From: Jukka K. Korpela (jkorpela@cs.tut.fi)
Date: Wed Oct 26 2005 - 16:56:58 CST
On Wed, 26 Oct 2005, Michael Everson wrote:
> At 00:11 +0400 2005-10-27, Andrew S wrote:
- -
>> That it will "disrupt everything" and "invalidate oceans of existing data"
>> I still don't understand.
>
> A, B, C, D, E, and F are already used by *EVERYONE* for representing
> hexadecimal data. My Calculator on OS X can even perform calculations using
> them. Therefore, hexadecimal numbers are already represented by the letters
> A, B, C, D, E, and F, and the proposal is a bad one.
The proposal might be a bad one, but I think the rationale for rejecting
it has not been sufficiently explained. How do the statements about
disrupting everything and invalidating existing data follow from
established use of A to F as hexadecimal digits?
The issue might be important to future discussions, and clarifying the
situation could help to prevent proposals that have no chances of getting
approved, or at least have them rejected in a manner that better convinces
their advocates.
As far as I have understood, the proposal was about adding new characters,
for use as hexadecimal digits, and normally with shapes similar to the
letters A through F. And unless I'm missing something, the reason would be
semantic disambiguation. (The proposal seems to present arguments related
to font issues, too, but I must admit that I lacked motivation to read the
rejected proposal in any detail.)
If the (main) reason for rejection was that new characters will not be
added to Unicode and ISO 10646 just to make semantic distinction, when no
visual difference (normally) exists, then I think it would have been
useful to say so. The same applies if the reason was that characters can
be so added but only if a demonstrated need for the disambiguation, or at
least tangible usefulness, is demonstrated. In that case, people
would know that in cases like this, good arguments will be needed. But
if the point is that semantic distinctions alone won't do at all,
then it would be futile to collect and formulate such arguments.
However, would the proposal, if accepted, have _invalidated_ existing
data? In which sense? Even if it had said that the new characters are the
recommended characters for use as hexadecimal digits, I do not see how
existing data would have become invalid. Moreover, the way I see it,
the proposal was just about semantically unambiguous _alternatives_, just
as the hyphen is a semantically unambiguous (or at least considerably less
ambiguous) alternative to the hyphen-minus. I don't see complaints about
existing data being invalid just because it uses hyphen-minus.
I don't see how the addition of new characters could _invalidate_
existing data. Is this a matter of interpreting words like "invalidate"
differently?
I can see many practical reasons for rejecting the proposal. Existing
software would need some updates _if_ it wanted to handle the new
characters e.g. when reading data in hexadecimal notation. But that would
be optional. Someone might wish to convert existing data to use the new
characters, but that would be his choice.
-- Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
This archive was generated by hypermail 2.1.5 : Wed Oct 26 2005 - 16:57:44 CST