Skip to content

report skypeconf 12 July 2012

hamial edited this page Jul 12, 2012 · 9 revisions

Participants:

Dario Fabbri (FAO), Jacques Joyeux (FAO), Erik van Ingen(FAO), Matt Nelsson (Metadata Technology), Jens Dosse (OECD), Stratos Nikoloutsos (Agilis), Spyros Liapis (Agilis), Al Hamilton (ABS), John Csaki (ABS).

SDMX-JSON

Jens expressed that SDMX-JSON is very high on the prio list of OECD, the .Stat platform, ABS and Statistics NewZealand because of the open data initiative.

CORS versus JSONP

An ABS partner has chosen the CORS part. JSONP could be less secure. Additional info from Jens: CORS relationship to JSONP: CORS can be used as a modern alternative to the JSONP pattern. While JSONP supports only the GET request method, CORS also supports other types of HTTP requests. Using CORS enables a web programmer to use regular XMLHttpRequest, which supports better error handling than JSONP. On the other hand, JSONP works on legacy browsers which preclude CORS support. CORS is supported by most modern web browsers. Also, whilst JSONP can cause XSS issues where the external site is compromised, CORS allows websites to manually parse responses to ensure security. The problem of lack of support for CORS in browsers will disappear over time. In SDMX-JSON the use of CORS is suggested.

Concepts to be used as header information in the SDMX-JSON message

  1. Title (concept 63) ; 2. Comment (concept 10); 3. Source of data (concept 56) ; 4. Contact (concept 14) ;5. Last update date (under concept 27);6. Observation value (concept 40);7. Time period (concept 61);8. Unit of measure;9. Unit multiplier (concepts 64 & 65);10. Release calendar (or date of next update – concept under concept 51);11. Adjustment (concept 3);12. Reference area (concept 49); 13. Base period (concept 3); 14. News release (concept 30)

The message can have a header. The header gives high level info. Other data can be used Alistair will check why the URI is missing in the COG.

Non-IDitic JSON representation

There are 3 possible message formats: A) Array approach. This does not group the data. Added structural metadata. Is similar to the json-stat approach. There are many ways of describing. Matt did index the observations. Actual structural metadata is held as a separate object. B) Compact message groups the data. C) Compact message based on Odata. Matt worked on A) Array approach: Goes back to our priorities. Webdeveloper or the message as small as possible. The visualization tool will decide on how to slice the data. Jacques: the cpu is a factor, the simpler the better. Matt: CPU is important. Include a library, the underlaying format does less matter. We need to start testing on how responsive the different formats are. Jens: It might not always be possible or wished to use such a library (e.g. dataexchange usecase). Erik: Can see in the future also a replacement of XML by JSON for the dataexchange. Matt: In that case we need to talk how JSON relates to streaming. The meeting agreed that the SDMX-JSON group would primarily focus on the data visualization usecase, the dataexchange usecase can be discussed later.

Multiple datasets in a single message?

This topic was discussed without a clear decision. There was a tendency towards simplicity, therefore for the option of having only 1 dataset in a single message.

Index on the dimensions

Matt has proposed to index the dimension in order to be able to deal with parse Cubes. Should this be used, and if so, should it be optional or not? Stratos will investigate more whether there is a clear problem of sparse data or not. Jens: Having such an option increases complexity, and web clients and services would not necessarily know which format would be more appropriate for a specific dataset. It was recommended that the performance should decide whether it would be useful to use an index or not. Matt proposed to write client code to test performance. Jens offered to use that test code to test performance on larger OECD data.

Next meeting

Skype conference call on the 31th of August 2012 14:00h - 15:00h (Roman time) The suggestion is that the SDMX-JSON group meets in Paris on the 12th of September, just before the SDMX expert meeting on the 13th and 14th.