On Jun 26, 2009, at 12:56 AM, Doug Schepers wrote:
Art, Chaals, Mike, and I discussed this yesterday, and we agreed
that this seems like the best solution. Like the Widgets work, a
deliverable doesn't necessarily have to be in a single spec, so we
believe there is sufficient justification for this in the charter.
The plan of record would be to split out the SQL Storage section
into its own spec, with an alternate spec edited by Nikunj, and to
publish an updated draft of Web Storage that points to both those
other drafts. This way, all parts of the web storage deliverable are
put on a level playing field to be judged on their individual
merits, and subject to being edited and updated individually.
I don't think the Web Storage draft (I assume by this you mean the
remaining draft that would define LocalStorage and SessionStorage)
needs to link to either of the other drafts. I also don't think
proposals for other storage APIs need to be "alternate"; they could be
orthogonal non-overlapping specs and implementors could choose which
make sense for them.
I should add that I still think SQL Storage is a good technical
solution to the problem of structured client-side storage. Web
developers who are specifically targeting mobile devices, or in
particular iPhone, have given extremely positive feedback about both
LocalStorage and SQL Storage, as well as the HTML5 Application Cache.
On general-purpose Web sites, of course, uptake is limited by the lack
of other implementations so far. But Web developers seem positive
about it as a technology, based on feedback from presentations. All of
this makes me doubt that a fundamentally different model for
structured storage is needed or would be significantly better.
That being said, I would like to see the technology succeed on its own
merits and not by being tie to the so far more widely deployed
technology of LocalStorage/SessionStorage.
Regards,
Maciej
