The secure hash of a document can serve many of the functions of a name.
It fails to suggest the topic of the document but so do many other naming schemes.
Its great virtue is that it is easy to see if a document matches its name.
The name is uniquely determined by the document and for practical purposes documents with different content have different names.
It can only name an edition and not a work.
Indeed the name is known only after the document is complete.
I think that it could serve as an URN for immutable things.
There can be no fights over who gets to use a name.
If two people produce the same document then it should have the same name.
Here are a few ideas suggested by hash names:
After a decade of use by many people there might accumulate about a trillion emenda.
This is not an onerous data base.
The indexer of emenda can make money by selling access to the emenda and also perhaps by demanding $0.001 per emendum to add one to the permanent collection.
- An index mapping hash names to locations can be maintained by indexing engines as a byproduct.
- Editors that make small changes to documents can emit small (~30 byte)
emendum packets that include before and after hashes and the details of the change.
Such details can be by absolute address in file.
They can be unidirectional or bidirectional.
These emenda are thrown to the wind and collected and indexed by hash.
If I find the hash name of a document that no-longer exists, but is like an existing, indexed document except for a comma, and the emendum that inserted or deleted the comma is indexed, then I can restore the lost document.
Some purposes might be served by signing emenda but it is not normally necessary.
- Trees or even cycles of variations on a document become available if just one version can be found and the arcs of the tree are available as emenda.
Probably those who collect emenda will organize them into nets and perhaps consolidate emenda.
- This protocol is somewhat synergistic with this business service which you can use to prove that you had a particular document at some particular time.
This fits well with the Data Bank.
Note that there is no problem with “counterfeit emenda” except as they clutter public indexes.
Each emendum authenticates itself.
The simplest thing that I can think of to serve as the name of a mutable work is the public key of a key pair where the author of the work holds the corresponding private key.
The private key is used to sign editions of the work.
It might be strategic to have a convention to the effect that a particular edition supplants some other edition identified by secure hash.
This project addresses a side-set of these issues.
There is available code!
The above is based on ideas introduced here.