[Proposal]: Doesn't use any dataset or model that does not compatible with Open Source

[wannaphong]

I think we should doesn't use any dataset or model that does not compatible with Open Source in our project (PyThaiNLP). If the license that doesn't compatible with open source, we should ignore theirs datasets/models.

[bact]

+1

For data,
see some licenses that compatible with OKFN Open Definition here https://resources.data.gov/open-licenses/

For code,
see all OSI-approved licenses here https://opensource.org/licenses

[wannaphong]

+1

For data, see some licenses that compatible with OKFN Open Definition here https://resources.data.gov/open-licenses/

For code, see all OSI-approved licenses here https://opensource.org/licenses

@bact Do you think about LST20 Corpus? It is free for open source project only. I thinking remove all LST20 model and replace with Blackboard Treebank. 18c8c50

[wannaphong]

Blackboard Treebank license is CC-BY. https://www.facebook.com/dancearmy/posts/10158423653343284?_rdc=1&_rdr

LST20 Corpus (free for non-commercial and open source only). https://www.facebook.com/dancearmy/posts/10157641945708284

[bact]

NC (non-commercial) clause makes things not fully open. OKFN Open Definition states that to be open it must also be open "for any purpose".