On 08/04/2013, at 20:01 , Leif Isaksen <[log in to unmask]> wrote:
> Hi Scot
>
> just to pipe in and agree with Tom and Sean on this - it's best not to
> get too hung up on the vocabulary [sic]. It is one of the great
> ironies of the semantic movement that many of the terms used have
> somewhat opaque definitions (unfortunately that includes some formally
> defined ones as well). As a general rule it helps to think carefully
> about the structural role different RDF/URI resources are trying to
> provide. Some focus on cataloguing instances of stuff (Pleiades for
> example), others focus on verbs and/or categories to connect them
> together (like the CIDOC CRM), and others have very specific
> vocabularies for connection which do one thing but do it well (Open
> Annotation; SKOS, etc.). Wikipedia is often a good place to get a
> summary overview of each of their uses. In case it helps (it may not)
> there's a summary of Semantic Web developments for archaeology and
> cultural heritage that isn't quite up to date but may be handy as a
> primer in the early chapters of my PhD thesis:
>
> http://eprints.soton.ac.uk/206421/
>
> All the best
Thanks for the information. Speaking as someone with a computer science background (before becoming a classicist), and thus having a considerable pre-existing understanding - with its own specialised and related vocabulary - I find I need to be able to map the concepts in the 'RDF' and similar data models onto my own existing understanding of domain modelling before I can fully understand it. I guess I am used to using another, related but subtly different, set of terminologies to describe a metadata for describing a classical data set (or any data model).
I suspect, that the problem comes from the background of having two specialised jargons where I could previously reason about one domain with the other (perhaps not unsurprisingly, in either direction, sometimes), but now I find, in order to conform with other people's pre-existing work, that in actual fact, one of the specialised jargons has to be mapped to a different jargon. So I have a dual translation problem.
I think also, in the field of applied computer science, which is where most computer scientists end up working (i.e. in business of one kind or another), most people rarely have any qualms about starting at the particular (let's say for the sake of an argument, a descriptive language to categorise 'Livy') and working up to the abstract (a language to categorise 'Latin literature'). It may seem strange, but many of practical experience with modelling systems behaviour has taught me to be suspicious of abstract models: I've often found it's better to be specific with your model first, as they're nearly always unique.
I look at something like the TEI (the basis of the data in Perseus) and I see something much more familiar. A DTD/XSD/Relax NG format that describes a document structure, rather than a DTD that describes a document structure to describe a document structure (where 'document structure' here means a domain model): which in my world of pre-existing experience, has seemed to jump to the abstract far too quickly).
Normally, in the past, I would walk into a new business client and expect to be confused as all hell about their systems of business (their 'domain') but not all about my own. Here, I might have reasonably expected to have a good grasp of both domains, to a greater or lesser extent (I'm really a literature guy, not an archaeologist of any form, and in computing, I'm foremost a systems analyst and programmer, not for example, an operations guy or a database modeller), but I find I'm thrown into a loop!
Hence, my profound confusion at being profoundly confused! For which I must humbly ask your indulgence.
regards
Scot
|