Re: Qualifier Combinations in CQL (Was: Bib-2 and the DC-Lib)

On Wed, Apr 24, 2002 at 09:21:54AM -0400, LeVan,Ralph wrote:
> > Another way to achive the above sort of thing is using the attribute
> > overlay scheme I talked about earlier. Its not perfect, but you could
> > express the above in (extended) CCL with our system as
> > 
> >     utilCreator= @personal(@agentDetail(Kernighan))
> 
> No!
> 
> The way to do that is to define an Index in an IndexSet somewhere where you
> combine the attributes that you want.  Then you would just say
> DCLIB.author=Kernighan and define in the DCLIB IndexSet the Index author as
> a combination of the AccessPoint Creator from XC, the SemanticQualifier of
> Personal from Bib-1 and the AgentDetail of Whatever from DC-Lib.

For this particular example, I agree this would a good and reasonable
solution. However this cannot be done for all attribute type/values.
<, >, >=, = etc is one example of where you want to change an attribute
type value without affecting anything else. Stemming I believe is another.

I don't actually know what personal and agentDetail is. But if they are
modifiers to the base attribute list, then I don't think its reasonable
to enumerate every single base attribute list with every single combination
of attributes (for example, I think it would be wrong to so PubDateEq,
PubDateGt, PubDateLt, PubDateGe etc). So I was not trying to say
personal is like >, but rather *if* personal was a modifier, then the
above *could* be used.

But I have just had my first go at a Bath conformant database definition,
and did not need to use @func() type stuff at all - just relations and
truncation.

Alan

Received on Wednesday, 24 April 2002 11:03:14 UTC