[OAI-implementers] Empty ResumptionToken

Tim Brody tim@tim.brody.btinternet.co.uk
Mon, 8 Jul 2002 12:35:54 +0100

----- Original Message -----
From: "Alan Kent" <ajk@mds.rmit.edu.au>

> On Fri, Jul 05, 2002 at 12:16:51PM +0100, Tim Brody wrote:

> > an empty resumptionToken is just plain wrong ("The format of the
> > resumptionToken is not defined by the OAI-PMH and should be considered
> > opaque by the harvester", but OAI-PMH would appear to "define" an empty
> > resumptionToken - as Homer puts it, "Doh!".)
> I thought the reason for allowing resumptionToken to have empty content
> was so that the attributes of the element could always be present:
> expirationDate, completeListSize, cursor. That is, so the last request
> was not a special case where these attributes would be missing.

I think you are correct that an empty token was introduced to allow
cursor/list size in the final incomplete list.

The problem is this is altering the use of a mandatory component (resumption
token) for an optional one, cursors, list size. As I said in my previous
mail I don't think cursors/list size will be used in the majority of cases,
so an empty resumption token (which just looks odd) is mandated to account
for the minority situation.

Perhaps, in hindsight, mixing cursors and list sizes with resumption token
was not the best solution?
(It always struck me that putting a cursor at the end of the response, which
is actually the value at the beginning was slightly odd)

<ListRecords cursor="50" completeListSize="200000"> anyone?