Empty fragment

[Harlan Messinger]

Is the following a legal URL?

http://www.invalid.com/path#

This underlies my real questions:

1. Is <a href="#"> a *valid* (whether or not intentionally) way to go to
the top of a page.

2. Whether or not valid, do the usual browsers actually support this?

I'm asking because I'm reviewing a large set of pages created by several
other people, and many of them are using this device to return to the
top of the page. (No need to reopen the argument about not using Back to
Top links at all--not my decision!)

[David Håsäther]

Harlan Messinger <hm*******************@comcast.net> wrote:

Is the following a legal URL?

http://www.invalid.com/path#
Yes.
This underlies my real questions:

1. Is <a href="#"> a *valid* (whether or not intentionally) way to
go to the top of a page.
AFAIK, the behavior is undefined for empty fragment identifiers. I'm
not sure though, so you may want to look into RFC 2396 (somewhere
around section 4 maybe).
2. Whether or not valid, do the usual browsers actually support
this?

IE, Mozilla and Opera does to my knowledge, and I think this is the
most common behavior for browswers.

--
David Håsäther

[Harlan Messinger]

David Håsäther wrote:

Harlan Messinger <hm*******************@comcast.net> wrote:

Is the following a legal URL?

http://www.invalid.com/path#
I guess that should be either www.example.com or www.something.invalid.

Yes.

This underlies my real questions:

1. Is <a href="#"> a *valid* (whether or not intentionally) way to
go to the top of a page.

AFAIK, the behavior is undefined for empty fragment identifiers. I'm
not sure though, so you may want to look into RFC 2396 (somewhere
around section 4 maybe).

Yeah, it still doesn't seem to answer the question.
2. Whether or not valid, do the usual browsers actually support
this?

IE, Mozilla and Opera does to my knowledge, and I think this is the
most common behavior for browswers.

I just checked: even Lynx does.

[Jukka K. Korpela]

Harlan Messinger <hm*******************@comcast.net> wrote:

1. Is <a href="#"> a *valid* (whether or not intentionally) way to
go to the top of a page.

AFAIK, the behavior is undefined for empty fragment identifiers.
I'm not sure though, so you may want to look into RFC 2396
(somewhere around section 4 maybe).

Yeah, it still doesn't seem to answer the question.

RFC 2396 has been superseded by RFC 3986 (STD 66) in January 2005.

I haven't studied RFC 3986 in detail, but it seems to have obscured
things somewhat. The one-character string "#" is surely a syntactically
correct URL, but its meaning is not explicitly defined.

P.S. "Valid" is not the right word here; e.g.
<a href="! ¤¤ ##%">foo</a>
is valid but incorrect.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[John Dunlop]

Harlan Messinger wrote:

Is the following a legal URL?

http://www.invalid.com/path#
doesn't violate STD66 anyway.
This underlies my real questions:

1. Is <a href="#"> a *valid* (whether or not intentionally) way to go to
the top of a page.
valid, and moreover conforms to the text of the spec, read
loosely.

My take on it is, in a text/html doc, an empty fragment
designates an element named with the empty string (so that
element can't be named with 'id'). If that element doesn't
occur, the semantics of the fragment identifier are unknown.
2. Whether or not valid, do the usual browsers actually support this?

My Lynx (2.8.3) moves me to the top, no matter if an element's
named "" or not; but my Opera (7.54) does not move me to the
top, although if an element is named "" (even with 'id', which
violates the spec), it moves me there.

--
Jock

[Big Bill]

On Tue, 12 Jul 2005 20:07:30 GMT, John Dunlop
<us*********@john.dunlop.name> wrote:

Harlan Messinger wrote:

Is the following a legal URL?

http://www.invalid.com/path#

doesn't violate STD66 anyway.

This underlies my real questions:

1. Is <a href="#"> a *valid* (whether or not intentionally) way to go to
the top of a page.

valid, and moreover conforms to the text of the spec, read
loosely.

My take on it is, in a text/html doc, an empty fragment
designates an element named with the empty string (so that
element can't be named with 'id'). If that element doesn't
occur, the semantics of the fragment identifier are unknown.

2. Whether or not valid, do the usual browsers actually support this?

My Lynx (2.8.3) moves me to the top, no matter if an element's
named "" or not; but my Opera (7.54) does not move me to the
top, although if an element is named "" (even with 'id', which
violates the spec), it moves me there.

How come you're sensible over here?

BB
--
www.kruse.co.uk/ se*@kruse.demon.co.uk
seo that watches the river flow...
--