Maybe.  But, as the service gets more mature, the changes should be smaller
and maybe the choice will become easier.  My hope is for rapid stability.

Ralph

> -----Original Message-----
> From: Matthew J. Dovey [mailto:[log in to unmask]]
> Sent: Wednesday, July 30, 2003 4:40 AM
> To: [log in to unmask]
> Subject: Re: 1.1 Versioning
>
>
> We can do that now, but that just delays the discussion. At
> some point we will need a smooth versioning transition.
>
> Matthew
>
> ________________________________
>
> From: Z39.50 Next-Generation Initiative on behalf of LeVan,Ralph
> Sent: Wed 30/07/2003 04:19
> To: [log in to unmask]
> Subject: Re: 1.1 Versioning
>
>
>
> I think we said from the beginning that version 1.0 would be
> replaced, not
> upgraded.  I don't think we ever promised smooth upward
> compatibility.  The
> first version was an experiment that would lead, within one
> or two more
> versions at most, to a stable final version.
>
> Let the 1.0 clients go.
>
> Ralph
>
> -----Original Message-----
> From: Robert Sanderson [mailto:[log in to unmask]]
> Sent: Tuesday, July 29, 2003 5:23 PM
> To: [log in to unmask]
> Subject: Re: 1.1 Versioning
>
>
> > The problem with doing this is if a 1.1. client send such a
> request to a
> > current 1.0 server, the server will return an error. This
> also includes
> > any request to get the explain record, so we can't work out
> the version
> > from that (or just keep trying different versions until we get
> > something?)
>
> Why can't we have version 1.0 servers and version 1.1 servers?  If a
> version 1.0 server gets a 1.1 request it's going to fail it
> at the toolkit
> level because it doesn't conform to the spec.  1.0 talking to 1.0 will
> still work, as will 1.1 to 1.1
>
> The only advantage is a 1.0 client can still successfully
> send to a 1.1
> server. At the expense of interoperability nightmare of having to
> explicitly namespace every single element in the both request and
> response.
>
> Clients are easy to write and this is simply a matter to tweaking the
> namespace up to http://.../srw/1.1/ and everything will go back to
> working as expected (with the exception of Explain, and do any clients
> self configure from it apart from my XSLT based one anyway?)
>
> Are servers expected to support 1.0 and 1.1 at the same time?
> So if you
> get a 1.0 request for explain you send back the xml, and if
> you get a 1.1
> request you send back the record wrapper around it?  This
> seems a rather
> unuseful proposition to me, considering the current
> implementors are all
> on the list (as far as we know)  It's not like HTTP 1.0 and 1.1 where
> there are a bazillion clients already in existence.
>
> What do other people think?
>
> Rob
>
> --
>       ,'/:.          Dr Robert Sanderson ([log in to unmask])
>     ,'-/::::.        http://www.o-r-g.org/~azaroth/
>   ,'--/::(@)::.      Special Collections and Archives, extension 3142
> ,'---/::::::::::.    Nebmedes:  telnet: nebmedes.o-r-g.org 7777
> ____/:::::::::::::.                WWW:
http://nebmedes.o-r-g.org:8000/
I L L U M I N A T I