From: "Don Osborn"
Subject: RE: Google data
Date: Tue, 3 Jul 2007 15:08:28 -0400
Hi Mark, all,
I have a general comment about the degree to which the process relies on self-appointed experts, which, if I understand correctly is heightened in the case of use of the Google tool.
In some contexts, such as wikis, the approach of having volunteer contributions can ultimately lead to good if not excellent results (though with the constant risk of new edits reducing rather than enhancing quality). In the case of locale data and other standards-related data, it would seem necessary to avoid the overwriting of good data with bad, and to foster a process of continual improvement. The remark about quality vetting seems to be made with this concern in mind.
Another part of the problem is that the process of writing locales will encounter issues particular to specific languages, and even dedicated and skilled experts without knowledge of the languages cannot be expected to be able to know all these (if for no other reason than they cannot learn every language for which lovcale data is submitted).
At the same time there are bonafide experts in various languages who are not part of the process, and likely have no knowledge of what locale data or CLDR are. It would seem that one way to help assure best results in the process of writing and imporving locales would be to find ways to involve such experts and control the process.
This actually may be more of an issue for languages that are less widely spoken and tend to have less resources - just the kind of languages that we will hopefully see increasing locale activity on.
How to develop such expert groups? This is admittedly not an easy process and might require a more involved (as in individual contcts, invitation, explanation) approach than one of opening a web-based tool and sending out notices on public lists. Actually these approaches should be complementary, but as it stands now we have mainly the latter.
Just my 2 F CFA . . .