Subject: Re: from AbiWord to AbiSuite [was Re: A new draw on XP refactoring
From: Martin Sevior (msevior@mccubbin.ph.unimelb.edu.au)
Date: Fri Feb 01 2002 - 16:30:46 CST
On Fri, 1 Feb 2002, Tomas Frydrych wrote:
>
> The discussion of the drawing/SVG application is quite exciting,
> and I certainly do not wish to stifle it, but there is one issue that
> slightly worries me. The AW 1.0 release is not going to be the end,
> more like the beginning. One of the reasons we are releasing 1.0 is
> that we have learned a number of lessons, and among them that
> the current layout engine is reaching its limits and that to
> implement some features that the punters want, we will have to
> seriously redesign it, too seriously to do it overnight between 0.9.x
> and 0.9.x+1. The 1.0 version will be rock solid in what it does, but
> it will not be feature-full from many a user's point of view. It is only
> when we have at least table and footnote support (and perhaps
> some other stuff) that we will have a wordprocessor that will stand
> a chance when measured against TheOtherWordprocessor (and
> that day is coming!).
I fully agree Tomas. It will be important to be not too distracted from
what we need to do in abiword. However we really need this vector
graphics/equation editor/simple drawing code/ to be a competitor in the
field of the "The world's best word processor". The fact that there is so
much nice GPL code out there already in this helps us enourmously.
>
> What worries me is that the current development team is rather
> small, there is only a handful of people who contribute code on
> regular basis. I would hate for our resources to be to much further
> reduced by starting work on another member of the suite before the
> 1.1 code base is well on the way.
Yes. I know I'm jumping the gun a bit on this but I think we should agree
on the most pressing issues for post 1.0 abiword and to see if there are
volenteers to implement them.
For me these are:
1. Move to GTK 2.0 and redo our font mess to take full advnatage of the
much better architecture in GTK 2.0.
2. 16 bit unsigned char => 32 bit unsigned char to allow 100% unicode
compliance.
3. Redo containers/page to handle tables/frames.
4. Implement component embedding/embeddable abiword.
This point really needs a working vector graphics implementation.
5. Argue about file format.
I don't think we should adopt the full OO spec. This would slow down our
native file saves (which are blindingly fast right now.) We would also
have work around the fact that we do things like lists much differently
than OO. I think we should steal things from the OO file format that we
need instead. Maybe their Tables/Frames description? I don't know yet.
I haven't had time to evaluate what they've done there.
Cheers
Martin
This archive was generated by hypermail 2b25 : Fri Feb 01 2002 - 16:30:53 CST