RE: Explain-lite (Explain discussion)

Thought it about time I threw my penny's worth in!!

Whilst I agree with Sebastian's view that Explain is the way it is for a
reason and probably has a number of advantages over Explain-lite, E-Lite has
the advantage that it is being used whereas however technically good Explain
maybe it just hasn't been adopted (strains of O/S2 vs Windows, BetaMax vs
VHS, etc.). I think part if this is that Explain by addressing the more
complicated requirements is too complicated for the simple application which
is what most people (at least initially) would want to implement. E-Lite
fills this need as in can be implemented quickly without major overhaul of
the Z39.50 server. I don't think anyone intends Explain-Lite to be a
replacement for Explain but perhaps we need to investigate more closely the
relationship between them (where one should be used rather than another,
ensuring that one is a true subset of the other etc.)

Sebastian is his comments has indicated that some of the complexity is there
for a reason. Now, it maybe that we need to reinvestigate those reasons, and
not be afraid to throw something away if it isn't working - take the work on
the attribute architecture as an example of where the Z39.50 standard has
been reworked. However, was is missing from the standard is an historic
perspective. To someone trying to work with the consequences of these
decisions there is no clue as to way things are the way they are. I have an
e-mail from Ray somewhere over some other issue in which he replied
(roughly) "we don't know. The people who made the original proposal probably
had some need for it, but they have probably long left the ZIG, and no-one
nowquite remembers what the reasons were". I think a lot of the protocol
suffers from this, and we aren't going to developed the standard for the
better if we don't have this perspective.

This brings me to my final point. It, I think, has been recognised that for
Z39.50 to survive we do need to rebadge and market it better. If we sit in
our ivory tower, proclaiming ourselves to be the experts with 20 years
experience, the world will just pass us by. Z39.50 can be daunting, but to
promote it we need to explain why the standard is the way it is - just
saying we know best isn't the way. Others will think they know best and go
their merry way (take XML and SGML as an example). 

We also need to recognise that there is a world outside Z39.50. Sebastian
raises the point of having a Z39.50 explain server to act as a directory of
Z39.50 services. Laudable, but there are other services beyond Z39.50 and
various activities building directories of library services (including ISO
ILL, and NISO CIP) and e-commerce services. Explain-lite currently has the
advantage that it can be used outside the Z39.50 standard - mounted on a web
page, stored in an LDAP directory of library services or located in a UDDI
directory of e-services. It recognises that we need to fit Z39.50 into the
world amongst other standards and initiatives not fit the world into Z39.50.

Matthew

Received on Sunday, 19 November 2000 09:51:12 UTC