> > > But does that mean SRW should use a ZeeRex record as the SRW explain
> > > mechanism? Or should SRW define functions such as
> > > giveMeAListOfIndexNames() returning an array of strings
> > > giveMeIndexDescription(String name) returning description as string
> > See also Explain Classic which effectively did this, but via ASN/BER
> I think Explain Classic failed for different reasons. The main reason
> is its sheer complexity. There are so many fields! We have implemented
Oh I agree, but part of that complexity was that it wasn't just another
Z39.50 database, it had rules all of its own. The same applies here I
think -- if people can just implement one bit and get both (ie a normal
search will return the Explain info) then that's easier than implementing
an API interface.
> So I don't agree that Explain Classic can be used as an argument to
> say that APIs are the wrong way to go. In fact, Explain Classic was
> not an API based scheme. Each record had a deep and complex nested
> structure ... just like a ZeeRex record. :-)
Give me a list of index names is effectively what the attributeDetails
search is in Explain Classic. And while E.C. didn't go down to index
level, targetInfo is similar in scope to give me a description of this
index.
You need special handlers to do that sort of thing, and that is where the
complexity of E.C. came in. ZeeRex abstracts this behind an XML record,
but you can still do at least the first of these requests via scan. In
this increasingly XML world, I think (prod Sebastian) that there should
also be better document fragment support in Z for XML so that you /can/
retrieve just the index descriptions if that's what you're after.
Rob
--
,'/:. Rob Sanderson ([log in to unmask])
,'-/::::. http://www.o-r-g.org/~azaroth/
,'--/::(@)::. Special Collections and Archives, extension 3142
,'---/::::::::::. Twin Cathedrals: telnet: liverpool.o-r-g.org 7777
____/:::::::::::::. WWW: http://liverpool.o-r-g.org:8000/
I L L U M I N A T I
