Re: Proposal: make init optional

Bob

how would you propose that some of the negotiation about buffer sizes and
services supported (plus protocol version) be handled if there is no init
- Im not necessarily opposed to this - just interested in how you see
those types of things being handled

mark


On Mon, 8 Jan 2001 wald@library.ho.lucent.com wrote:

> 
> 
> Coming out of the ZIG meeting decided it was time to propose this; that
> is make the init PDU optional.
> 
> why: two reasons.
>   1. by far the main one is that it would allow a good quality Z39.50
>      session to be done in one round trip. THat is, a search PDU to
>      the server, and a search response with piggy-backed results.
> 	Note this would closely resemble an HTTP search, but with a
>      high quality search and retrieval specification.
> 
>   2. the second reason is minor, and it is that it is not needed for
>      many (all?) sessions - and is yet another hold on to "older"/OSI
>      thinking.  I believe by the way (but no assurances) that my server
>      does not require an init - but certainly might break it.
> 
> 
>   Note this change requires little or no change to the actual protocol -
> the one curious question is what a server that requires an init does when
> it receives a search PDU without an init? Choices I can think of are:
>   1. send an init failure PDU
>   2. send a search failure PDU
>   3. Be a good OSI implementation and shut down.
>   4. Or I sort of like: send a new generic failure PDU that replaces all
>      failure PDUs.
> 
>   bob waldstein  wald@lucent.com
> 
>

Received on Monday, 8 January 2001 10:02:47 UTC