At 08:52 AM 6/29/01 +0200, Jeremias Maerki wrote:
>Just an idea, but I don't have enough overview over FOP right now to know
>if it's feasible:
>Why not take a little time to encapsulate the parts to be redesigned, so
>the FOP 2 people can work on a separate repository (or just another
>package subtree)? Consequences:
>+ Encourages loose coupling and exchangability of algorithms
>+ Improvements on either side in the common stuff helps both sides.
>+ Less work to sync common stuff between projects
>+ This is a chance to tidy up some dirty spots and dependencies.
>- Some effort is neccessary before FOP 2 can start.
>- Probably more work for the FOP 2 people because they have to maintain
>  compatibility.

This is _my_ personal inclination, what you describe above, particularly if 
we were to go with the separate package(s) view. I am not anxious to do FOP2 
as a CVS branch in any case - I think a completely new sub-project or new 
drop-in modules is called for. I personally think new package(s) are enough. 
A modified build.xml with new targets should be enough to handle building. I 
don't think we have to duplicate the majority of FOP, so a new _CVS_ module 
is probably unnecessary.

FOP2 is the term for a completely new layout engine, and so maybe we don't 
really want to call the thing "FOP2". Suggestions? I'll call it FOP1+ for now.

(Note: we can certainly envisage a FOP2, which would be a much more thorough
rewrite and redesign, but that concept has not yet been described)

FOP as it stands is pretty good with modularization - not great, but pretty 
good - at a high-level, although it gets rapidly worse at finer levels, 
which is one big reason for FOP1+. I'd like to see us agree on interfaces, 
first off, between the layout/formatting component(s) and everything else, 
that both FOP1 and FOP1+ will obey. We're actually close to that now, I 
think, as far as FOP1 is concerned. That way we can, as you suggest, just 
drop in the complete new layout/formatter when the time comes.

>To ensure that as many people as possible help bringing FOP up to a
>version 1.0 or even 2.0, I think it's crucial to maintain an up-to-date
>list with things to do and with indications where and how non-committers
>can help best. Then people have to take ownership of their tasks which
>also should be tracked.

I had some ideas occur to me last night about this, and will try to develop 
them over the next few days.

We also have a number of FT (or nearly FT) developers at the moment - I 
think their input on this is critical.

Arved Sandstrom

Fairly Senior Software Type
e-plicity (
Wireless * B2B * J2EE * XML --- Halifax, Nova Scotia

To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to