Subject: Re: Commit: The futures so bright I gotta wear shades :-)
From: Martin Sevior (msevior@mccubbin.ph.unimelb.edu.au)
Date: Mon Nov 20 2000 - 22:50:01 CST
On Mon, 20 Nov 2000, Eric W. Sink wrote:
>
> I hate to dampen your enthusiasm, but I am far less optimistic
> about the use of sections/columns to simulate tables. More
> specifically, I would agree that your approach would work, but
> only for a severely constrained implementation of a tables
> feature set.
>
> For example, how would your idea handle a cell which spans 3
> rows?
>
Whoops! "My" approach would not work here. Clearly I need to study some
code. Luckily there is lots to be had. I'm still happy though that the Abi
layout code can place multiple containers on a page correctly. This alone
makes me happy. Now at least I know we can mostly reuse that code to place
containers where we we want.
> Most of my experience with tables and their design comes from
> web browser implementation. Support for the HTML table
> feature set is actually quite complex. Spyglass found a
> linear algorithm which requires 4 passes. Most approaches
> are far worse.
>
> The table model used by Microsoft Word is quite a bit more
> complex than the HTML model.
>
> Bottom line: Anyone with experience in the design or
> implementation of a table layout system is very likely to
> perceive your remark as somewhat naive. No offense is
^^^^^^^^^^^^^^
No offense taken, it is true I know I am. I have found though that
sometimes it helps to not have pre-conceived ideas about what can be done.
> intended. I would love to be wrong, but my own experience
> tells me that implementing tables is fantastically more
> complication than it appears.
Complex Maths doesn't scare me all:-) I know a lot about that.
> BTW, I *can* empathize with the experience you've just had
> regarding the piece table problem. That code is very, very
> dense. Any problem with the piece table tends to look like
> an immovable obstacle, and I have found myself quite
> depressed when facing one. And then, when the bug gets
> fixed, it feels like the weight of the world has been
> lifted! :-)
>
Luckily it wasn't a Piece Table bug :-). Of those I've found none that I
didn't introduce working through Fields.
Cheers
Martin
This archive was generated by hypermail 2b25 : Mon Nov 20 2000 - 22:50:57 CST