DASL and GILS alignment

Hello, I've just joined this list. If you'll excuse my 
brashness as a newbie, I would like to offer some thoughts 
about DASL and GILS alignment.

I understand there has been some discussion here about 
possible overlap of DASL with Z39.50 and the STARTS work. 
In view of the hugely valuable content accessible through 
Z39.50, interoperability with Z39.50 would be a huge win!

If you are not familiar with Z39.50, it grew up over the 
last ten years primarily in the library and information 
services communities. It is often used inside major services 
that charge for use, but you can see a sampler of free 
Z39.50 Internet resources at <http://www.gils.net/list>. 
Some of these resources are themselves referral sites to 
dozens of other Z39.50 resources in nations throughout the 
world.

One of the often-stated objections to Z39.50 is its use 
of a binary encoding on the wire. This encoding follows 
from using Basic Encoding Rules (BER) in association with 
the fact that the Z39.50 protocol is specified in Abstract 
Syntax Notation (ASN.1).
   
At the QL '98 Workshop, I presented work over the last 
month that resulted in a proof-of-concept for representing 
Z39.50 in XML using XER (XML Encoding Rules), described at 
<http://asf.gils.net/xer>. With this new development, it 
becomes quite reasonable to look at making DASL interoperable 
with Z39.50. (The XER is still formative--we definitely need 
help to address XML namespaces, appropriately use element 
attributes, and how best to internationalize the XML tags.)  

Another common objection to Z39.50 is that is a very complex 
protocol with many options. To address this issue, those of 
us involved in the Global Information Locator Service (GILS) 
developed a simple subset GILS Profile. GILS-compliance 
involves a minimal set of commonly used attributes much like 
those adopted in STARTS and DASL. 

The GILS Profile of Z39.50 requires just 3 services (Init, 
Search, and Present). For those applications where statefulness 
is infeasible and network round trips must be minimized, it is 
possible to collapse the messages into a single "Init-less 
Search with a piggy-backed Present".

My intent here is just to suggest the possibilities I sense.
There are many Z39.50, GILS, ASN.1, STARTS, XML, and DASL 
experts who can do a much better job of elucidating the 
technical issues--and those of you on this discussion list 
know this already :-)

Thank you...
   




Eliot Christian, US Geological Survey, 802 National Center, Reston VA 20192
echristi@usgs.gov  Office 703-648-7245  FAX 703-648-7112  Home 703-476-6134

Received on Friday, 11 December 1998 07:53:16 UTC