[POC] Added JCR specification

[dantleech]

I have written a small script to convert an HTML version of the JCR spec (as exported from libre office using the original doc file) to RST.

Its not perfect, as images are not preserved and the references are broken. I'm sure we could do better.

[dantleech]

@dbu @lsmith77 wdyt? worth having?

[dbu]

sorry, i have missed the original PR notification.

why would we not include either a pdf or just link to http://www.day.com/maven/jcr/2.0/ ?

[dantleech]

The benefits are:

• Its referencable from within sphinx
• It is contained within the PHPCR website (so user can easily navigate without getting lost in the web)

The disadvantages:

• The conversion is not 100% perfect
• It is one big page
• It duplicates the day docs.

I the first two points can be fixed - the last is true, but seeing as day is no longer a company (?) And its a JCR spec which I guess is not affiliated with a company, it gives us an advantage as PHPCR is a canonical place for PHPCR.. if you get what I mean.

David Buchmann notifications@github.com wrote:

sorry, i have missed the original PR notification.

why would we not include either a pdf or just link to
http://www.day.com/maven/jcr/2.0/ ?

---

Reply to this email directly or view it on GitHub:
#8 (comment)

Sent from my Android device with K-9 Mail. Please excuse my brevity.

[dbu]

okay. i can agree to merge this if you can specify everything we need to re-generate this documentation (maybe in a build.txt file in the same folder as the generated doc, or something like that, and add your script or whatever). so that we can fix and rebuild should a problem be discovered even if you are not available.

[dantleech]

Have written a new converter, which uses the HTML version of the spec:

https://github.com/dantleech/jcr-283-doc-convert

This version splits the chapters, generates a TOC tree and It also pulls in the pictures. Next step is to try and guess the code blocks..

[dbu]

hm, did you try to ask the jcr people if they somehow have the source of the docs available? probably that would be easier than trying to scrape it.

[dantleech]

Where / who should I ask? But the originals are distributed in a word document and the online HTML reflects this.

[dbu]

maybe @bdelacretaz knows if the jcr spec "source" document really is a word file?

@dantleech would converting the wordfile then not get you closest to the original data?

[bdelacretaz]

I don't know myself but I'll ask someone who should know to have a look at this and hopefully comment here.

[dbu]

thanks bertrand!

[dbu]

@bdelacretaz any luck with finding the source of the spec document?

[dantleech]

I also wondered if there are any copyright issues with copying the java spec?

[bdelacretaz]

Looks like I forgot to ask or never got a reply - I have asked again if we can make the spec available in a more structured format than word.

[dbu]

as this is very old and we did not follow up, i close this now.