://github.com/kedder/ofxstatement/blob/master/src/ofxstatement/ofx.py13:24
joarbkuhn: Regarding licensing, for example the ofxstatment project only mentions its license in one place, and that's in https://github.com/kedder/ofxstatement/blob/master/setup.py#L49. Is that enought for the project to be considered AGPLv3?13:34
joarI suspect it's a fluid line, and that would generally not be seen as enough.13:35
* bkuhn loks13:38
bkuhnjelmer: BTW, are you around?13:38
bkuhnjoar: are you asking: "is that text copyrightable"?13:38
bkuhnThe answer there is surely yes.13:39
joarI'm not sure I understand?13:39
jelmerhi bkuhn, joar13:39
joarhi jelmer!13:39
joarI was wondering about the license of ofxstatement, and saw that it was "licensed" under AGPLv3, but it only makes it clear in that one place.13:40
bkuhnjelmer: Hey, so did you see my introduction to scott_gonzalez yesterday?  scott_gonzalez is Javascript hacker I know who has done a lot of work on jQuery and is  working on a side-project for bug tracking software.https://github.com/scottgonzalez/debt  You and I had discussed bug tracking software recently and thought you and Scott might have a lot to talk about.13:40
joarreading http://oss-watch.ac.uk/resources/opensourceyourcode#applying-the-licence it seems that there's some things missing13:40
bkuhnRelevance to this project is we haven't picked a bugtracker yet and we want a good one that's Free Software. :)13:40
jelmerbkuhn: no, I missed that - scrolling up now - thanks!13:41
bkuhnjoar: So, the URL you pointed me to is someone else's code, right?13:46
bkuhnSo I thought you were asking me about that.13:46
joarI have some work to do with applying the license correctly to accounting-api too. I will do that now.13:49
bkuhnjoar: actually, I was just looking that tissue13:51
bkuhnhere's what I suggest13:51
bkuhnI don't suggest copyright notice in each file13:51
bkuhnrather a top level file13:51
* joar was just preparing for a find -name '*.py' | xargs13:52
bkuhnlet me add a COPYRIGHT file now13:52
bkuhnjoar: ok, I just pushed my commits14:03
bkuhnjoar: so, I suggest adding this to each file that you've written:14:04
bkuhnPart of accounting-api project:14:04
bkuhnLicense: AGPLv3-or-later14:04
bkuhnat the top14:04
joarwill do14:10
bkuhnjoar: I noted the Spinx stuff you included mentions BSD license14:28
bkuhnyou should copy their LICENSE file into something like 3-Clause-BSD and change it to refer to that rather than LICENSE14:28
bkuhnadd to LICENSE file info that says:14:28
bkuhn"Some files in the docs/ directory were borrowed from Spinx, which is licensed under 3-Clause-BSD [or 2-Clause-BSD... I didn't check].  See the file [N]-Clause-BSD for details14:29
joarconcurrent work15:11
joarbkuhn: should I add a license notice pointing out the Sphinx license in docs/source/conf.py and docs/Makefile or should I add that pointer in LICENSE?15:16
bkuhnI'd say it makes sense to add it to LICENSE15:16
bkuhnjust to keep all LICENSE info in one place15:16
bkuhnyou could say in LICENSE:15:16
bkuhn"Spinx's license can be found in docs/source/conf.py"15:17
bkuhnor somesuch15:17
bkuhnIt's ok if there's a lot of pointer indirection. :()15:17
joarbkuhn: this is the sphinx license: https://bitbucket.org/birkenfeld/sphinx/raw/c411c343dd216f41d263b5f5a14ad14d3ff5dad7/LICENSE15:21
bkuhnThe top license (the one labeled "License for Sphinx") is 2-Clause-BSD15:22
joarshould I include the entire file or just the 2-clause BSD?15:22
joaronly the 2-clause, since the files in accounting-api does not have anything to do with the software incorporated in sphinx.15:24
joarbkuhn: pushed15:30
bkuhn(and you answered your own question correctly, as I think you know :)15:31
joarbkuhn: one thought I had was to simply reverse the transaction instead of updating it.16:26
bkuhnjoar: ugh, it drives me nuts when accounting systems won't allow you to change a transaction.16:26
bkuhnThat's such a hold-over from paper-based ledger systems.16:27
bkuhnbetter to keep a log of what happened.16:27
joarin sweden you aren't allowed to change transactions in your accounting system.16:27
bkuhn(there are probably some accounts who would see that as a feature rather than a bug, though, I should admit.)16:27
bkuhnI mean, GAAP in the USA tends seems to indicate that it's acceptable to change it if the books for that period aren't closed and there's a record of who did it.16:27
joarThat's what I hear from my father and brother-in-law at least16:27
bkuhnjoar: I'm not entirely surprised... I suppose it's a 'mode' we should support... but I don't think we want to insist that it works that way, either.  What do you think?16:29
bkuhnilan: I'm sorry to say that I was really busy this weekend and missed the SCALE deadline.  But I just submitted a talk on this project there, so16:29
bkuhn... I just submitted it.  If it's too late, I understand.16:29
joarbkuhn: if we implement both then people may decide for themselves.16:36
joarreversal would be technically easier with ledger, and probably easy to do with every other backend, but not everyone likes that behaviour.16:37
*** nesciens has joined #npoacct16:37
*** bkuhn is now known as bkuhnIdle20:33
*** bkuhnIdle is now known as bkuhn21:07
*** zernia has joined #npoacct22:32
*** zernia has quit IRC22:44

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