[OAI-implementers] Results for various sites trying POST requests

Tim Brody tim@tim.brody.btinternet.co.uk
Mon, 11 Feb 2002 12:21:04 -0000

Hi Alan,

I don't believe the issue of "naming and shaming" repositories has come up
before in OAI. I believe the theory is that the validation carried our
during registration should make sure that publicly-accessible repositories
are sufficiently error-free to be harvested.
(so I don't know whether your results reflect more upon repository's
implementations, or the OAI testing process ...)

What would be very interesting would be to correlate your results with how
OAI-PMH has been implemented at the repository - whether the admin has used
one of the available APIs, languages, and so on.
(It's a shame that many repositories seem to code their own implementation.
Perhaps OAI should be more pro-active in promoting the development of, and
use of standard APIs, including standard error messages)

Might I suggest an addition to your testing, that should be relatively
simple (from the last new problem I found with a repository):
Do records have an identifier and datestamp?

I noticed there were some problems with resumptionTokens - how does the
repository behave when its given escaped/unescaped resumptionTokens?

I guess the verb problem is related to POST (and a misleading error

The required metadataPrefix/resumptionToken problem is because the protocol
document isn't very clear. No arguments, apart from the verb, should be
given when a resumptionToken is used. This exclusivity is confusing when the
document says metadataPrefix is required.
(perhaps a note should be added, making this clear?)

All the best,

----- Original Message -----
From: "Alan Kent" <ajk@mds.rmit.edu.au>
To: "OAI Implementors" <oai-implementers@oaisrv.nsdl.cornell.edu>
Sent: Monday, February 11, 2002 1:08 AM
Subject: [OAI-implementers] Results for various sites trying POST requests

> I have built up a HTML page of interop tests of a simple client
> I was trying to write (ie: might be buggy from my end) when trying
> to POST to various other servers out there. I have not tried GET
> requests instead. I supply the VERB in the POST data (not on the
> URL) which I think caused many of the faults.
> There are several sites that I have not yet sucesfully completed
> a download of. A network problem or similar occurs before they
> finish, so I have to restart from scratch.
> If people are interested, the results can be found at:
>     http://www.mds.rmit.edu.au/~ajk/oai/interop.htm
> In the SOAP interop test bed (which I have been involved a little bit
> in), people with clients and servers tend to host such pages which
> has helped a lot in achieving interoperability. Nothing like airing
> dirty laundry to get people to clean up their act! :-) From my
> reading so far of OAI, I suspect there may not be the same push
> behind OAI.
> If I have missed a site, please let me know and I will add it.
> The table is currently manually maintained - I am still investigating
> OAI to see if its worth pursuing further. If people are interested
> (including me! :-), I could investigate generating such tables
> automatically including more detail (wire dumps). A sample page
> for SOAP interop is at
>     http://www.mds.rmit.edu.au/~ajk/soap/interop-results.htm
> Alan
> ps: The product I work on is currently undergoing a name change from
> SIM to TeraText, so I probably will mix up the names until I brainwash
> myself to the new name! :-)
> _______________________________________________
> OAI-implementers mailing list
> OAI-implementers@oaisrv.nsdl.cornell.edu
> http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers