- From: Martin J. Dürst <duerst@it.aoyama.ac.jp>
- Date: Sat, 27 Aug 2016 14:44:19 +0900
- To: Annette Greiner <amgreiner@lbl.gov>, "Phillips, Addison" <addison@lab126.com>
- CC: Phil Archer <phila@w3.org>, Deirdre Lee <deirdre@derilinx.com>, Bernadette Farias Lóscio <bfl@cin.ufpe.br>, "ishida@w3.org" <ishida@w3.org>, "public-dwbp-comments@w3.org" <public-dwbp-comments@w3.org>, www International <www-international@w3.org>
Hello Annette, I think the data in your example is locale-neutral, because Unix time is the same in all locales. It's not necessarily the most frequent locale-neutral representation, and it's not human readable, but these are separate issues. Regards, Martin. On 2016/08/27 02:31, Annette Greiner wrote: > Addison, or anyone in i18n that knows, > > I have a question about how to make data locale neutral when all the values are of the same format. It seems a hard sell to tell people to expand every datetime value into a value and a format when the format is always the same. For example, if I have taken a terabyte worth of data (not at all uncommon where I work) that consists of pairs of a datetime plus a sensor reading, and the time is always in UNIX format (seconds since January 1, 1970), is it still considered locale-neutral if I simply indicate in the column metadata that the column holds UNIX time values? > > UNIX time, sensor reading (mV) > 1471995721,4.7 > 1471995731,7.5 > 1471995721,6.2 > > If not, is there a practical way to make it locale neutral without repeating the format for each value? I’m concerned that, for scientific datasets especially, it makes little sense to inflate the size of the dataset with repeated information. For large data, that becomes impractical, which tempts me to suggest that we say in the BP that locale parameters should be used only when a locale-neutral representation is not *practical* (rather than not *possible*), unless my example above would qualify as locale neutral. > -Annette
Received on Saturday, 27 August 2016 05:45:06 UTC