Re: rationalizing identifier [was Re: Rationalizing the term URI]

I'm not in favor of redefining the term "URI" in terms
of byte sequences instead of 'sequence of characters'.

I think there is a lot of strength in the fact that
a URI encoded in EBCDIC is the "same" 
as one encoded in ASCII and with the "same" characters. 
(in this case, same[URI] is the baseline equality
for URI equivalence, and same[characters] is the
equivalence of characters after transcoding.)

As long as URIs have a restricted repertoire
of characters, there is no problem.

Part of the work for IRIs is to insure that 
two IRIs which have the same characters
are equivalent even when represented in two different
charsets, especially when the charsets have different
character repertoires or non-1:1 mappings.

Larry
-- 
http://larry.masinter.net

Received on Friday, 24 January 2003 02:13:04 UTC