> -----Original Message-----
> From: Mike Taylor [mailto:[log in to unmask]]
> Sent: Wednesday, January 07, 2004 5:40 AM
>
> In all HTTP versions except the broken first one, every HTTP request
> and response explicitly states its version. In the same way, in every
> SRW version except the broken first one, a client MUST explicitly
> state what version of the protocol it is using; and a server MUST
> include in the response the version of the protocol that it's using.
> These must be mandatory in the good old-fashioned sense of "may not be
> omitted", rather than the "Mandatory Lite" sense of "you can leave it
> out and the server can do what the hell it likes" (i.e. what we used
> to call "optional"). Otherwise we're digging ourselves a hole in
> quicksand.
Sigh.
You are absolutely right.
The point really is to handle the few people, like me, who hand enter SRU
URL's. No real user will ever be inconvenienced by a <input type="hidden"
name="version" value="1.2"> statement hiding in a form somewhere.
Programmers who generate URL's in their code will hard-code the value and
forget it.
Now, I also hand enter http requests through telnet sessions (sometimes you
really need to see the raw data on the line) and remembering the 1.1 in the
request IS a pain, but it's also the rule and I do it.
Maybe the issue would go away if someone put together a generic SRU
interface that I could use instead. I keep meaning to put something
together, but it has been too low a priority for me to do.
Anyway, I pledge to be conformant with the mandatory nature of the version
parameter.
Sorry for having contributed to the mini-furor. The good thing about
digging a hole in quicksand is that they tend to fill themselves in quickly.
:-)
Ralph
|