<br><br><div><span class="gmail_quote">On 4/13/07, <b class="gmail_sendername">John Cowan</b> <<a href="mailto:cowan@ccil.org">cowan@ccil.org</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Mark Davis scripsit:<br><br>> That means that in ISO 639-2 the following are collections:<br><br>> mul Multiple languages<br><br>Yes, unfortunately. 639-3 takes a more sensible stand, and calls<br>'und', 'zxx', and 'mul' special (as distinct from individual language,
<br>macrolanguage, collection, or private-use).<br><br>> The language codes clearly do not form a partition, since some<br>> collections encompass other codes.<br><br>True. The collection codes, however, form a rootless hierarchy,
<br>a rather flat one, of which "mis" is one top-level component.<br><br>> The collection codes that are tagged with (Other) are clearly meant<br>> to be the remainder of partitions, but for the collections that are
<br>> not tagged with (Other) there is no evidence that they were intended<br>> to exclude other cases -- if anything, the contrary -- if they had<br>> meant to be the remainder of partitions, they would have said (Other).
<br><br>I think that's reading too much into them; the entries are the work<br>of many hands. FWIU, all the "(Other)" epithets are being removed,<br>anyhow.</blockquote><div><br>What is in the specification is what we can use to establish the semantics of the code at introduction into BCP 47, which then determines whether any successive changes to the Description are allowed (broadenings) or not.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">> Moreover, while some may be perfectly willing to have stability go by<br>
> the wayside, it is extremely important to us, and that is one of the<br>> guiding principles of and reasons for BCP 47. That means that if I<br>> validly and correctly tag content with "mis", that application cannot
<br>> be made incorrect by any future change to BCP 47. That is why we can<br>> broaden the application of codes, but cannot narrow them.<br><br>+1<br><br>> There is no evidence in BCP 47 that I cannot correctly tag the content
<br>> "kind" with "mis".<br><br>There is no *evidence* that you cannot tag "George slipped on a banana<br>peel and fell down" as "fr". It's a quality of implementation issue.
</blockquote><div><br>I disagree. Any reasonable person with a knowledge of French and looking over the text of the standard would say that tagging that sentence with "fr" is a mistake. The same is not true of 'mis'.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">--<br>Eric Raymond is the Margaret Mead John Cowan<br>of the Open Source movement.
<a href="mailto:cowan@ccil.org">cowan@ccil.org</a><br> --Bruce Perens, <a href="http://www.ccil.org/~cowan">http://www.ccil.org/~cowan</a><br> some years ago<br></blockquote></div><br>
<br clear="all"><br>-- <br>Mark