by Michael S. Kaplan, published on 2006/06/04 09:23 -04:00, original URI: http://blogs.msdn.com/b/michkap/archive/2006/06/04/616814.aspx
If you follow Shawn Steele's blog you will have seen these already, but if you don't and you care about globalization support in Windows or the .Net Framework you will want to take a look:
There is a lot to grok in there, particularly the second one where backcompat runs headlong into working with standards and also working with a community that had been going in a slightly different direction than Microsoft was already on.
The old model for names was definitely built along the same architecture as LCIDs originally were. And the underlying resource model does not currently change at all. Also, underneath it all data is still for the most part structured the same way.
But the change (going from a language[-region][-script] to a language[-script][-region] ordering in the names) is a conceptual shift that will better match the way that people often look at locale and language identifiers.
As a rule, these types of changes are painful, since for every customer who is glad Microsoft will finally match standards efforts that for the most part came after them, there might be someone else who really only cares that backcompat is broken and feels that standards folks can get bent.
But they are important changes to do without delaying any further -- while the real impact can be minimized since so few cultures are affected.
(And as Shawn points out, those who are relying on the old behavior can always use custom cultures to get back the old (non-conformant) names they need!)
Definitely worth a read or three... :-)
This post brought to you by ܔ (U+0714, a.k.a. SYRIAC LETTER GAMMAL GARSHUNI)
referenced by
2007/07/11 Offense in depth
2006/06/04 Developers vs. Users?