WG4, the group tasked with maintaining IS 29500, met last week in Helsinki. This was my third WG4 meeting, so I think I’m starting to get the hang of it.
If I had to choose the topic that dominated this WG4 meeting, it would be extensions. This is a hard problem: standards compliance shouldn’t stifle product innovation, but standards bodies also shouldn’t feel pressed into constantly just updating standards to reflect implementations. How to allow implementers to make fun stuff while still making the standard a collaborative effort with the appropriate reviewing and authoring procedures?
We discussed the topic very thoroughly – we talked about how to extend the standard without breaking conformant applications, and we talked about some extensions which looked like good targets for standardisation, and how WG4 might go about standardising them. As an implementer who’s already used the extension mechanisms built into the standard, we (Microsoft) are very interested in working out what the right thing to do is here, and I hope that we (WG4) can come up with a methodology which will stand the test of time. Extending any standard without breaking conformance for existing documents is not an easy feat and, although I think IS 29500 is designed well for this, it’s a complicated issue and there is a lot of talking still to be done. The Japanese National Body will soon propose the creation of a new multi-part standard just to contain extensions to IS 29500, so it’ll be interesting to see what happens here. We made some good progress this week – Japan’s NWIP will be under ballot when WG4 next meet in Tokyo so we won’t be able to discuss its content, but we’ll still have plenty to talk about because of this last week’s second main topic: dates.
Some background: Spreadsheet applications (and operating systems) generally use “serial date values” to track dates at runtime – these values represent a number of days since a particular historic date (the “epoch”). IS 29500 allows you a choice of two epochs – December 30th 1899, or 1st January 1904. To see serial dates in action, load your favourite spreadsheet app, type “40349” into cell A1 and format it as a date. It’s 20th June 2010, or 40349 days after the 30th of December, 1899. IS 29500 also allows you to turn on or off the mythical 29th of February 1900 (there’s been plenty written about that “feature”, so I’m not going to go into the whys and wherefores of it here).
Back at the BRM, when IS 29500 was in the approval process, a lot of discussion centered around dates. There’s an existing ISO standard regarding use of dates and times electronically (ISO 8601) and many changes were made at the BRM to add support for that standard to IS 29500. Several of the larger differences between the Strict and Transitional IS 29500 conformance classes are around Strict using ISO 8601 dates, and Transitional using the runtime serial values.
Since the BRM a few defect reports have come up regarding the details of ISO 8601 support. Gareth Horton and I are the current project editors for one date-related change (the removal of ISO 8601 support from Transitional documents) and the intermingling of this project with other date-related DRs has made the editing process very complex. On Tuesday, Gareth and I presented to WG4 our current thoughts on this project – we believe the best plan is to expand the project scope to cover the other date issues. We’ll continue to work on this with WG4 through the conference calls, and we’ve pledged to have our project at the final draft stage in Tokyo in September (at the same time as the project scope can be altered to cover the extra work).
Of course, we also did a little sightseeing. WG4 went to a dinner reception on the island of Suomenlinna, sponsored by Microsoft and SFS (the Finnish standards agency). Given the WG4 fascination with cameras this went down well, and there are several pictures around the place of the rainbow we saw on the way back.
Although I’m still the new kid on the block, I felt like the meeting went well and we got some good productive work done. Tokyo looks like it’ll be focused on breaking the back of the outstanding defect reports and reviewing the final version of the date Project. I’d better get typing…