Re: [Koha-zebra] .abs file and subfield ordering

[Sebastian Hammer]

Joshua Ferraro wrote:

> On Tue, Feb 14, 2006 at 11:47:09PM -0500, Sebastian Hammer wrote:
>  
>
> > Of course MARC must die... but I don't think this is a major issue. I've 
> > had extensive conversations with staff at the LoC about Zebra's 
> > weaknesses as pertains to MARC, so I'm acutely aware of the weaknesses, 
> > but I'm not sure this is one of them. Most of their concerns have 
> > concerned the ability to combine multiple subfields into one phrase 
> > index for scanning or complete subfield searching... I've never heard of 
> > a wish to control indexing based on what subfield follows what other 
> > subfield... others on the list may have other experiences (indeed, other 
> > formats than MARC21 may pose other challenges here).
> >    
> OK ... I won't worry about the ordering prob then ... with regards
> to the multiple subfields in one phrase search, I have actually 
> been saving that one :-). A while ago I found a thread on the lists:
> http://lists.indexdata.dk/pipermail/zebralist/2005-August/000875.html
>
> That seems to indicate there is a solution for that -- or am I
> reading it wrong?
>  
Some Russians have developed functionality that allows you to combine 
subfields into an index and do other fancy stuff. I haven't looked much 
at it myself.

The up-and-coming XSLT-based indexing system will allow *any* kind of 
crazy logic on the records to support indexing, including mapping the 
whole damn thing to sound like a pirate, aargh.. It should be easy 
enough to write some perl to map an old-style .abs to an XSLT-based 
filter if/when the need arises.

--Seb

>  
>
> > > Also, speaking of mappings, I've got a few CQL questions. So I 
> > > understand the notion of a 'context set' and 'indexes' within
> > > each context set. I'm not clear on what the best context set
> > > would be for the MARC records in the libraries using Koha. bath?
> > > cql? 
> > >
> > >      
> > I'll defer to Mike, since he's on the editorial board. I imagine this 
> > might be something that'd be good to bring to the ZING list.
> >    
> Right ... I'll do that then.
>
>  
>
> > > Also, though the bath context set defines indexes, it doesn't clearly
> > > specify mappings for those indexes into any specific record format
> > > like MARC. Are there specifications anywhere that define such 
> > > mappings? Bib-1 maybe? MODS? Any suggestions?
> > >
> > > Finally, in some of the ABS files included with Zebra I see a
> > > ? after the tag like in this entry:
> > >
> > > elm 245                 title           -
> > > elm 245/?               title           !:w
> > > elm 245/?/a             title           !:w,!:p
> > >
> > > Is this a 'WildThing'? could someone explain what that means? :-)
> > >
> > >      
> > I think Bath has been roundly criticized for stopping short of mapping 
> > USE attributes (to say nothing of indexes) to MARC fields. The Bib-1 
> > semantics document makes a half-baked effort, but I don't think anyone 
> > would consider it an authority today. I've heard some folks say that 
> > either the US national profile or the ZTexas profile actually specify 
> > mappings to MARC fields for the basic stuff.
> >    
> Thanks, I'll check those out.
>
>  
>
> > But the marc21.abs file that comes with Zebra should be a really good 
> > starting point for the most simple and obvious stuff. I developed that 
> > over several iterations with Larry Dixson of the LoC, and it was 
> > directly based on a set of requirements that they developed for their 
> > LIS vendor when they migrated to Voyager (of course, the vendor totally 
> > ignored them). And the folks at the LoC sure do know their MARC.  :-)
> >    
> OK ... I'll check this out as well.  Thanks!
>
> Cheers,
>
>  

--
Sebastian Hammer, Index Data
address@hidden   www.indexdata.com
Ph: (603) 209-6853