- From: Charles McCathieNevile <charles@w3.org>
- Date: Mon, 10 Jan 2005 09:33:07 -0500 (EST)
- To: Thomas Baker <thomas.baker@bi.fhg.de>
- Cc: Dan Brickley <danbri@w3.org>, "Dickinson, Ian John (HP Labs, Bristol, UK)" <ian.dickinson@hp.com>, public-esw-thes@w3.org
On the whole I find names that are derived from the end of the URI identifying a property are not very clear. Fortunately, in the design of a vocabulary language for RDF, smart people like danbri have given us the properties label and comment, which not only let us use a couple or three words as appropriate, but also make it trivially easy to provide these in multiple languages. And tools like Ontaria even use these properties to present things in a legible way. I would argue from long and mostly bitter experience that minimising URIs is something that is done for convenience, and that one should recognise that almost any label short enough to be convenient will be hard to understand for most users (possibly including the users of your vocabulary). The time taken to find a name that seems self-explanatory is, in my experience, much less than the time taken afterwards explaining it. All the more so if serious effort was put into thinking up the name. Using real words and comments, and examples, is easy enough even with meaningless names like "term1" "term2", "mp00a2e", etc. And valuable to users who have trouble asking you in IRC for an explanation. cheers Chaals On Mon, 10 Jan 2005, Thomas Baker wrote: > >On Thu, Jan 06, 2005 at 07:35:23AM -0500, Dan Brickley wrote: >> Some RDF vocabs do go down this route, with 'hasAge', 'hasName', >> 'hasHomepage' instead of 'age', 'name', 'homepage'. While >> backwards-named properties like 'subClassOf' complicate the picture, ... > >Dan, > >I agree about backwards-named properties, though it isn't >immediately obvious to me how one might rename a property >such as http://purl.org/dc/terms/isPartOf -- hasWhole? >Maybe that's what you mean by just "minimising" such names, >i.e., not avoiding them altogether. > >But I would argue that "hasAge" is more helpful than just >"age". ...
Received on Monday, 10 January 2005 14:33:07 UTC