[osis-core] Creating large numbers of ID's

[Kirk Lowery]

As part of the Linguistic Annotation module for OSIS, we are creating a file of "universal" labels for the 
grammar of (first of all) the biblical languages and then for the rest of the world's langauges. Just a small 
goal! ;-)

To accomplish this and in order to normalize data, I am creating what TEI calls "feature structure libraries", 
"feature libraries" and "feature value libraries". All of these are XML elements that need IDs so that we can 
construct, for example, complex parsings of words in marking using these IDs for pointers to the actual label 
strings.

What is relevant to this group is the general question, how does one generate very large numbers of IDs? I'm 
assuming that one does this in a principled way, and in such a fashion that the coder (me, in this case) can 
remember and use the IDs without having to looking things up in a label-ID dictionary, i.e., using some kind 
of mnemonic. I don't expect anyone will recommend that I start with ID="000001" and simply increment from there.

Patrick tells me there are some differences in approaches among us. So allow me to kick off the debate.

How would any of you with strong feelings about this subject advise me to proceed?

<putting on my flame-proof suit> :-)

Blessinngs,

Kirk
-- 
Kirk E. Lowery, Ph.D.
Director, Westminster Hebrew Institute
Adjunct Professor of Old Testament
Westminster Theological Seminary, Philadelphia