Re: [metadata] Launching DPIG Metadata Task Force

Thanks Bill!

I'll try and collect the thoughts into something a little more coherent for
the wiki :)

And I agree that what, not how, is the right approach to take.  I'm too
quick to jump to "solutioning" as my wife might say in management-speak.
 Or, to make sure we're on the same page, the focus is to collect
requirements, we don't necessarily make recommendations for solutions.

Further inline below...

On Fri, Jan 24, 2014 at 10:19 AM, Bill Kasdorf
<bkasdorf@apexcovantage.com>wrote:

>
> One other way to look at this—and one that’s consistent with the mission
> of this DPIG—is to say “what metadata do publishers need to incorporate in
> their documents, and does the OWP give them a way to do that?” Echoing
> Ivan, that’s not a _*how*_, that’s a _*what*_.
>

My only concern here is being a bit clearer on the extent to which we need
to answer the second part, as to whether the OWP provides a way to do that.
 Clearly, the OWP/web architecture includes resources that can describe
others [please, no http-range-14 discussions!].  Is that sufficient, but
provides no practical recommendations, or otherwise how do we avoid getting
into "how"?





>   So for bibliographic metadata, I would point you and others to
> CrossRef, which is the means by which scholarly literature is interlinked.
> While the DOI is the mechanism behind that (and yes, I know your view of
> the DOI!), it is the bibliographic metadata _*associated*_ with the DOI
> that makes it work.
>

:) We can agree to disagree on the current DOI implementation, but I think
we do agree on the need for a publication level identifier that isn't
necessarily associated with any one representation.


> Well, what bibliographic metadata do publishers in general think is needed
> to “make it work”? HOW it gets made to work is outside the scope of the
> DPIG; but surfacing the use case from publishers _*that*_ they want this
> to work is what we’re about. The reason I point to CrossRef is that, at
> least for journal articles (and increasingly for books) they have evolved a
> practical complement of bibliographic metadata that is sufficient to “make
> it work” for what scholars and scholarly publishers need to accomplish. No
> more than that. Worth looking at.
>

Definitely. After I hit send on my mail I thought about the crossref
metadata as another example of existing practice.  Also the headers in
Elsevier's XML representation of articles, via CrossRef's Prospect API.


 As to your question regarding the “line” between descriptive and
> bibliographic: okay, it may be a fuzzy line. I deliberately used generic
> terms (even avoiding the loaded word “semantic”) because I want so focus on
> the concepts and the use cases, not the technologies, at this stage. Here’s
> a simple top-of-the-head distinction:
>
> --Bibliographic metadata describes what the publication _*is*_, who
> authored it, who published it, etc.
>
> --Descriptive metadata describes what the publication is _*about*_:
> subject information, primarily.
>
> There is probably a better term than “descriptive” because “describing” is
> fundamentally what all metadata does, but that’s the sense I was using it
> in.
>

Okay, thanks! For curiousity, do you consider the abstract bibliographic
(this is the abstract of the publication) or descriptive (this abstract
describes the publication)?

Perhaps the descriptive metadata is that subset which could legitimately be
added by a reader, rather than the publisher?  In that view, an abstract is
bibliographic (it IS the abstract) where as a summary, that could be word
for word identical to the abstract, is descriptive?



>   Most of what I was intending by those distinctions was just to give us
> some related “buckets” to discuss these issues so there isn’t a major
> mish-mash that becomes impossible to sort out. Mostly based on a MODS-like
> way of looking at characterizing different types of metadata, but again,
> keeping it generic, not wanting to make the discussion MODS-driven.
>

And please not MODS/RDF driven.
Wiki-wise, is it worth pulling those buckets over to separate pages?


>  Hope this is helpful. One other comment: I do think that this kind of
> discussion on the list, complementing the somewhat more structured
> “discussion” on the wiki, are both important parts of the process.
>
>
>
> But to reiterate what I think is fundamental to what we’re about here: we
> need to focus on the what, not so much on the how. What do publishers need
> to do regarding metadata in the context of the OWP (as opposed to all the
> other places metadata lives in the ecosystem)?
>

And why as a non-publisher, I'm happy to comment from a distance :)

Rob