Re: [idn] IDN WG Last Call on two major changes to Stringprep

["Adam M. Costello" <idn.amc+0@nicemice.net.RemoveThisWord>]

I wrote:

> I have reviewed and understood the changes to Stringprep, and approve
> of them...

I forgot about one minor technical issue that I starting discussing with
the authors a while back, but that discussion fizzled out.

One of the changes to Stringprep is the addition of two new prohibited
characters, U+0340 and U+0341 (combining grave/acute tone marks).  These
characters are prohibited because they are deprecated, and for no other
reason.  The previous draft of Stringprep also prohibited deprecated
characters, but the reason for their prohibition was something else
(they "change display properties"), and the fact that they were
deprecated was incidental.

I see no real harm in prohibiting these two characters, and no real need
to prohibit them either.

But Nameprep (or any other profile) will not be able to follow this
precedent of prohibiting deprecated characters in the future.  If
subsequent versions of Unicode deprecate more characters, Nameprep will
not be able to prohibit them, because section 7.1 forbids later versions
of a profile from prohibiting characters that were allowed in earlier
versions of the profile.

Why start down a path that you can't stay on?  Perhaps deprecation alone
should not be considered sufficient reason for including a character in
the prohibition lists.

As I said, this is a minor issue.  I don't think it really matters
whether these two characters are prohibited or not, I just think they
might end up looking like anomalies in the future.

There is a related issue that is purely editorial, not technical.  If
these two characters are prohibited, I don't see why the categories
"change display properties" and "deprecated" should be merged into a
single section, as they are in the current draft.  They have nothing
to do with each other, except that there happen to be some characters
that could fall into either category.  This is not the only instance
of overlap.  The "change display properties" category overlaps with
the "control characters" category, and yet those two are not merged.
Similarly, the "inappropriate for plain text" category overlaps with
the "control characters" category, and those two are not merged.  For
consistency, I think the "deprecated" category, if it is retained,
should get its own section.

AMC