confused by use of "xexp" in your docs...

54 views
Skip to first unread message

'John Clements' via dev-redirect

unread,
Mar 3, 2015, 1:41:47 PM3/3/15
to Neil Van Dyke, d...@racket-lang.org
In trying to diagnose the recent issue with the sxml parsing libraries, I was super-confused by your package’s use of the term ‘xexp’ to mean… well, something very much like sxml. Given the name, I assumed it would instead be something very much like x-expressions.

I’ve just taken a quick look at Oleg’s sxml specification, and I find, to my mild disappointment, that it’s extremely well-written and precise, and I now understand why you’d be reluctant to describe your format as being sxml. It’s pretty darn close to SXML/1NF, though. It’s just those darn entities that are going to be problematic.

Anyhow, in my opinion, naming your function

html->sxml

rather than “html->xexp” would reduce confusion; the documentation is already quite clear about the minor differences between SXML/xexp and SXML.

Just my two cents, I suppose.

John

Neil Van Dyke

unread,
Mar 3, 2015, 2:11:52 PM3/3/15
to John Clements, d...@racket-lang.org
A high priority for me is to not dilute the meaning of Oleg's
well-specified SXML by calling things "SXML" that are not SXML. Which is
why I started temporarily saying "SXML/xexp" in docs, and "xexp" in
identifiers.

I'm trying (well, I let this go to the back burner) to get the SXML
changed or to otherwise get Oleg's blessing.

I should take that up again with Oleg, before I consider renaming
anything and confusing the meaning of "SXML".

Neil V.

Neil Van Dyke

unread,
Mar 22, 2015, 7:29:47 PM3/22/15
to John Clements, d...@racket-lang.org
FYI, for now, I shouldn't call my tools' existing SXML variant "SXML".
(I haven't yet heard back from Oleg.)

Regarding moving my tools' formats back to 100% SXML: I want to specify
things so that the symbols like "*top*" can be lowercase (original SXML
spec was ambiguous, IIRC), and I'd really like to keep a reasonable
ASCII representation for HTML character entity references, like `(&
rArr)`, especially for handwritten HTML. (I think my original HtmlPrag
parser from 2003 supported that character entity extension to SXML, it's
been useful, and I'd really like to keep using it.) I'm more ready to
give up my other tweaks, like limited to 1NF, and the alternate to the
somewhat problematic `@` character.

I'll try Oleg again in a month or so, if I don't hear back before then.

Neil V.
Reply all
Reply to author
Forward
0 new messages