[ZOOM] Some 1.4 options questions

Mike Taylor mike at indexdata.com
Tue Feb 3 12:17:16 CET 2004


> Date: Tue, 3 Feb 2004 18:36:01 +1100
> From: Alan Kent <ajk at mds.rmit.edu.au>
> 
> I just got the latest 1.4 abstract API spec from the web site to
> work out all the options (I hope its up to date). Reading it, I need
> to rework what I have done. Accessing unknown options and the X-*
> stuff needs a bit of work. Just pottering around.

Hi, Alan.  It's always profitable to hear from you, but rarely easy.
Why don't you ever ask nice, simple questions?  :-)

Yes, realistically v1.4 of the AAPI is what you, and all ZOOMers,
should be using now.  It's still labelled "IN PROGRESS", but really I
ought to promote it to Official, and start work on 1.5 "IN PROGRESS".

> * Why is there a 'host' option but not a 'port' option?
> * Why is there a 'proxy' option holding a host name, but no proxy port?

This is because the option-list I started with, and massaged into
what's currently in v1.4, is the one from YAZ's ZOOM-C documentation,
and ZOOM-C as it currently stands smashed "<host>:<port>" pairs
together into single strings.  That should at least be documented in
the AAPI; and perhaps it should just be changed.  Any thoughts, Adam?

> * Is it intensional that databaseName appears twice in the table,
>   once at the Connection level and once at the Record level. One is
>   a '+' separated list of databases to search, the other is the
>   actual database name returned in a record. I am guessing yes.

Yes, it's intentional.  But a case could be made that it would be more
intuitive to use a different name for the latter one.  How about
"recordDatabase"?

> Other stuff (which have been discussed before, probably without me fully
> understanding all the implications).
> 
> * The small set/large set stuff I assume most would not implement -
>   it would seem better to use the presentChunk and elementSetName
>   options and hide the protocol details from users.

I think so.  To be honest, despite Adam's having explained it more
than once, I still don't fully understand how these two sets of
options interact.  I have an email message from him on this subject,
buried deep in my mailbox, which will eventually be editied into a
part of the AAPI.

> * My reading of options and inheritance is if you do a search, the
>   Result Set object created will use the current element set name of
>   the Connection.

Yes.

>   So if you start a new query etc and change the element set name,
>   any old Result Set objects will be affected too.

Yes.

>   This feels bad to me. (Might be necessary for conformance, but
>   still feels bad.)

Sorry, that's how it is.  I think it's too late the change now without
a very compelling reason.  ("Feels bad" won't cut it! :-)

>   That way fiddling with the Connection class when there are
>   multiple Result Set objects wont do strange things.

"Don't do that, then."

> * The table of options and their scope may be clearer if it listed
>   for each option which classes can get and set the option.  Eg:
>   elementSetName can be get&set at Connection and Result Set.  The
>   only purpose at the Connection level is to provide a default for a
>   Result Set. But can you get it for a Record? Or do you have to
>   have the Result Set the record came from. (not stressed, just
>   wondering) Not a big deal though.

I take your point, but I am keen to the amount of verbiage down.  I
think it's clear from elsewhere in the spec. that anything that can be
set into an RS can also be set into a Connection, and conversely that
anything that can be read from a Connection can also be read from an
RS.

I suspect that the single greatest reason for the success of ZOOM is
that the AAPI document is so short.  My browser makes it 13 pages, as
compared with for example the OAI-PHM specification at
	http://www.openarchives.org/OAI/openarchivesprotocol.html
which actually says much less than the ZOOM AAPI but takes 28 pages to
say it.  Or compare with the fairly minimal Draft OpenURL standard,
which is available in three PDF parts from
	http://library.caltech.edu/openurl/Standard.htm
totalling 37+90+42 = 169 pages.

So my point is, of course, that I don't want to get into the thing
where the AAPI feels it has to point every little thing out
explicitly.

Thanks, Alan.

 _/|_	 _______________________________________________________________
/o ) \/  Mike Taylor  <mike at indexdata.com>  http://www.miketaylor.org.uk
)_v__/\  "Adjectives are like spices, to be used for flavour, with
	 care: not to kill the essential flavour of the sentence" --
	 Rheal Nadeau.

--
Listen to my wife's new CD of kids' music, _Child's Play_, at
	http://www.pipedreaming.org.uk/childsplay/






More information about the ZOOM mailing list