Re: Comments, parsers, XPointers
John E. Simpson (simpson@polaris.net)
Wed, 20 May 1998 15:22:14 -0400
At 01:34 PM 1/1/80 -0500, you wrote:
> [various helpful stuff]
>(comments have no purpose on the production/browsing side).
I was wondering about that. Specifically, I was wondering about a "view
source"-type feature in XML browsers, perhaps with a show/hide comments
toggle. I've no experience with SGML, but I (and a lot of -- most? --
other HTMLites) have frequently made use of other developers' comments for
learning purposes, outside the context of formal training. Assume there
were some convention for associating a comment with a particular element or
other component of the logical (even physical) model (probably a big
assumption!); some facility for locating element X's comment, if any, would
be extremely helpful for such purposes. A
docname.xml#child(1,elementname).(1,#comment) sort of XPointer seems like a
natural construct in this case.
Of course excessive comments add to network load, invite abuse and so on (I
still get the willies when I see scripting embedded in HTML comments); but
they aren't always noise, and aren't always useful only to their author.
>In the longer term, what we need is an official definition of an XML
>information set, specifying (for example) that reporting comments is
>optional, while reporting the start and end of elements is required.
>Once such a beastie exists, many vexing questions about (and
>inconsistencies among) the DOM, XPointers, SAX, XSL, etc. will
>disappear.
Yes to all!
Thanks, David.
John
John E. Simpson | It's no disgrace t'be poor,
simpson@polaris.net | but it might as well be.
| -- "Kin" Hubbard