Date: Sun, 09 May 1993 14:09:40 -0400 (EDT)
From: John C Klensin <KLENSIN@INFOODS.MIT.EDU>
Subject: Re: Wrappers for URLs
In-Reply-To: <9305081825.AA07142@wilma.cs.utk.edu>
To: moore@CS.UTK.EDU
Message-Id: <736970980.746103.KLENSIN@INFOODS.UNU.EDU>
John Curran wrote...
>> we recognize that we are eternally condemming said character to an
>> inferior representation (%xx). Some characters that would not make
>> good choices for this honor include period (used in hostnames) and
>> slashes (used in filenames).
And Keith Moore responded...
>Well, I used underscore = 20 hex in RFC 1342, so I guess I don't mind
>it too much...and it's certainly better than not encoding them.
>
>Is there anybody who cannot live with this?
While there are few of them on the Internet at this point, underscore is
a very popular separator in file names on some operating systems.
Treating it as isomorphic with space won't work for them, as some of
them can (reluctantly) handle file names with embedded spaces.
None of this is a problem as long as we insist that URLs are
abstractions, and not file names. But, if they are really abstractions,
then the issue of what to "condemn to an inferior representation"
becomes moot.
The search for what character we can live without always walks down the
path of "what hasn't been used for anything else important" and we may
have exhausted the list. On logic similar to John's, I'd suggest that
none of _ \ - ! % are really good candidates either. Perhaps "*" or
"?", if no one would confuse them with wildcard symbols, but they are
not especially readable as intuitive space-surrogates.
john