From: msevior_at_physics.unimelb.edu.au
Date: Fri Mar 12 2004 - 21:25:28 EST
> On Fri, 2004-03-12 at 19:07, Hubert Figuiere wrote:
>> > > Anyway, I'm in favor of putting native AWML data into the clipboard
>> for
>> > > our internal copy/paste.
>> >
>> > I'm not. We *need* excellent RTF import/export as our primary
>> > compatibility with other WP's. Using RTF on the clipboard enforces
>> this.
>>
>> No. It just screw us up and make us have a data model that stick to
>> RTF's, which by anyway is not the best idea.
>
> Not that my opinion counts for much (yet) but I agree with Hub,
> especially (but not just because of) the following reason:
>
>> RTF import / export should be considered as a lossy process.
>
Actually where AbiWord has features in excess of RTF I've speficically
extended RTF to make the process lossess for AbiWord. Up and including 2.0
I think saving to an RTF was pretty close to lossless.
> Shouldn't AWML be at the center of everything since surely all
> import/export algorithms are available to/from AWML?
>
> That way you only have 1 format to deal with internally, whilst
> everything else is simply an import/export filter away. I assumed that
> was how AbiWord worked - but Martin is implying otherwise.
All imp/exp from AbiWord goes through filters. However the AWML format is
just an XML-ized markup of the AbiWord document, (What we call the
piecetable), so it's extremely easy to write lossless imp/exp for it and
we can extend the format as we wish as AbiWord acquires new features.
When I did the AbiWord table implementation, I actually wrote the importer
before implementing "Insert Table" from the GUI. It took less than 1 hour.
Doing the RTF imp/exp took well over 1 month.
Martin
This archive was generated by hypermail 2.1.4 : Fri Mar 12 2004 - 21:27:18 EST