comments on HTML 2.0 Spec

Warren Wiscombe (wiscombe@climate.gsfc.nasa.gov)
Mon, 30 Jan 95 10:02:34 EST

Hello,

I picked up the HTML Specification - 2.0, dated Oct 17, 1994, and wanted
to send a few comments. It is very nicely written overall. There are two
important omissions, however:

** in sec 3.13 FORMS, the example in sec 3.13.1 shows <input name="family"
type=int> but later there is no description of INT as one of the allowed
TYPE values; the summary in sec 2.1.6 also omits this possibility; it
makes me wonder whether there is also a TYPE=REAL, etc.

** in sec 3.13.1, the <P> tag is shown with an option ALIGN=CENTER, yet in
the formal definition (sec 3.12.1) this option is not mentioned; it would
be very useful to know this, since centered paragraphs are quite handy at
times

Other omissions:

** in sec 2.1.5, ISMAP and SRC are on the same line; it looks like they
should be on separate lines, with a definition following ISMAP

** in sec 2.1.7, the quote character was not mentioned

** in sec 3.4, it says each HTML document should start with the prologue

<!doctype ....>

but no further mention is made of this, and none of the examples use it,
even though the following text says "this prologue must be prepended"; is
this required or not? I have never seen it in any HTML docs on the Web

** sec 3.4.6: apparently a lot more than "historical implementations" use
> to end a comment; my HTML editor uses > in preference to --> because it
says MOST browsers still make this "mistake"

** sec 3.5.2: the Macintosh WWW browsers seem to handle relative filename
addresses incorrectly, and I always wind up having to specify the full path
to any file; somewhere this ought to be mentioned, perhaps here under BASE

** sec 3.5.3: the discussion of ISINDEX left me confused; an example
would help

** sec 3.5.4: it is not clear why the LINK element is necessary, unless it
is because of the final sentence in this section -- but that is not stated;
links are more naturally put in the BODY, so the reader is left wondering
why they would want to put them in the HEAD

** sec 3.6.3: the reader is left wondering if he should use TITLE or not,
because it is called "dangerous"; some better guidance is needed here

The following elements are poorly defined, so that it was not clear to
me how one would use them:

** <SAMP>, sec 3.8.6; what is a "sequence of literal characters"? Or more
to the point, how is that different than any other text?

** ISMAP, sec 3.10: a pointer to information on how to create an imagemap
is needed, at least

** <DIR>, sec 3.11.2: it was not clear how the columns would line up in a
browser, even for the simple example shown; would they be left-justified,
fixed width columns, etc.?

** <MENU>, sec 3.11.3: (if this just makes an OL or UL without blank lines
between list items, it hardly seems worth a separate tag; it should just be
an option on OL, UL)

Also, the FORMS discussion in sec 3.13 would be much improved by the
following:

** in sec 3.13.3, the description of ACTION was not really complete enough,
and only when I asked a colleague did I find out that your word "location"
usually meant an executable script or program, and that there is no need to
"elicit a response" since the script can just, for example, file the
information on the FORM without sending any acknowledgment (rude, I know,
but possible)

** in sec 3.13.3, the description of METHOD is incomplete; it would be
nice to have a short description of what POST and GET mean, and what METHOD
the average user should choose

** nowhere is it said that the FORM becomes STDIN for the program which is
the target of the ACTION field; this is important information for the
person writing the program to process the form!

** it would be nice to show a sample shell script for processing an
incoming FORM, say a script that just puts the information in the FORM onto
a file, or e-mails it to somebody; then a user could just copy the script
and start practicing with FORMs immediately (right now he would have to
consult other sources before getting started)

--------------------------------------------------
Dr. Warren J. Wiscombe, Goddard DAAC Project Scientist
NASA Goddard, Code 913, Greenbelt, MD 20771