Thread: Ambiguity in description of ws/availability

Started: 2012-10-06 05:02:20
Last activity: 2012-10-06 06:48:00
Topics: Web Services
Joachim Saul
2012-10-06 05:02:20
Hi,

in http://www.iris.edu/ws/availability/ there is an ambiguity with respect to the correct value for the parameter "output" if bulkdataselect list format is requested.

Under "Query Usage:" there is

output-options :: [output=<xml|bulk|query>]

whereas below in the table it is not "bulk" but "bulkdataselect".

Which is correct? The IRIS implementation appears to accept both but IMO in the specification it should either be one of these or it needs to be stated explicitly that both "bulk" and "bulkdataselect" have to be accepted by the server. I don't want to split hairs but in a real world the current specification might lead to compatibility issues.

Cheers,
Joachim

  • Chad Trabant
    2012-10-05 21:16:47

    Hi Joachim,

    Thanks, I've changed the query usage part to "bulkdataselect" so it is consistent through documentation.

    As you say this might cause problems, but only if there is more than one implementation. Are you planning on implementing this interface? Our long term plans are to include data availability into an FDSN version of the station metadata service (current under development), when this is new service is established it will be a superset of the functionality provided by the current ws-station and ws-availablity.

    cheers,
    Chad

    On Oct 5, 2012, at 1:02 PM, Joachim Saul wrote:

    Hi,

    in http://www.iris.edu/ws/availability/ there is an ambiguity with respect to the correct value for the parameter "output" if bulkdataselect list format is requested.

    Under "Query Usage:" there is

    output-options :: [output=<xml|bulk|query>]

    whereas below in the table it is not "bulk" but "bulkdataselect".

    Which is correct? The IRIS implementation appears to accept both but IMO in the specification it should either be one of these or it needs to be stated explicitly that both "bulk" and "bulkdataselect" have to be accepted by the server. I don't want to split hairs but in a real world the current specification might lead to compatibility issues.

    Cheers,
    Joachim
    _______________________________________________
    webservices mailing list
    webservices<at>iris.washington.edu
    http://www.iris.washington.edu/mailman/listinfo/webservices



    • Joachim Saul
      2012-10-06 06:48:00
      Hi Chad,

      Chad Trabant [05.10.2012 23:16]:
      Thanks, I've changed the query usage part to "bulkdataselect" so it is consistent through documentation.

      That was quick, thanks!

      As you say this might cause problems, but only if there is more than one implementation. Are you planning on implementing this interface?

      It exists already: an IRIS-compatible WS implementation for SeisComP. We currently use (and thereby test) it internally (with ObsPy on the client side) and that's how I came to notice this issue in the specification. Whether or not there will be a public GEOFON WS using the IRIS interface is a different issue and not yet decided.

      Our long term plans are to include data availability into an FDSN version of the station metadata service (current under development), when this is new service is established it will be a superset of the functionality provided by the current ws-station and ws-availablity.

      OIC. We will certainly follow these developments with interest.

      Cheers,
      Joachim


Page built 04:39:34 | v.88350d0a