Re: SoLiD, and LDP dependency

On 24 July 2015 at 00:37, Amy G <amy@rhiaro.co.uk> wrote:

> Hi all,
>
> I raised concerns recently on github about SoLiD's dependence on LDP
> perhaps making it unsuitable as an API candidate for the WG:
> https://github.com/linkeddata/SoLiD/issues/38
>
> My concern is that LDP over-specifies implementation details, ie.
> restricting which technology stacks are viable, which is not the role of
> this group, and doing so would significantly impede uptake.
>
> Options could be either abstracting LDP-specific parts out of the SoLiD
> spec and considering it on that basis, or reframing it instead as a layer
> between the Social/Federation specs (whatever they end up looking like) and
> LDP for implementers who *do *want to use LDP as the basis for their
> server (the latter being beyond the scope of this WG).
>
> Slightly more expanded thoughts in the thread on github, anyway. I'd be
> interested to hear what others think.
>

-1

Thanks for raising concerns.  It would be more helpful if it were possible
to offer concrete ways to improve things.

To say "LDP is over specified" doesnt really say how it could be improved.

It's clear that LDP is standards quality work, for reading and writing data
in general, standards compliant, and a relatively minimalistic way, as it
went to REC status.  What would you cut out.  How could it be done better?
What specific improvements could be made?

Saying "rewrite the spec" or "make it more abstract" isnt really much to go
on, given that LDP WG spent 3 years trying to create consensus here.
Perhaps Arnaud could offer more insight as he was chair.

I think comments like "significantly impede uptake" are subjective and
perhaps not as constructive as they could be.  On the web it's not easy to
predict what will get uptake, and it's natural for everyone to have a view
and for those views to often conflict.  What is perhaps more practical is
to select a technology that is able to address the use stories in a
standards compliant way, aligned with the deliverables of this WG.

I do think LDP does that, with the caveat that it is more oriented to
turtle (a W3C REC) than to JSON LD.  This is perhaps something that could
change tho, with this WG being one of the motivations.


>
> Amy
>