--On Thursday, April 24, 2014 08:39 -0600 Peter Saint-Andre
<[email protected]> wrote:
> IMHO the WG has worked hard to get things right - we've had
> two separate implementations comparing their output, etc. But
> as we know i18n is difficult, and perhaps only John Klensin
> has the expertise to get this right anyway. :-) That's partly
> why in general we based PRECIS on the same technical approach
> taken in IDNA.
I know only three things:
(1) How little I actually know, an assessment that gets worse
with each passing week and problem.
(2) How badly some approaches can lead us astray and into
problems, including the Law of Least Astonishment one and its
relationship to profiles in this area.
(3) Following up Peter's comment, how dangerously close to
burning out I get at regular intervals, sometimes unfortunately
resulting in tuning out as a self-preservation measure. The
latter got me several times during PRECIS discussions, in large
measure because the whole approach that could lead to "one
application, one profile" struck me as so profoundly wrong. I
couldn't get the WG to address that issue and investing energy
in fine-tuning documents that reflected that approach just
didn't seem productive.
john
p.s. The W3C i18n WG has had an action item on its agenda since
late (Northern Hemisphere) summer to note the differences
between the evolving PRECIS approach and that of the Charmod
revision, differences in string matching, and some normalization
concerns. From the viewpoint of my "too many profiles"
concerns, Charmod is yet another profile approach as is UAX31
("Unicode Identifier and Pattern Syntax"). I did mention on
their call today that the action item is probably OBE but, if
anyone wants to pursue those differences, the contact points are
Addison Phillips <[email protected]> and
Richard Ishida <[email protected]>
_______________________________________________
precis mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/precis
