- From: Kevin Gamiel <kgamiel@islandedge.com>
- Date: Wed, 21 Jun 2000 13:36:38 -0400
- To: Robert Waldstein <wald@library.ho.lucent.com>
- CC: www-zig@w3.org
I wasn't at the meeting either, but that's never stopped any of us from commenting before:-) How about well-known database(s) that serve xml metadata (RDF?) about known servers, profiles, etc? IMHO, we need to stop forcing so much into the core protocol and use the transport mechanisms we have in place for higher order purposes. This is a good example of that. In the proposed scheme, I assume that the client has to request the information in the init request? If not, I definitely disagree with that sort of information being forced on the client. If so, that is a burden on developers, having to get their hands dirty with ber (a real pain for those who rely on third-party libraries). In my proposed scheme, *any* client can at least search and retrieve the information without being touched, and usable by the human. Of course, more advanced clients would tightly integrate the response records into the client functionality. Just a thought, Kevin Robert Waldstein wrote: > > > Review of the discussion in San Antonio: We need a mechanism for a client to > > dynamically discover Z39.50 servers. We briefly considered "broadcasting" and > > "robots", but these ideas were passed over in favor of a simpler mechanism, a > > "distributed Z-directory". The approach is to define a structure that a Z39.50 > > server may include within the Init response -- an otherInfo item that points the > > client to one or more servers. The client may then initiate associations with > > these servers, possibly for the sole purpose of discovering more servers, > > I tried to avoid commenting - given that I wasn't at the meeting, nor am > willing to read the minutes (sorry). But this bothers me. What I understand > is my init may get back a complete list of all the servers/databases in > the world, unsolicited, assuming the server wants to. Yuck. > > Why is more and more being dropped in init - because EXPLAIN is still not > being implemented? I have already had one server blow out my client cause > they upped the buffer size to send me a large init - full of stuff I don't > want and will ignore. > > Re the discovery being discussed above - assuming you don't want to use > EXPLAIN, I would say don't use z39.50; how about every HTTP server in the > world have a http://server/z39.50databases.txt -- okay, not serious; but not > happy with the init solution (don't care about the XML issues). > > You may take this as an uninformed rant, but really doesn't seem like > a good idea to me of adding this discovery info to potentially every Z39.50 > init... > bob > -- > Robert K. Waldstein Email: wald@lucent.com > Bell Laboratories, Room 3D-591 Phone: (908) 582-6171 > 600 Mountain Avenue > Murray Hill, New Jersey 07974 -- Kevin Gamiel Email: kgamiel@islandedge.com Island Edge Research, Inc. http://www.islandedge.com Kill Devil Hills, North Carolina Phone: +1-252-449-8969
Received on Wednesday, 21 June 2000 13:46:20 UTC