Re: More HTML3 DTD

Michael Johnson (
Thu, 01 Dec 94 13:43:38 EST

Dave Raggett writes:
>Michael Johnson thinks its difficult to implement VALIGN in table cells:

Yes, I do.

>There are no HEIGHT attributes defined in the draft HTML 3.0 dtd for rows
>or cells for that matter.

Which was my error and which does not mitigate the problem with VALIGN at all,
it simply reduces the list of possible algorithms for handling it.

> The vertical offset for each cell is worked out
>trivially at display time, as by then the row height is known (you don't know
>it until you have formatted all cells in a row).

Trivially? What are you going to do, adjust the position of the formatted
elements when you get the window exposure for the table? If that's not what
you mean, then where is the difference between formatting and display?

I can think of one possible shortcut for this problem, which is to store
references to the first and last visible element in each cell as a row is
being formatted, and then to go through and adjust the vertical position of
those elements when the whole row has been formatted, and the total row height
is thus known. It's ugly, but it would work.

> This is another example of
>how in computing you should put until tomorrow what you can't do today!

No kidding.

Michael Johnson
Relay Technology, Inc.