Jonathan Rochkind wrote:
>
> But I agree that encoding format is not our fundamental problem. First
> we've got to be clear about what data we have/want/need, and THEN we can
> decide how to encode it.
Good, one step before the next. So let's really stop the encoding
debate here and now and concentrate on the question of what it is
we need to have in our records.
Years ago I had a chat with Stu Weibel and asked him what implementors
did with the 15 DC elements. He said well, mostly what they do is
just keyword index the whole stuff. OK, said I, then let's now jot down
a new model on this beer mat (that was in a pub in Göttingen, Germany)
and call this new model the "Göttingen Kernel": consisting of 1 element.
Unfortunately, we got carried away on other matters before actually
scribbling the new model down, so it got lost.
B. Eversberg
Received on Wed May 23 2007 - 09:19:27 EDT