RE: Init is dead?

No Init is not dead yet. Not for me at least.
I do not agree with the idea that the Init contains only useless
information.
The PreferredMessageSize negotiation is very useful.
Most of our applications use the authentication through the Init.

I haven't been in Washington, so I don't know what the discussions were
about, but personally I don't think we gain much this way.
Does the omission of Init imply that there will not be a Present anymore? I
have the impression. No session. No ResultSet anymore.

The people who complain that Z39.50 is too complex, will always find it too
complex even we omit the Init. (The attribute set structure is quite complex
compared to the Init.)
Anyway who cares, the Init is only 50 bytes, and it's easy to implement.

People who want to do Internet-like searching can do that already now:
For every search you do Init Search Present Present etc. and you close the
connection.
That's easy, we use that here a lot.

Pieter van Lierop


> -----Original Message-----
> From: LeVan,Ralph [mailto:levan@oclc.org]
> Sent: Friday, January 12, 2001 7:35 PM
> To: 'ZIG'
> Subject: RE: Init is dead?
> 
> 
> I agree that we'll be waiting a while for this all to be 
> formalized, but I
> think we've now authorized some experimentation.
> 
> Ralph
> 
> > -----Original Message-----
> > From: Ray Denenberg [mailto:rden@loc.gov]
> > Sent: Friday, January 12, 2001 12:01 PM
> > To: 'ZIG'
> > Subject: Re: Init is dead?
> > 
> > 
> > "LeVan,Ralph" wrote:
> > 
> > > Wow!  Was my argument for making Init optional so 
> > persuasive that you're all
> > > convinced and it is now a dead subject?
> > 
> > Yes as a matter of fact I'm prepared to support the idea, 
> > based both on your
> > reasoning and Bob's.  Re-evaluating my earlier argument, 
> the case for
> > encapsulation is that when you actually do need Init, 
> > encapsulation enables
> > single-round-trip. However, I'm prepared to concede that my 
> > agrument that
> > encapsulation means that you should include Init even when 
> > you don't need it was
> > overstated.
> > 
> > Note of course, when we're talking about making Init 
> > optional, we're talking
> > about version 4.
> > 
> > --Ray
> > --
> > Ray Denenberg
> > Library of Congress
> > rden@loc.gov
> > 202-707-5795
> > 
> > 
>

Received on Monday, 15 January 2001 06:27:42 UTC