On 04/26/2001 06:14:21 PM "William Overington" wrote:
>Peter Constable asks "If I write "chat", do you know what I mean?".
>
>Hmm, let me ponder! :-)
>
>Is it possible that you are referring to the answer that an Australian
>numismatist might give if asked what is the bird on the reverse of a
British
>farthing coin of the mid-twentieth century?
>
>(I found that in the dictionary just now. A useful dictionary, I checked
>how to spell numismatist as well. :-) )
Actually, it was the Romanisation of a word in Thai that I had in mind. You
didn't know what dictionary you should have been looking in.
>The prior agreement that the original author of the file of unicode plain
>text needs to have with the person seeking to read that file of unicode
>plain text need not be between the two people directly.
Of course; the two people need not know that the other even exists. But the
receiver has know that the hypothesized author intends that the characters
be interpreted in some way. You have suggested that we have some coding
mechanism that can be used to identify what interpretation convention is
intended. My point is that you as a receiver are *assuming* that the
characters you receive and interpret as that coding mechanism were intended
by the author to be interpreted that way. You haven't solved the problem;
you merely introduced a level of indirection.
>Sentence B. If I send you PUA characters in a plain unicode text file,
even
>in context you don't know what I mean unless there exists a separate
>document that explains what the characters mean.
That's not all. (a) That document must exist; (b) you have to know that I
intend for you to interpret in terms of that document; (c) I need to
communicate that intention to you some how, and (d) you have suggested some
encoding mechanism for the purpose of exactly that communication. But then
(a) there needs to be some document that specifies that convention, and (b)
you have to know that I intend for you to follow that convention, and so
(c) I need to communicate that to you some how... Do you want to loop
through (d) and back to (a) again, or should we simply interrupt that
process, step outside the data stream and communicate that some other way
(i.e. with metadata)?
>Now if you had stated sentence B then I would agree with you entirely...
>Are we both in fact saying the same thing or am I missing something?
Direct vs. indirect interaction has nothing to do with what I'm saying.
You've said that we can engineer some solution that encodes the identity of
an interpretation document into the text stream. I've said that the
recipient needs to interpret *that* encoded messages that was added to the
text stream. Unless you get conventions for that interpretation
standardized, which will never happen, then your solution has not solved
anything unless we have both agreed (with or without direct interaction) to
adopt that convention, *and that has to be done externally to that
mechanism*. Since we have to work outside the mechanism at some point, why
not skip that mechanism.
- Peter
---------------------------------------------------------------------------
Peter Constable
Non-Roman Script Initiative, SIL International
7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
Tel: +1 972 708 7485
E-mail: <peter_constable@sil.org>
This archive was generated by hypermail 2.1.2 : Fri Jul 06 2001 - 00:17:16 EDT