RE: [ONT] Alternative property names : Normative or Informative

[moving to public list]

I believe the property identifiers should be normative, yes. The alternative names (for use as identifiers in various programming languages) should be informative, in the form of "best practices", and with the strength of "SHOULD" as per RFC 2119.

While we may be the authority on ontologies/vocabularies for device descriptions, we cannot claim to be the authority on the appropriate conventions for various programming languages. We are nevertheless reasonably experienced, so we offer some strong suggestions as information for those who will implement in those languages. Indeed, I would prefer not to mention programming languages in the ontology at all. Instead I would refer to naming conventions only. So the alternative names would be categorised as:

 "CamelCase"
 "lowerCamelCase"
 "Underscore_Delimited"
 "Space Delimited"
 "alllowercase"

We could even describe a simple algorithm which says that you take the "Space Delimited" name and replace the spaces by your chosen delimiter (or delete it) and then adjust the letter casing appropriately (e.g. lowering the first letter). I wonder if this would work for all naming requirements?

It would be up to individual implementators to decide which of the various conventions best suit their programming language.

Or they could completely ignore our advice, in which case they'd have to work out their own mappings to the normative names, and provide documentation to their users, who might be confused by the deviation from our suggestions.

---Rotan

-----Original Message-----
From: member-ddwg-request@w3.org [mailto:member-ddwg-request@w3.org] On Behalf Of José Manuel Cantera Fonseca
Sent: 18 June 2007 17:41
To: DDWG
Subject: [ONT] Alternative property names : Normative or Informative


Hi,

I think there is one more point to be discussed around the alternative 
names for properties. The point is that these alternative names caught 
by the ontology should be normative or informative.
I think that they should be informative but I would like to know other 
opinions from the group. If this is the case I think a resolution on 
this should be taken.
The only normative property names are the property identifiers in the 
ontology.

Best regards

Received on Monday, 18 June 2007 21:48:37 UTC