Re: SV: Z39.50 on the web (and in print)

Rob,

with explain being there for all those years in the 1995 standard:
- there has been an explain test bed driven by Makx back in 1995 - 96 - result - very little participation or usage;
- there was the ONE project - resulting in a handful of explain-aware servers in existence since 1997 but almost no usage;
- we had a completely free explain add-on component on our ftp server for over 5 years - only a handful of downloads;
- the number of commercial library management systems supporting explain in their Z servers or clients is probably none;

- all in all a huge amount of nervous energy - enthusiasm from ZIG folk - discussion over the past years  etc but a  tiny, tiny takeup.

To me that says something -  despite how easy it might be to portray or use or say how useful it would be. 

From a commercial point of view - I see no justification in creating an explain server or an explain client when you can 
use alternative options  - ie registries, phone, web sites, robots etc to find the same information. In theory it sounds nice - in practice people must be saying "why bother" when I can currently get away with alternatives. The alternatives are not perfect - but they get the job done.

Even if there was enough ZIG energy to do explain properly via a profile etc. - chances are that someone will want the ASN changed
and make it different - or realise that there are things missing reflecting today's technology needs that will result
in a re-write for toolkits, and the very few existing applications etc.

We know that explain lite is not perfect - and it was argued against as "madness" because explain was there.
But at least explain lite can be used far easier than explain - it is human readable - it can be on the Z-init and/or on a web page  - it does not need any ASN modifications - it uses flavour of the decade technology - XML.

Explain lite is not - or ever intended - to cover all the fine details of the complex servers - it is there to try and please some of the people some of the time.

And in 2 years of explain lite existing there is probably more usage than there ever were explain servers.  For the requirement of improving automation and configuration - that to me is a good start.


Rob

-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-
  Rob Bull    bull@crxnet.com         Crossnet Systems Limited
  tel +44 (0) 1635 522912             Unit 41 Bone Lane, Newbury
  fax +44 (0) 1635 522913             Berkshire, RG14 5SH, 
 United Kingdom
-+-+-+-+-+-+-+-+-+-+-   A member of the DS Group  -+-+-+-+-+-+-+-+-+-+-+-
  ----- Original Message ----- 
  From: Robert Sanderson 

  To: Mike Taylor 
  Cc: www-zig@w3.org 
  Sent: Friday, February 22, 2002 2:12 PM
  Subject: Re: SV: Z39.50 on the web (and in print)




  > The obvious way forward, then, is for us all to swallow our
  > computer-scientific objections and jump aboard the Explain-lite
  > bandwagon, pausing only to cast a swift regretful glance back at
  > Explain Classic as it gets dumped in the same bin as Betamax and,
  > maybe one day, non-MS operating systems.

  I disagree entirely. (  Especially about non MS Operating systems :)  )
  Explain Lite is woefully inadequate and needs rewriting from the ground
  up.  It demonstrates fundamental lack of knowledge about the way XML
  works, and can't describe any non trivial Z servers.  See:

  http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0111&L=dner-tech&F=&S=&P=301

  For a more complete discussion, schema/dtd/etc, email me.

  Note Well: My initial reaction is to follow -the standard- and use Explain 
  proper.  As this is not likely to get anywhere, the only other option IMO 
  is to have something that does all of explain's job but can be put into a 
  normal Z39.50 searchable database.  


  > (To be clear: I'm not arguing that Explain-lite is a perfect
  > solution.  I'm not even saying it's a _good_ solution.  But it's one
  > that has the very useful property of actually existing :-)

  So does the nuclear bomb. Doesn't mean it's something we should use ;)

  We have successfully built just the system described using XplainML.
  http://gondolin.hist.liv.ac.uk/~cheshire/warm/

  This locates servers that support the query types, how they map the query 
  attributes, then fans out the searches to them.
  This is just a prototype/demonstration, and still needs a lot more work, 
  but it does exactly what the original post was about. In our practical 
  experience, this could not be done with Explain Lite.

  Rob


  -- 
        ,'/:.          Rob Sanderson (azaroth@liverpool.ac.uk)
      ,'-/::::.        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