Dear all,

Perhaps a real world analogy might be useful. There are comparisons between this discussion and the DNA barcoding debate, particularly on the question of ascribing identity and separating notions of identity from concepts about that identity. In DNA barcoding a string of DNA bases can be generated from a  biological specimen, much in the same way that a taxonomist might generate or apply a taxonomic name. The DNA barcode (like a taxonomic name) might not be unique, but quite separately we can apply concepts to these strings that add meaning (e.g. these DNA barcodes / taxonomic names do / do not refer to the same 'thing', what ever that 'thing' [e.g. a species] might be). The first step to doing this is to uniquely identify all the elements (DNA barcodes or taxonomic names) by applying GUID's to them. Keeping up with the DNA barcoding analogy, NCBI already do this with Genbank accession numbers that uniquely identify DNA sequences. Genbank will give any DNA sequence an identifier independently of any taxon concept ascribed to a particular sequence or even if a DNA sequence is a duplicate (although they try to avoid the latter). The number is just an identifier. 

Given that generating GUID's is easy, and can be done in a distributed manner (e.g. with LSID's) there is no reason why we can't do this for taxonomic names. As in DNA barcoding the hard part will be in ascribing concepts to taxonomic names (e.g. does this group of DNA barcodes / taxonomic names identify the same concept). Some of this can be automated by applying rules across the metadata (i.e DNA sequences or authority data for a taxon name) identified by GUID's. There will always be disputes about these rules, and to some extent there are real question over how much of this can be automated for taxonomic names (uBio seem to be making a lot of headway with this). For me the real debate is over these rules and the concepts behind them, not the (Globally Unique) identifiers.

Regards,

Vince Smith