Re: Defining Profiles

Tend to agree with Antoine here - understanding what a profile is is best
supported by understanding the practical value (functionality) - including
content negotiation, validation and things like form configuration, but
also discovery modes.

Negotiation is probably the simplest of these - it has only a few
significant aspects AFAICT:
1) identity of profiles
2) possibly dereferencing identifiers to get more information
3) substitutability of subprofiles

The other requirements for profiles need more description (and hence
guidance).

Splitting the discussion is only sensible if negotation is a non-negotiable
requirement :-) In that case delegating these aspects of defintion and
behaviour to the negotiation team to define is the appropriate response, we
cant afford the time to allow separate threads to reinvent different sized
wheels here.

Rob


On Wed, 10 Jan 2018 at 06:56 Antoine Isaac <aisaac@few.vu.nl> wrote:

> Hi Karen,
>
> I'm jumping in without having had the time to read the long threads
> entirely nor to react to them... but if indeed there are two definitions
> that the group can't seem to reconcile right now, then it sounds a good
> idea to split the discussions. And btw perhaps try even not to have a
> discussion on profiles in the context of DCAT update at the same time (as
> it appears in the last thread).
>
> In terms of schedule though, I'd be tempted to suggest doing the other way
> round. I.e. wrap up the Content Negotiation deliverable without too much
> commitment on what a profile is. This way the Guidance document could be
> instructed by a practical approach - I believe that the 'functional' aspect
> will be easier to determine with practical technology (i.e negotiation)
> being specified. In this approach, it would make sense to make a lot of
> progress on the DCAT update before pushing on the Guidance too.
>
> On the other hand I see that there is much value if we could get an
> agreement for the Guidance doc soon enough, and if the Guidance doc
> specifies how to document a profile and what sort of spec is in a profile,
> then this is excellent 'functional' progress too.
> So maybe in the end the timing/sequence doesn't matter much. What is
> important is that we accept to be very relaxed about discrepencies and
> vagueness in our definition(s) of profiles for some months.
>
>  From an editorial perspective this can be done with a big fat warning in
> front of each document, saying that the definition of profile is not yet
> stabilized - and perhaps even calling for readers to help us on it.
>
> Cheers,
>
> Antoine
>
> On 09/01/18 18:59, Karen Coyle wrote:
> > All,
> >
> > I'm beginning to wonder, after reading and answering comments on our
> > profiles discussion page, that it may be clearer to separate the two
> > deliverables, Profiles Guidance & Content Negotiation, until we have
> > clarified them independently. My feeling is that we are using the term
> > "profile" in different ways for the two deliverables, and that is
> > causing confusion.
> >
> > Our timeline (and perhaps simple logic) requires the Guidance
> > deliverable to precede the Content Negotiation deliverable. If we can
> > reach agreement on this in principle, and work to develop a definitional
> > and perhaps also functional basis for the Guidance deliverable, we can
> > then determine how that fits with the Content Negotiation deliverable,
> > and if they can share the definition of "profile" or if adjustment needs
> > to be made.
> >
> > Note that it is instructive to look at the requirements for profiles in
> > the UCR.
> >
> > Let me know if you think this makes sense as a workflow.
> >
>
>