Here's a specific proposal which is consistent with Sebastian's comments below:
Create a new Git repo "spdx/license-list-data" containing a README file to describe the repo and the following directories: html - A straight HTML version of the license text (basically, just escaped text + the paragraph tags) template - The SPDX license templates for all of the licenses complying with the format specified in the SPDX version 2.0 appendix text - Simple text (encoded in UTF-8) for all licenses website-rdfa - All of the pages generated for the website - these can also be parsed using RDFa json - JSON format of the licenses A bit of background - all of the above is already generated by the output tool in a build directory. It is just not pushed back to the git repo. Each of these formats has been requested to be generated at some point in the past. I'm still thinking it is worth considering a directory structure in the license-list repo where the above subdirectories are included under a "data" directory. There are two reasons - it would be easier to keep the input and output versions in sync since they are part of the same repo and it would be less work for me ;) That being said, I'm OK going with two repos is folks think it is easier for the users of the data. Feel free to comment on the above. Gary > -----Original Message----- > From: Schuberth, Sebastian [mailto:[email protected]] > Sent: Thursday, April 14, 2016 1:05 PM > To: Gary O'Neall; [email protected] > Subject: Re: Git repository for JSON license files > > Usually I'd agree the license-list repo would be the best fit. However, > that repo already contains a wild mix of files. For example, "Updating > the SPDX Licenses.txt" is in the same directory as the license .txt > files, making it very hard for scripts to parse the license .txt files. > Basically, you need to maintain a blacklist of .txt files that do not > describe a license, and you also need to filter out files like > "spdx_licenselist_v2.4.ods" / "spdx_licenselist_v2.4.xls". > > That said, I did not want to "pollute" the license-list repo even more > by adding the JSON files. That is, unless we do a cleanup and add some > directory structure to the license-list repo first. On the other hand, > people who are interested in the JSON files are probably not interested > in the "spdx_licenselist_v2.4.*" files etc., and we should make it as > easy as possible to get only what you need. > > Creating a new repo that contains just the output formats seems like a > good compromise. It still might be more than needed for people who are > only interested in the JSON files, but at least all output formats > would be in a single place then. But this new directory would certainly > need some directory structure (one per format or so), too. > > PS: I just noticed version 2.4. of > http://spdx.org/licenses/licenses.json contains a "releaseDate" of > "null". > > PPS: Also, the license terms in the individual license files (like > http://spdx.org/licenses/Glide.json) contain escaped Windows line > endings ("\r\n"). Shouldn't that probably be Unix line endings ("\n")? > > ________________________________ > Sebastian Schuberth > Lead Engineer, Chief Architect Office > > HERE Deutschland GmbH > Invalidenstr. 116 > 10115 Berlin > Germany > > ________________________________________ > From: Gary O'Neall <[email protected]> > Sent: Thursday, April 14, 2016 17:31 > To: Schuberth, Sebastian; [email protected] > Subject: RE: Git repository for JSON license files > > This should be pretty straight-forward to produce. > > It would require an extra step during release, but I agree it would be > beneficial. > > There are a couple alternatives to consider: > - Create a separate git repository that contains the output JSON, RDF > and template formats (all 3 output forms produced by the tools) tagged > by release > - Add a build directory to the existing spdx/license-list repo that > would contain all of the output files > > The latter would be an easy change - I would just need to remove a git > exclude of the build directory. It would also keep all of the license > artifacts together in the same repo which may make it easier to > maintain. > The only thing I don't like about the build directory is it feels like > we are including compiled artifacts in a source repository. Perhaps we > should rename the directory from build to something else? > > Let me know your thoughts. > > Thanks, > Gary > > > > -----Original Message----- > > From: [email protected] [mailto:spdx-tech- > > [email protected]] On Behalf Of Schuberth, Sebastian > > Sent: Thursday, April 14, 2016 4:03 AM > > To: [email protected] > > Subject: Git repository for JSON license files > > > > Hi, > > > > in addition to hosting the license files at [1] and the individual > > details at [2], how about storing them in a separate Git repository > at > > GitHub? We already have license-list [3] and license-list-XML [4], so > > license-list-JSON would probably be a good name. > > > > The benefit would be that projects hosted in Git that make use of the > > license files could easily include that repository as a Git submodule > > instead of downloading the files separately. > > > > [1] https://spdx.org/licenses/licenses.json > > [2] https://spdx.org/licenses/<spdx-id>.json > > [3] https://github.com/spdx/license-list > > [4] https://github.com/spdx/license-list-XML > > > > Regards, > > Sebastian > > > > > > _______________________________________________ > > Spdx-tech mailing list > > [email protected] > > https://lists.spdx.org/mailman/listinfo/spdx-tech _______________________________________________ Spdx-tech mailing list [email protected] https://lists.spdx.org/mailman/listinfo/spdx-tech
