At 04:49 PM 4/30/01 +0100, Tomas Frydrych wrote:
>For my purposes until AW 
>supports footnotes it is not a usable wordprocessor, and for many 
>people the same goes for tables. 

Unfortunately, that kind of argument leads you down a very slippery slope.  
Everyone has their favorite feature, and we've all been trained to not pay 
attention to products which don't have the kitchen sink and two swiss army 
knives thrown in.  (But wait, why use that product when mine has a third 
swiss army knife?  Yeah, well *mine* has three swiss army knives *and* a 
spork!)

When talking about specific features, it would be nice to have market 
research data, but we don't.  So here's a free substitute.  

Most academics who use word processors need footnotes.  Most word processor 
users are not academics.  Footnotes are used outside academia, but not that 
often.  My spouse is an academic, and I'd love to have footnotes for her, 
but I'll ship 1.0 without them.

Tables are slightly more plausible, but in practice they tend to be very 
hard to use.  (By comparison, styles are much easier to implement and use, 
but most people don't use them either.)  Think about the UI choices a user 
faces when they want to create simple tabular data.  

  - If single-line cells and regular layouts will do, then a tabs-based UI 
    is much easier to learn.  (As Martin has recently demonstrated, just 
    draw lines around the result and you're happy.)

  - If you want a few big parallel blobs of text, then a columns-based UI is 
    easier to learn, too -- especially when you learn that you can do 
    section breaks on the same page. 

The place when you really need and want to learn how to use a true tables UI 
are when you want to do rich complicated tables, with overlapping cells, 
multi-line cells, fancy borders/shading, etc.  And as soon as you get 
started, you're not likely to be satisfied with anything less than a 
full-blown tables implementation, a la Word or HTML.  (To be honest, I'm not 
sure which is worse.)

The day we say we support tables in AbiWord, I guarantee you that someone 
*will* dig out the nastiest table layout they've ever seen in a Word 
document and try to open it in AbiWord.  If we aren't pixel-perfect on that 
day, our product doesn't Just Work.  Indeed, on that day, they should also 
be able to:

  - save it back out and view it in Word again,
  - copy it to the clipboard (in a lossless way), and perhaps even
  - cleanly import an HTML table copied from the clipboard. 

Anything less simply won't make tables users happy.  

To date, anyone who's seriously investigated the work required to add 
full-fledged tables support to AbiWord comes away with a healthy respect for 
what a sizeable job it'll be.  (Doable, but with a *lot* of backend work and 
even more UI work.)  Unless you've been looking at the problem to say 
otherwise, I'd expect that tables could easily take as much as a a year to 
reach that point.  

In the mean time, I'd hate to force everyone who doesn't really need tables 
to do without a rock-solid version of AbiWord.  

Paul

PS:  One of the benefits of releasing 1.0 without tables support is that 
it'll attract attention from developers who want to prove that I'm wrong, 
and that tables are much easier to implement than I think.  I *love* to be 
proven wrong like that, but they will need to prove it.  ;-)  


Reply via email to