Keeping xml ids the same on round-trips? (was: Re: "Uncompressed XML Files" format variants)

[Friedrich W. H. Kossebau]

Am Montag, 12. November 2012, 22:42:32 schrieb Boudewijn Rempt:
> On Monday 12 November 2012 Nov, Friedrich W. H. Kossebau wrote:
> > > The xml id regeneration does spoil the diffs. But I think that Calligra
> > > formats the XML nicely enough that a line based VCS is still usable. And
> > > I
> > > also don't think that it will make the diffs unreadable and hide the
> > > real
> > > content changes.
> > 
> > In the long run we need to make sure the xml ids are consistent anyway
> > (like Jos said), so that was just an argument for now.
> > Though looking at diffs on the text-file level is not what we should
> > optimize for, or? Still nice to have, I agree, if only for
> > as-small-diffs-as-possible.
> Actually, regenerating the xml-id's on saving was a conscious decision back
> then... One that has cost us a lot of effort in implementation work,
> actually. I don't totally remember the arguments for regeneration, just
> that I have spent hours and hours when working on RDF to make it work.

Oha. Anyone remembers why that conscious decision was done back then?

Right now naively I cannot imagine why one would like to have all ids to be 
changed just because e.g. there was a typo fixed in one sentence and then the 
document saved again.

But I am only thinking of use-cases where one wants to reference an object 
from outside of the document (e.g. a text paragraph or image or table) 
disregarding of the position in the document and its version. There might be 
more, and they might have influenced that decision. Though that solution 
actually disables the usecase as given above :/

Cheers
Friedrich