Increase interoperability

[linonetwo]

Please look at https://github.com/linonetwo/solid-tiddlywiki-syncadaptor

Let's continue discussion here instead of #1 (comment)

[linonetwo]

I'd like to use bag for the path in POD this tiddler should store in.

Just like what tiddlyweb does.
https://github.com/Jermolene/TiddlyWiki5/blob/c05c0d3df66e587f35c5cd3eedcac432b1eed012/plugins/tiddlywiki/tiddlyweb/tiddlywebadaptor.js#L229

---

update:
Thanks to Jeremy Ruston and PMario, I know it's ok to use bag just in sync adaptor.
https://groups.google.com/forum/#!topic/tiddlywikidev/SZ2eV8_X22U

[bourgeoa]

I understand that bag has been added to tiddlyweb to store the path from where a tiddler comes from to be able to edit it and store it back to the good location. It's a way for a tiddlywiki to be multi-POD.
As such I think it is not to be stored in pod but only on local storage.
I wanted to use it for a multipodwiki.

I think it is a good idea to use it as from the beginning

[linonetwo]

It's unable to put file's metadata into container, see:

So I decided to use /public/tiddlywiki/main/index.ttl to store SkinnyTiddlers.

And for the text field of tiddler, I will just put them into files inside /public/tiddlywiki/main/, so inside /public/tiddlywiki/main/aaa.txt there will not be metadatas, there will just be contents (aka, text field).

[bourgeoa]

I don't think it is a good idea.
I feel tiddlers should be kept with a .TTL

As for the index file. Either you keep it or rebuild it when syncing from pod to tiddlywiki, because it is redundant. The point is performance, you need to read all tiddlers to rebuild it.

Actually I only load tiddlers as needed.

Last and most important I am not at all sure that all tiddler's fields but text are included in the index.

[linonetwo]

I'm not either, I've just asked on the forum https://groups.google.com/forum/#!topic/tiddlywikidev/b1InjYGd95Y

How to rebuild it on the fly? NSS 5.0 currently seems can't support SPARQL on multiple ttl files, so we can't just perform a SPARQL on the container, and return an array of all files, that have tiddlerfields without the text field.

I think the way to build an index on the fly is:

1. GET container to read the file list
2. perform SPARQL on each file, return all tiddlerfields without text field ( so the server won't need to return text field (possibly very long)