To HTML or not to HTML?

[brender at gemevn dot zko dot dec dot com (Ron 603-884-2088)]

Gentleman,

The DWARF 2.1 Draft 1 document now out for review was created with the
expectation that we would also want to provide an HTML file that matches
the .pdf. Now that I have seen how the .pdf comes out, I am wondering if
that should still be a goal.

There are several reasons for not providing a separate .html file:

  - It appears that the .pdf can do everything that .html can do wrt
    hot links and the like. Thus an .html adds no value. (I think the
    only thing lost is automatic resizing to fit your browser window
    -- which is no big deal as far as I'm concerned.)

  - It is yet another file to keep track of, pass around, generate,
    make sure it matches the .pdf, etc.

  - HTML is in fact less expressive than the .pdf in a number of ways,
    which may or may not be regarded as important. Examples:

      . There are two footnotes in the 1993 spec (see pages 9 & 10).
        Because footnotes can't occur in html, the Draft 1 document
	turned this text into italicized commentary which begins with
	"Footnote:" to reflect its origin. If we don't do HTML then
	we could turn this text back into footnotes.

	Alternatively, perhaps it would be best of all to just keep the
	text, drop the leading "Footnote:" and make this matter moot --
	because we don't need two separate ways to add commentary in the
	first place. (I favor this latter.)

      . The Figures in Draft 1 uniformly use tables in which all of the cells
        have boundary lines on all sides because that is as good as you can
	do in HTML (the only option is no boundary lines at all in a complete
	table). So, the .pdf is left looking like the .html would have to
	look.

	We could improve the .pdf tables to look better by judicious removal
	of some boundary lines [and I think this could also help reduce the
	size of some of the tables]. Then the .html and .pdf would not
	look quite the same. Do we care? If we don't do .html, this
	question is moot.

  - Word 97 claims to be able to create .html, but in fact it does such
    a lousy job that the result is unusable (and far too botched to even
    consider trying to fix it up by hand). [Headings get translated to
    numbered lists (<ol>), not headers (<.h?>) but not even consistently
    at that.] It is rumored that Word 2000 does a much better job, but
    I am reluctant to "get out front" in using MS's latest bugs. I would
    probably try it if we really, really want to have .html. But if we
    don't do HTML, then there is a lot less hassle for me... 

  - The multiple versions of Acrobat Reader on various systems are
    probably much more compatible than the various browsers, so we
    will not have to worry about IE vs Netscape vs ... differences.

We don't need to make a decision right away, but we probably ought to
talk about this when we review Draft 1.

Thanks,
Ron