Added data transmission info

[trestletech]

Added examples of data transmission, description of bandwidth benefits, and propose a compacted, mixed-array-based format for data interchange.

[rufuspollock]

I quite like this - @mk270 your thoughts?

[mk270]

None of this stuff belongs in the JSON Table Schema specification.

No-one should be changing that text without changing the version number.

This material, on how you might use the schema, does not belong in the document telling you what the schema is!

[rufuspollock]

@mk270 - OK, could it not go in an appendix and some of the submission is simple enhancement of examples :-)

[mk270]

To be fair, I think a bit of thought on recommended transfer protocols is a good thing, grateful that you're looking into t and I'm happy to send you a pull request splitting this out from the schema spec document.

[mk270]

where by you I mean @trestletech

[trestletech]

Not sure I understand the proposal -- would I then re-submit the revised pull request? That's fine, just making sure I understand the expectations.

And if this doesn't belong in the spec, I can just post it on my blog so I can find it later. 'Just attempting to be a good "open" steward and submit my thoughts back to the project.

[rufuspollock]

@trestletech really appreciate the enhancements and example. Would you possibly be able to submit 2 pull requests: one with the fixes / enhancements to core spec and one with the data transmission example?

We can then merge the first immediately and discuss the second :-)

[trestletech]

Just merged your request and re-instituted the changes to the example that got overwritten in the merge. If that was intentional, I can go back to the version you had. I thought having two rows in the example were a bit more clear. Your call.

[mk270]

It was intentional - the spec shouldn't change without the version number / date info changing.

[trestletech]

OK. I reverted back to your version of the example. Only remaining contribution in the spec is the addition of the data transmission example.

[rufuspollock]

@mk270 i don't think updating the examples is changing the spec and would be happy for those mods to go in IMO :-)

If you really want we can bump the spec number, but i feel this is still draft (and we could put v1.0 draft if that would be useful ...)

[mk270]

Well, I don't feel it's still a draft. That's what having "v1.0" and a date means. There are two independent implementations. And the stabler the spec, the better and more numerous the implementations will be.

[trestletech]

I'll create a separate pull request for the changes to the examples with incremented version number. Is there anything else I need to do on this one to get it closed?

[rufuspollock]

@trestletech don't believe so. @mk270 can we get the example changes in (we can update version number if you wish).

[trestletech]

All,

Would it make sense for us to branch our work into a new spec? It seems like here hasn't been much enthusiasm here for either the mixed-array format nor for foreign keys, which are the two primary ways we're using the protocol.

Perhaps it would make more sense for us to create a new spec which builds on top of these ones you provide? I'd be interested to hear your thoughts.

[mk270]

really sorry not to have got back to you; however I have actual time now to look at this, so give me an hour or two

[trestletech]

Sure thing. I don't mean to be a pest, I was just trying to get a feel for whether we were headed in the same direction or if our use case was sufficiently different to merit a separate protocol.

[rufuspollock]

@trestletech I definitely think we need to sort foreign keys #23 and primary keys #21 one way or the other and preferably in an update of json table schema (though it could be a side spec that you could use with it ...)

Suggestions in #23 and #21 as to approach would be very welcome (I can also add) and what you have there already looks good ...

[rufuspollock]

I've reviewed and I think this looks pretty good. merging now.