[rfc-i] ISBN and DOI

[John R Levine]

> I suggest that we don't pay ransom for codes. We should end the racket, not
> perpetuate it.

ISBNs and DOIs aren't codes.  They're database keys.

When you buy a bunch of ISBNs from Bowker, you don't just get 100 
thirdeen-digit numbers, you get the right to add 100 entries to Books in 
Print, and to add the prefix of those numbers to the BIP publisher 
database, which is still what bookstores use to find out how to order 
books.  If you're a book publisher, those are quite valuable.  If you're 
the RFC editor, maybe, maybe not, although I think there is some value if 
it lets people in other fields looking for RFCs find definitive online 
copies, rather than ordering $50 printed copies from Global Engineering.

DOIs are also database keys which you can look up online in various ways 
including via the proxy at dx.doi.org.  The name structure is more 
flexible than ISBNs, but it's the same idea, every DOI registered lets you 
add an entry to the DOI database.

The database has a bunch of stuff including a forwarding URI 
for material that's available online.  DOIs seem mostly to be aimed at 
documents which are sold by the copy, whether paper or online, e.g., 
standards and articles from tech journals.  CNRI gives away software 
libraries can use to track what items they've already bought, so when a 
library user looks up the DOI in a catalog, it checks to see if the 
library already has a copy before going out to the pay site.

Again, it's not obvious how valuable this is for RFCs, but again, if it 
lets people find the definitive free copies RFCs more easily, maybe it's 
worth 75c.

R's,
John

Fun fact: the DOI group claims they want doi: URIs, but apparently nobody 
involved has gotten around to writing a draft and running it through the 
process.  Maybe we could offer to do it for them, in exchange for 20,000 
free DOIs for the next decade's RFCs.