Test Data Gone?

[Steven R. Loomis]

> On Nov 18, 2015, at 11:16 AM, Doug Ewell <doug at ewellic.org> wrote:
> 
> Steven R. Loomis wrote:
> 
>>> At the time we retracted it, it didn't appear that there was a lot of
>>> usage, and you really get a much more thorough test by comparing to
>>> ICU's implementation.
>> 
>> Right. An idea at IUC was rather than trying to scope test data as
>> cldr conformance test data, to have a new effort that simply and
>> explicitly records ICU's result  for a certain Icu/cldr version
>> somewhere for certain input values and certain formatting routines.
>> People are doing this already, just combine efforts. 
>> 
>> Maybe the results would be an Icu-maintained file  instead of cldr,
>> like a sample app. 
> 
> I've always been uncomfortable with the idea, which has been expressed
> from time to time, that CLDR is merely a data format for ICU, instead of
> being generally usable by applications and libraries that have nothing
> to do with ICU.

Doug,
 I am uncomfortable with that as well, and always glad for non-ICU users of CLDR.

 It was (non-ICU) implementors of CLDR which asked for the data.
 CLDR had not had a set of maintained conformance data, so the idea I mentioned was to simply make available “ICU’s results” in a non-normative way.
 The first preference is of course to improve the documentation so that it is usable as is. However, the question came up, “why not just dump ICU’s output so people have some examples to look at”.  My comments above reflect a struggle to find a home for such data.

 By the way, that reminds me of the implementer’s guide which was started here - http://goo.gl/zAfDt <http://goo.gl/zAfDt>   

 Do you have any other recommendations on how to make CLDR more generally usable?

> Mark and Markus will recall that I expressed similar discomfort back in
> 2002, when BOCU-1 was specified in terms of ICU-based "sample" code
> instead of an actual specification, and when the conformance test was
> "make sure your implementation generates the same output as the sample
> code.”

Right. ICU must not become the spec for CLDR.

-s

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://unicode.org/pipermail/cldr-users/attachments/20151118/715efce5/attachment.html>