cburgmer
LICENSE change, things alive here?
@cburgmer
Can we change the license at the software-level to BSD, MIT or Apache?
My reasons are for the ones stated here: https://github.com/ScottDuckworth/python-anyvcs/issues/32#issuecomment-28528142.
Due to the nature of cjklib being python and the data libraries being useful in pieces, a simpler license would be a more helpful measure at this point.
I'm going to cross-post this to the google group (https://code.google.com/p/cjklib/issues/detail?id=23&thanks=23&ts=1386125345)
At the time of creating cjklib I wanted to make sure that people using the library were contributing code back. As of now I don't believe the LGPL can guarantee this any more than the more permissive group of BSD, MIT or Apache.
I would happily relicense this library's code under one of those licenses. However I am not the only contributor to some parts and for those files we would need to ask the specific contributors for their consent.
I will try to compile a list of contributors here:
- Shanghainese character pronunciation (https://github.com/cburgmer/cjklib/blob/master/cjklib/data/charactershanghaineseipa.csv) is Copyright 2010 by Kellen Parker and Allan Simon (http://sinoglot.com/wu/tools/data/), see there for license and ways of contacting them.
- StrokeOrder (https://github.com/cburgmer/cjklib/blob/master/cjklib/data/strokeorder.csv) and CharacterDecomposition (https://github.com/cburgmer/cjklib/blob/master/cjklib/data/characterdecomposition.csv) contains data from at least two other individuals that I would need to contact (Gavin Grover being one).
- Reading syllable data has been compiled by me
- Data from Unihan.txt is of course Copyright by Unicode, Inc.
Back then I was unsure whether the data should've been released under a CC license or a more source code typical license. I still don't know what is better, but feel that for this kind of data Creative Commons is probably better. Would that work for you?
@cburgmer
Python code
MIT for python code, definitely. It's simplest and most free.
Datasets
For data packages, do you have time to contact contributors or just want to put a list up? I think firing them an email if you still have the email thread it wouldn't hurt to ask if they're willing to pick one:
- ODC / Open Data Commons Attribution License (ODC-By) v1.0 - http://opendatacommons.org/licenses/by/summary/ - http://opendatacommons.org/licenses/by/1.0/ - Simple, guarantees attribution.
- CC0 - If attribution is not important - http://creativecommons.org/publicdomain/zero/1.0/
@cburgmer how does that look? For that matter maybe it's worth making a spreadsheet blog post on these chinese data sources and see if we can have them covered by ODC / CC0 / something clear and giving attribution.
I contacted Gavin and Uriah who contributed to the character decomposition/stroke order set.
Thank you. I followed up with some more information I found after finding the project's license. Let's see what his response is
On Thu, Dec 5, 2013 at 6:20 PM, Christoph Burgmer [email protected]:
I contacted Gavin and Uriah who contributed to the character decomposition/stroke order set.
— Reply to this email directly or view it on GitHubhttps://github.com/cburgmer/cjklib/issues/6#issuecomment-29886427 .
@lsmith77 good to know. Sadly doesn't work here, as contributions on data didn't happen through Git commits.
I see .. you could potentially still make i work, by simply identifying the commits in question and updating the rows in question inside the database. Not sure if this is supported in the hosted version but it should be easy enough in the self hosted version.
As an update, the decomposition information is now available under:
- the MIT license http://opensource.org/licenses/MIT
- the Open Data Commons Attribution License (ODC-By) v1.0 http://opendatacommons.org/licenses/by/1.0
Thanks to Gavin Grover for generously providing this great data set.
See: http://cjkdecomp.codeplex.com/wikipage?title=cjk-decomp
@cburgmer :
Datasets, this leave us with:
- PinyinSyllables
- PinyinInitialFinal
- WadeGilesSyllables
- WadeGilesInitialFinal
- GRSyllables
- GRRhotacisedFinals
- GRAbbreviation
- JyutpingSyllables
- JyutpingInitialFinal
- CantoneseYaleSyllables
- CantoneseYaleInitialNucleusCoda
- JyutpingYaleMapping
- WadeGilesPinyinMapping
- PinyinGRMapping
- PinyinIPAMapping
- MandarinIPAInitialFinal
- JyutpingIPAMapping
- CantoneseIPAInitialFinal
- KangxiRadical
- KangxiRadicalIsolatedCharacter
- RadicalEquivalentCharacter
- Strokes
- StrokeOrder
- CharacterDecomposition - Copyright Gavin Grover license MIT / ODC-by
- LocaleCharacterGlyph
- MandarinBraileInitial
- MandarinBraileFinal
- ShanghainesePASyllables - Copyright 2010 by Kellen Parker and Allan Simon (http://sinoglot.com/wu/tools/data/), see there for license and ways of contacting them.
- CharacterShanghaineseIPA - Copyright 2010 by Kellen Parker and Allan Simon (http://sinoglot.com/wu/tools/data/), see there for license and ways of contacting them.
Which of these are compiled by you? Which one's arent?
And moved forward, the as datasets, is MIT / ODC (IANAL, but put simply: basically attribution, don't sue me) ok for datasets?
Are you ok with MIT'ing the codebase?
@cburgmer Greetings, can I have permission to use https://github.com/cburgmer/cjklib/blob/master/scripts/convertdecomposition.py in https://github.com/tony/cihai under MIT or BSD license?
Convert Decomposition has to be updated because the old groovy.csv is updated to a new format.
Can other parts of your code that's python / non-dataset related be considered MIT/BSD now?
My status update: this is a big task.. I'm trying to make a pluggable version of cjklib with a simple API and more pythonic.
I have been racking my brain on this for a few weeks already. Particularly from an architectural perspective and how far I much can I realistically expect relational data to handle different things. Seeing what I can push out of sqlalchemy / UnicodeVariants / etc. to make puling things efficient. I'm trying to do it right. My current scribblings are at http://cihai.readthedocs.org/en/latest/spec.html.
I am probably bumping into some of the same issues you have had building cjklib. One big piece of help was not to have one big cjklib doing everything. I am going to try to make cihai just be an external API and let datasets plugin (expressjs style).
By the way, @nieldlr's excellent https://github.com/nieldlr/Hanzi is definitely worth a look at! I am trying to make sure any data I get working with cihai Niel / Hanzi can use too. If you haven't become acquainted already, it may be nice for each other to be acquainted :)
@cburgmer Any update on this?
Any chance you can release the LGPL from the data @ https://github.com/cburgmer/cjklib/tree/master/cjklib/data and make it CC-0 or MIT? I want to make them into Data Packages