Re: checklink: code 200.

On Fri, 19 Mar 2004, Olivier Thereaux wrote:

> On Thu, Mar 11, 2004, Pascal wrote:
> > I don't really understand what is "broken fragment".
>
> This comes form the term "fragment identifier", which is used for the
> part of a URI after the "#" character.

Or the fragment that appears after a URI. The URI specifications partly
use a bit weird terms, and technically they say that a fragment is not
part of a URI but part of a "URI reference". See the nasty details in RFC
2396, specifically section 4, "URI References", available at
http://www.cs.tut.fi/~jkorpela/rfc/2396/full.html#4

On the other hand, HTML specifications (and many other specs) say "URI"
when they apparently mean "URI or URI reference". I'm afraid there's not
much we can do to remove this confusion. (And "URI" is for most practical
purposes the thing that normal people know as "URL".)

> What checklink is telling you is that it found a link where
> "fragment identifier" (e.g <a href="#example">) that does not refer
> to any "name" or "id" in the document, so the link may not be broken,
> but the fragment is.

I would say that the link is broken too - though in a manner different
from a link with the URI proper broken. It's very useful that checklink
verifies such things. But maybe the wording could be somewhat different.
The phrase "broken fragment" sounds like a fragment that is syntactically
malformed, e.g. contains characters that need to be URL escaped.
Maybe "undefined fragment" would be better.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Friday, 19 March 2004 04:20:13 UTC