[Dspace-general] RE: Version Control

[MacKenzie Smith]

At 08:37 PM 4/5/2004 -0400, David Goodman wrote:
>The more difficult problem is where the author himself posts variant 
>version without indication of which represents the final draft; I can even 
>envision a situation where a multi-author team writes a paper, and 
>the  different authors post different final versions, with possibly very 
>substantial changes.

I think this is exactly the scenario we're talking about. Multiple authors, 
each depositing a copy of their paper in their institutional repository 
(and oblivious to each other, let's say, so without linkages between the 
copies). If they're identical then the MD5 checksums assigned during 
submission would help with subsequent duplicate detection, but the metadata 
will vary widely so that might be the *only* clue that they're copies of 
the same thing (and each copy will have it's own persistent ID, just as it 
might have it's own call number in the print collection). As you suggest, 
if they're "different versions of the same paper" then the library world 
considers them different papers and so would rightly assign different 
identifiers (but wouldn't necessarily know of the other versions to provide 
the links).

This scenario is both possible and likely to occur, but I just can't get 
excited about it as a big problem. Depositing multiple versions of papers 
into institutional repositories isn't something we're seeing a whole lot of 
(although faculty do want to know that they can). I'd advocate that we not 
worry too much about this scenario yet.

MacKenzie/


MacKenzie Smith
Associate Director for Technology
MIT Libraries
Building 14S-308
77 Massachusetts Avenue
Cambridge, MA  02139
(617)253-8184
kenzie at mit.edu