Re: How to organise?

> On 1 Dec 2015, at 05:13, Robin Berjon <robin@berjon.com> wrote:
> 
> How do you prefer to organise? Here are some proposals:
> 
> • We use a repo on GitHub and largely discuss through issues (at least
> for the concrete stuff). We can ask for w3c/scholarly-html and publish
> it off there. An alternative is spawning a new org, or creating a team
> in another one (we can do that under the scienceai org if you prefer
> that, but we don't want to make a landgrab).

I think GitHub is the right approach; many of us are already using it, so…

Regardless of the hat I wear I think having a w3c repo is the most appropriate. Although "only" a CG, we are nevertheless part of W3C after all…

If we decide to do so, I am of course happy to set it up.


> 
> • I would suggest we use ReSpec as the format. We could also use the SH
> as defined by itself but the problem with that is that every time you
> change a definition you have to edit the parts of the document that use
> it. We just did that... it's not always fun.

We could (probably should…) produce a version of the document in our own dogfood at the end. But that is down the road, I think using Respec as an intermediary is a good idea.

> 
> • Contributions should be made ideally through PRs more than through
> discussion — it makes for more concrete discussions. At the very least,
> if discussion then trying to root it is best.
> 

PR-s, lots of discussion through issues… I personally would not dismiss emails. I still find emails the best option for certain kinds of discussions; PR-s are more when things become very specific. I know I am old-skool, though…

I see some people have already been using gitter. Adding this to our toolset looks like a fine idea, although I am not a user of gitter so far (I see it as IRC better integrated with github, that is all…)

> • Should we start in stages, first building the HTML structure, then
> overlaying semantics, etc. or should we do all at once? I reckon we
> might get easier consensus if we move incrementally but that's just a
> gut feeling.
> 

*If* we go for RDFa as expression of semantics, then we may want to have at least a rough notion of what we want to express with that, too. In my experience, combining RDFa with HTML is not always obvious, barring the simplest situations, and doing it after the fact may be a real hurdle.

I say "if" because that may be a matter of discussion. We may decide to use microdata, we may decide to embed semantics in a script element using turtle or json-ld, instead of merging with the core html tag set, we may decide to concentrate on aria attributes instead...

Cheers

Ivan


> • We should pay particular attention to interaction with
> implementations. I know that we have tools that work around our SH,
> notably a parser, a set of React components that output it, parts of a
> common stylesheet, and ideas about validation and the such. Assuming we
> can all reach some rough consensus we'd certainly like to converge them
> on the common format. Are there any other consumers/producers here?
> 
> --
> • Robin Berjon - http://berjon.com/ - @robinberjon
> • http://science.ai/ — intelligent science publishing
> •
> 


----
Ivan Herman, W3C
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
ORCID ID: http://orcid.org/0000-0003-0782-2704