> Date: Mon, 26 Sep 2005 22:08:51 +0100
> From: Robert Sanderson <[log in to unmask]>
> 
>>> >I'm confused about the //explain/serverInfo/database element and its
>>> Not quite.  If it were being used to describe a non Z39.50, non SRW, non
>>> OAI database, it might be just an identifier for the database or set of
>>> records.
>>
>> Okay, so it's more overloaded than I thought.
> 
> Right.  'database' is probably best described as: the part of a URL
> which identifies the database that isn't transport/host/port.  So in
> Z3950r/z3950s, it's the database name. In SRW, SRU, OAI, FTP,
> OpenSearch, etc, etc, it's the path in the URL. But if one were to
> somehow use IRC as an information retrieval protocol (well, you
> -could-, I guess) it would be the channel name.

Nice explanation.  You should pop it in the commentary.

>> Seeing that the description is unchanged in that draft formal
>> standards document, perhaps you might consider splitting the
>> secondary (SRW/U) context meaning into a distinct optional element
>> "path_info" or somesuch
> 
> At the moment it's easier to fix the prose than have all of the
> SRW/U implementations rewrite a mandatory element in their
> description.

The current version is also better than the proposed change: it just
needs clearer documenting (such as you have supplied above).

 _/|_	 ___________________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
)_v__/\  Cab driver to Groucho Marx: "Have a nice day"; Marx: "I'll have
	 the hell sort of day I want"