HTML 4

I've been browsing through the HTML 4.0 document, and come up with
some questions/suggestions.


More entity names

     In the present definition, the entity names supplied are a subset
of entities present in UNICODE.  In my opinion this is a mistake.
Graphical user agents will often be capable of representing a wider
range of glyphs than has at present been named.  For instance, I would
make use of the following symbols: 'bottom' ( approximately _|_),
'top' (approx. ¯|¯) and 'mapsto' (approx. |->), and I can find none of
these in the present list of names, although they all exist in
UNICODE.  It would be better to supply names for as many entities as
can be thought of (whether or not the appear anywhere in UNICODE),
since this would allow authors to specify the correct entity, and have
it rendered correctly some time in the future.  For example, I had
occasion to use the square-root symbol in the past, and put &sqrt; in
the expectation that future versions of HTML would use that name.  Of
course it is now (quite correctly) called √ If this name had
been specified in HTML 2 (in the expectation that most UAs would have
rendered it unchanged), I would not have had to change the document.

In general, the use of numbers to refer to characters ought to be
frowned on, since this is poor practice from a computer science point
of view.  Authors will check using some specific browser, which may be
incorrect.


Clarification of the use of the Q entity

    (Hmm - seems to be something to do with Star-Trek!) In the present
specification, Q and BLOCKQUOTE are described together, and it is not
made clear what one should expect from <Q>quotations</Q> in the
default style.  I would be pleased if <Q>foo</Q> came out the same as
&lsquo;foo&rsquo; (or &ldquo;...) as this would solve the tedious old
problem of matching quotation marks with minimum fuss.  Since there is
no Q element in earlier versions of HTML, there would be no harm is
reccommending the addition of quotation marks for Q in the default
style (when not already inserted by the author), unlike BLOCKQUOTE.

Image alignment.

     Sometimes one wants to include an image that represents something
essentially textual.  Examples are monograms, typographic or
calligraphic specimens and certain forms of logotype.  In such cases,
the results for graphical output are usually poor with HTML < 4
because it is impossible to get the vertical alignment right.  One
wants the baseline of the text in the image to align with the baseline
of the document text.

A solution would be to include an attribute specifying the position of
the baseline relative to the bottom of the bottom row of pixels in the
image. This could then be used with the 'vertical-align' property
'baseline'. Note that although many of the cases where this might be
used can be done better by avoiding the use of an image altogether,
there will always remain cases where an image is the only way of
presenting the information wanted.


Tarting up bulleted lists

Whether we like it or not, a frequent usage is to use an image
(typically brightly coloured) in place of the bullets in <UL>.  This
tends to be messy since the indentation is not designed to cope with
it.  It would therefore probably be wise to include a proper facility
to do this.

  Jón

-- 
Jón Fairbairn                                 Jon.Fairbairn@cl.cam.ac.uk
18 Kimberley Road                                        jf@cl.cam.ac.uk
Cambridge CB4 1HH                      +44 1223 570179 (pm only, please)

Received on Monday, 11 August 1997 12:12:23 UTC