Terry> Which RFC is that?
appendix to 1738, but I think someone already said that.
Terry> One objection I have to that is that it's not eas easily
Terry> readable. The "<URL: ... >" adds no information to
Terry> "http://some.url.com" and is awkward when writing text to be
Terry> read by humans, as opposed to by machines.
URLs are meant to be read by machines. Anything that detracts from
those machines' ability to unambiguously parse them is a mistake, IMO.
If a human can read them, great. But the point of an URL is to tell a
computer where to go for a resource. The <URL:http://site/full/path/>
syntax does this.
This is especially true now when various web programs are using
various non-alphanumeric characters in URLs. It is pretty hard to
write a program that will pick out all URLs correctly with no extra
characters. That is, unless the URLs are in this form.
Terry> Yes, looking at the RFC, it does seem to be a mere
Terry> recommendation. I hate it, though.
Right.
But if you read it, it is being recommended for exactly the sort of
context we are talking about here.
[RFC 1738 snippet]
Terry> In addition, there are many occasions when URLs are included
Terry> in other kinds of text; examples include electronic mail,
Terry> USENET news messages, or printed on paper. In such cases, it is
Terry> convenient to have a separate syntactic wrapper that delimits
Terry> the URL and separates it from the rest of the text, and in
Terry> particular from punctuation marks that might be mistaken for
Terry> part of the URL. For this purpose, is recommended that angle
Terry> brackets ("<" and ">"), along with the prefix "URL:", be used
Terry> to delimit the boundaries of the URL. This wrapper does not
Terry> form part of the URL and should not be used in contexts in
Terry> which delimiters are already specified.
By specified, I am pretty sure that Tim Berners-Lee meant 'explicitly
specified', such as in a program that reads/writes URLs in some fashion.
Terry> I think that the last sentence's exception swallows the whole:
Terry> in most contexts, simple whitespace provides delimitation,
Terry> making the angle-brackets and "URL" redundantly superfluous.
Does it?
What about punctuation? '.' is a legal URL character, and I have seen
many URLs that contain '?' and other characters that are often used as
punctuation. The <URL:http://site/full/path/> syntax is not
ambiguous. 'Simple whitespace' is.
-Justin
[
Usenet Hypertext FAQ Archive |
Search Mail Archive |
Authors |
Usenet
]
[
1993 |
1994 |
1995 |
1996 |
1997
]
© Copyright The Landfield Group, 1997
All rights reserved