Re: TAP KB naming rules / names carrying (transitory?) type info

> I notice that TAP names often carry the (or a) type for the named thing as
> part of their name / ID.
> 
> I can see many cases where this might be a very useful optimisation, but
> wonder about the consequence for long-term use of these IDs.

It wasn't intended as an optimization, it was simply a convenience; we had
to pick something and that's what we picked. We were working largely with
public figures, and products, in those days. These things do not change type
as often as other things. 

> I'm currently for eg listed in TAP as
> 
> http://tap.stanford.edu/cgi-bin/kb.pl?node=W3CPersonBrickley%2C_Dan&op=show&syn=10&browse=all
> 
> ie Resource ID: 'W3CPersonBrickley,_Dan'
> 
> I don't expect to be a W3CPerson my entire life. What's the TAP plan for
> identifier persistence? Should I expect to stop using
> W3CPersonBrickley,_Dan as a tap-name for me, if/when I eventually move on
> to life after W3C? Will TAP-based systems be encouraged to unpack these
> structured names and conclude that anything named W3CPerson*,* will have
> certain characterstics (such as working for W3C)? In which case I'm wary
> of using your ID for me, since it'll cause no end of problems a few years
> down the line.

No, the ID's are most definitely NOT intended to be automatically packed or
unpacked. They're human readable for the same reason that URLs are human
readable, to facilitate easy copy and paste, and to enable them to appear
in print without looking like gibberish.

It probably would have been more consistent to pick 
ComputerScientistBrickley,_Dan since that's less likely to change than the
fact that you're at the W3C.

> Sounds like a cataloguing rules and 'best practice' issue rather than a
> deep technical problem. Filing me as PersonBrickey,_Dan would seem less
> likely to cause semantic bit-rot...

This is probably a better long-term plan.

> What's the plan? How have things worked to date?

Thus far we've been largely working with celebrities and popular figures,
whose types don't change all that often. So it hasn't come up just yet.

There is another question here, which is the persistence of facts in TAP.
TAP has no notion of time or former roles. If I want to represent Henry
Kissinger as a former Secretary of State, there isn't any notion of this
at the moment. You're either the Secretary of State, or you're not. TAP is
still evolving so we're creating conventions for these sorts of things
as they come up.

Received on Monday, 23 September 2002 15:36:36 UTC