From: Dom Lachowicz (firstname.lastname@example.org)
Date: Tue May 20 2003 - 12:33:12 EDT
I think that this discussion ignores the fact that
there are different types of ids, with different uses,
different localities, and different sources.
We have at least:
1) Lists - auto-generated list ids. Any unique
numbering scheme with regard to the set of list ids
will work here.
2) Objects - auto-generated ids. Any unique numbering
scheme with regard to the set of object ids will work
3) Headers/Footers - ...
4) Footnotes/Endnotes - ...
5) Bookmarks - the only interesting case I see. Here,
the ids have user significance (as opposed to
significance only within our piece table). It's not
clear how to handle conflicting names. All of the
following seem to have some merits and flaws to me:
a) Removing old bookmark, using new one
b) Ignoring new bookmark, use old one
c) Sequential numbering scheme within bookmarks
d) Not caring, because, honestly, how often does this
Let's not confuse cases 1-4 with 5 here. 1-4 can be
implemented with a simple global counter, or a set of
them (using 1 per case). Importers only need to call
getNextUID(), *regardless* of what the UID (if any)
might be in the originating file. I'd put the onus on
each importer to do all (if any) mapping between some
"original" ID and the UID the document is going to
Also, let's not just assume that 5.c is what we want
either. 5a-d are all fairly straightforward to
implement, and in my mind, any or all of them could be
the correct solution.
--- F J Franklin <F.J.Franklin@sheffield.ac.uk> wrote:
> On Tue, 20 May 2003, F J Franklin wrote:
> > The real problem is not the generation of unique
> IDs, which is simple
> > enough to do, but rather the merging of document
> fragments each of which
> > has its own set of IDs and ID references. When
> merging documents, the
> > safest method (though probably very difficult as
> long as we use linear
> > import) is to replace all IDs with new
> document-unique IDs and to map all
> > ID references to the appropriate new IDs.
> Thinking about it, what importers need to do is:
> every time an ID or ID reference is encountered, the
> ID is looked up in a
> map to see whether the ID was encountered previously
> and if so replace the
> ID with the appropriate mapping; if this is the
> first time the ID has been
> encountered then generate a new ID, add to the map,
> and replace.
> where it gets slightly interesting is where we have
> user-named IDs (i.e.,
> bookmarks) because how should these be renamed?
> should they be renamed at
> all? if you are copy&pasting then what?
> Ciao, Frank
> Francis James Franklin
> After we were introduced I noticed that both of them
> were carrying
> three-foot-long machetes. I hate hiking with
> convicts carrying machetes.
> - Susan
> Orlean, The Orchid Thief
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
This archive was generated by hypermail 2.1.4 : Tue May 20 2003 - 12:47:24 EDT