Re: Request to register "identifier" relation type

[Peter Williams <[e-mail]>]

>
> semantic purity may be an elusive goal to chase.
>

Completely agree. Striving for semantic purity, in either relation
definition or usage, is a fool's errand.

Personally, i think relations have usages more than definitions. There are
clear and confusing ways to use relations, not really right or wrong.

The more the usages of a relation diverges the less useful it is. This
makes it important to ensure that new relations are evocative and cohesive
enough that usages will tend to cluster together and that they are distinct
enough from existing relations that they will not reduce the usefulness of
existing relations.

Peter

On Wed, Aug 9, 2017 at 11:39 AM, Erik Wilde <[e-mail]> wrote:

> hello.
>
> On 2017-08-09 09:59, Peter Williams wrote:
>
>> Having multiple relations, one for each semantic, might be another way to
>> address the usability issues. Some of those use cases seem to be covered by
>> existing relations. For example, `canonical` seems tailor made for the
>> "Version Identifiers" use case. For other use cases, such as
>> "Multi-Resource Publications" and "Persistent Identifiers", there don't
>> seem to be any existing relations that would work. Relations for those
>> narrower use cases would be much easier to understand and use.
>>
>
> wrt to "semantic purity" of link relations: any attempt to avoid existing
> link relations because they are used in slightly incoherent ways and define
> new ones that are "more clearly defined" often just ends up adding to the
> overall noise.
>
> pragmatically speaking, i think we have to accept the fact that no link
> relation has a precise and closely followed semantic definition. it always
> ends up being a loosely defined concept, and then it depends on the context
> to let us know what exactly this link relation means given the context
> where it is used.
>
> i have no specific opinion regarding this case, but it just seems to me
> that the general path taken here (analyzing real-world usage of existing
> link relations and then deciding that some of the usage is not quite in
> line with the intended use in a new context) might lead to unnecessary
> duplication, and ends up with defining another link relation that over time
> will end up being "misused" anyway.
>
> this is probably more of a general "how narrowly can we pragmatically
> define and guarantee the semantics of link relations" rant. i know that it
> would be great to have semantically pure link relations that we can
> interpret out of context because they are used consistently everywhere. it
> just seems to me that given the experience of how link relations are
> defined (generally fuzzy) and used (generally based on context), semantic
> purity may be an elusive goal to chase.
>
> cheers,
>
> dret.
>
> --
> erik wilde | mailto:[e-mail] |
>            | http://dret.net/netdret    |
>            | http://twitter.com/dret    |
>