Re: ERB: decision and conundrum [URLs]

lee@sq.com wrote:
> On the subject of fragment IDs, there seems to be some confusion.

Which seems to have roots in the terminology.
> There are two kinds of fragments.

Or two things named "fragments".
> One is a standalone part of a larger document, where only the fragment
> is served.  

That is fragment as I usually think of it.  A separable piece of a

> That isn't the web fragment, though.  The other is a named
> part of a document, where you have the whole document, and want to go
> to a particular region (surrounded in HTML with <A NAME=xxx>....</X>).

That is a named location with an element to anchor the locator.  Call it
fragment if you like, but that isn't what one usually thinks of as a 
fragment in SGML.  Thats a position to scroll to.
> Please don't confuse them.

Sounds like it got confused in the W3C HTML WG.
> In a web world, you are supposed to do SGML Open chunks by giving each
> chunk its own URL, just as DynaWeb (there! I spelt it right!) does.

How does that work?  IOW, what gets returned?  The document, or the 
> Since # is for something different than chunks, it doesn't have to solve
> the chunk requirements, and therefore the fact that the #stuff doesn't
> get sent to the server turns out not to be a problem.

Are you suggesting a different URL convention for the SGML fragment?
Again, what gets returned?