New links spec
Congratulations on the new draft. Just a few comments:
The last para of 3.3 currently reads
"Note that many of the attributes may be provided for both the parent
linking element and the child locator element. If any such attribute is
provided in the linking attribute but not in a locator element, the value
provided in the linking element is to
be used in processing the locator element. In other words, the attributes
provided in the linking element may serve as defaults for the (possibly
many) locator elements."
The second and third sentences of this paragraph are conflicting. If "the
value provided in the linking elment is to be used" how can "the attributes
provided in the linking element ... serve as defaults". I believe that "is
to be" should be "can, if no other value is specified by a locator element, be".
In 5.3.2 the spec for child currently reads:
selects child elements of the location source"
I would like to see "all" added between "selects" and "child" to make it
clear that more than one element may be selected by this option.
In 5.3.3 the definition of val currently says
"If specified within quotation marks, the attribute-value parameter is
case-sensitive; otherwise not."
Neat trick - pity it won't work.
Consider the following attribute:
<!ATTLIST figure size NUTOKENS #REQUIRED --height and depth-->
I enter <figure size="12pi 7mm"> in my document.
As these are tokens name case rules apply, so after parsing lowercase
characters should be replaced by caps, to give <FIGURE SIZE="12PI 7MM">
Now how can I specify I only want to pick up figures of a particular size?
I can't put spaces in as they stand and I have to know what the current
capitalization of my attributes are. Instead of entering something like
(1,figure,size,"12pi 7mm"), which matches my input quite well I have, I
presume, to enter (1,figure,size,"12PI%207MM") or
(1,figure,size,12pi%207mm). I forsee this presenting problems to users who
are not experts in SGML/URL specifics.
Do negative values for PRECEDING and FOLLOWING indicate counts from the
start/end of the document? If so, we should say so.
With regard to the Document element it would be nicer if #PCDATA was used as
the model rather than empty as we could then use this element to define a
citation title that could be displayed to users.
Martin Bryan, The SGML Centre, Churchdown, Glos. GL3 2PU, UK
Phone/Fax: +44 1452 714029 WWW home page: http://www.sgml.u-net.com/