[OAI-implementers] OAI-PMH & SOAP

[Hussein Suleman]

hi

now this has me wondering - don't resumption tokens solve the "safe 
list" problem ?

if a client requests a chunk of records, the server can choose to send 
fewer records and send back a resumption token, with retry-after headers 
  if necessary (thus the server is safe). the client, on the other hand, 
can choose to terminate a connection if the server sends too much data 
(the client is therefore safe). in fact the client should always be safe 
by virtue of the fact that the protocol is almost stateless and strictly 
client-initiated.

of course terminating the connection is only necessary for servers that 
are implemented without a scaling policy; and this solution is still 
somewhat cooperative, without proper dynamic negotiation.

maybe i am just missing the point - if i am interpreting the problem 
incorrectly, please clarify

thanx
----hussein

Walter Underwood wrote:

> --On Sunday, February 3, 2002 5:30 PM +0100 Thomas Place 
> <T.W.Place@kub.nl> wrote:
> 
>> Walter Underwood wrote:
>>
>>  > In a safe list protocol, the client requests a number of results,
>>  > and the server is allowed to return fewer. That way, both sides are
>>  > safe.
>>
>> This is already available in a comparable protocol under development
>> by the Z39.50 community: SRU - Search/Retrieve URI service
>> (http://www.loc.gov/z3950/agency/zing/srw.html#sru). SRU is SRW
>> mentioned by Herbert in his last message, where SOAP is replaced by HTTP.
> 
> 
> SRU/SRW/Zing is still a moving target. There is a quite a bit of
> discussion about the purpose of that protocol. The direction so
> far looks promising.
> 
> My point is that the current OAI list design is unsafe, not that
> I would like it to work differently.
> 
> LDAP, SRU, and every web search engine use the model I suggested.
> The model in OAI is dangerous to clients.
> 
>> I would applaud if the Z39.50 (http://www.loc.gov/z3950/agency/zing)
>> and the OAi initiatives were coordinated.
> 
> 
> Yes, though I would be happy if OAI focussed on metadata access.
> General search is a different problem.
> 
> wunder
> -- 
> Walter R. Underwood
> Senior Staff Engineer
> Inktomi Enterprise Search
> http://search.inktomi.com/
> _______________________________________________
> OAI-implementers mailing list
> OAI-implementers@oaisrv.nsdl.cornell.edu
> http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers


-- 
======================================================================
hussein suleman - hussein@vt.edu - vtcs - http://www.husseinsspace.com
======================================================================