From: Ian G. <ilg...@ya...> - 2010-12-27 16:21:26
|
> It's going to take me another couple of evenings - I haven't disappeared, bear > > with... > Hi Yves, Michael and all, I've had a look at your implementation Yves, from a practical perspective using aggregation to preserve the original segment is a feature worth having, particularly to prevent undesirable spontaneous closing of editors. You are right in something that you said in a previous post - if we use aggregation with the SegmentLinker, there is no need in principle to have the LinkedSegmentReferencer. All the functionality can be dumped in the SegmentLinker, which should simplify things. I want to try some experiments with this approach. What I'll try to do over the next couple of days is to modify my branch to get Segments to be linked via the SegmentLinker, and refactor away the LinkedSegmentReference. If this turns out to be successful, I think this will provide a nice simple code we can both use well. I am aware that you need to know for the repeated segment case which segment is the master and which are the slaves. In my use-cases (ordinary linking of segments) there are no masters or slaves, but I'm aware that for repeated segments it's important that there's definite master and slaves. I guess that for repeated segments there should be different rules for the master and slave segments, in that the master segment can be moved but the slaves should not be able to be moved. Cheers, hope you're all having good Christmases and holidays, Ian. |