William Overington <WOverington at ngo dot globalnet dot co dot uk>
> I feel that what happened is very interesting and should go down in
> Unicode history as "The Respectfully Experiment".
> William published a list of code points for ligatures. Quite
> independently of each other, James used the list to add a code point
> for a ct ligature into his fount, Doug used the list to include a
> code point for a ct ligature in his posting. Doug posted to the
> Unicode list, not directly to James, without any direct prior
> arrangement with James over the use of these Private Use Area codes.
> The message was received by James' computer and displayed correctly.
> So, the meaning of the sender was communicated to the recipient
> using a code from the Private Use Area with a meaning obtained from
> a published list.
It was sheer luck that anybody on earth had a font that contained a
ct-ligature at U+E707 and was actually using it to view my message.
Certainly I did not expect it. So I suppose the words "except by
private agreement" in the Unicode Standard could be read as "except by
private agreement or by sheer luck," but that's a pretty crummy way to
exchange data. Notice that only one person, out of everyone on the
Unicode mailing list, was able to see the ligature. If we wanted to
rely on sheer luck for information interchange, we wouldn't need
Unicode. We could stick with the 80,000 privately defined encodings we
had in 1990.
> This is, I feel, an experiment which should be properly documented
> in the Unicode archives.
By default, this discussion has already been documented in the Unicode
mail list archives. In a "Congressional Record" sense (sorry, don't
know the Parliamentary equivalent), everything we write on this list
goes into a "Unicode archive."
If you mean the Unicode Consortium should document this "experiment"
with the PUA in some formal way on their Web site... uh, don't hold your
> On the matter of searching the archives for the word "Respectfully"
> and not getting a hit returned for that message, I wonder if that
> situation need persist for ever.
> Suppose that the software which copies documents into the archive
> automatically converted any use of a golden ligatures code point
> into the constituent letters and ZWJ characters and that when an end
> user requests a search, ZWJ characters are ignored by the search
Indeed, that's just what I might expect suffficiently advanced
Unicode-based software to do (perhaps without the ZWJ's) for OFFICIALLY
DEFINED UNICODE CHARACTERS. I would absolutely NOT expect publicly
distributed software to interpret Private Use Area characters. Is this
> I realize that that last paragraph misses out the issue of the
> non-uniqueness of a Private Use Area code point designation, yet if
> the ligatures were promoted to regular Unicode, that problem would
> not exist.
Please, please forget you ever heard the word "promotion" used to
describe the relationship between a character encoded in the PUA and
that same character subsequently (if ever) encoded officially in
Unicode. That is not the nature of such a relationship, and I wish
Unicode had chosen a different word.
These ligatures are almost certainly *not ever* going to be encoded
officially in Unicode. Is this clear yet?
> [98-word clause omitted]
> I am pleased to suggest the code point U+E7C1 WATERMARK-LIKE MEMORY
> THAT A WHOLE LIGATURE WAS ORIGINALLY USED FOR THE FOLLOWING LIGATURE
> in the hope that experimental databases produced by various
> experimenters might all be compatible one with another.
Now I'm beginning to think this whole discussion was all a big joke that
I didn't catch on to until now. Ha ha. Very funny. Egg on my face.
I'm afraid this my have to be my last post on this topic, since William
seems to regard every response (no matter how negative) to his proposals
for misusing the PUA as a semi-public standardization area as incentive
to expand the proposals in ever more ambitious and impossible
directions. It may indeed be true, as the Hollywood film producers here
say, that "there's no such thing as bad publicity."
This archive was generated by hypermail 2.1.2 : Mon Jun 03 2002 - 10:29:07 EDT