foaf:dnaChecksum

I love this property... :)

There are protein databases that use that sort of checksums of amino acids sequences. An example: Dystrophin, a protein coded by the gene called DMD (named against a disease caused by an incorrect version of the gene - Duchenne Muscular Dystrophy) at Swissprot (ID P11532).

Its amino acids sequence begins by: "MLWWEEVEDCYEREDVQKKTFTKWVNAQFSKFGKQHIENLFSDLQDGRRLLDLLEGLTGQ" (Methionine - Leucine - Tryptophan etc.). Swissprot gives "E7236AA93DFAB56F" as a CRC64 checksum of the sequence. Using FASTA format of the dystrophin sequence, and SPcrc, a C implementation of CRC64 by Christian Iseli, I verified the checksum, and found the same (ie "E7236AA93DFAB56F").

It's quite useless that way, except for checking protein sequence integrity after an HTTP tranfer or something like that. The interesting with that is that it may open a way to authenticate people, using their DNA sequence. Well, that's what they all say. I don't think so. It seems far too easy to get someone's DNA sequence... well, except the analysis :) But you can get someone's DNA kissing him (her), hurting her (him) etc etc etc.

Well, that doesn't mean that I won't trust at all DNA authentication. If you try to get someone's blood (verifying somehow this is truely the person's own blood), you can then believe that this person is truely who they pretend to be. Well, nearly. Ahem. Anyway. But I won't trust the "gimme your DNA checksum and I'll trust you" thing over the Net. Well... you can use this as an IFP. Not as a signature.

The positive thing is that we aren't not able now to get easily our DNA sequence or checksum. That leaves us some time for testing!

[PS]: for the one that are interested in computing DNA checksum, there are about 100.000 genes in the human genome, and around 3 000 000 000 base pairs. Using ASCII code, that leaves you with 21 000 000 000 bits. Even if you use a two-bits code (00 corresponding to A, 01 to T, 10 to G and 11 to C, eg.), you get 6 000 000 000 bits. 46 875 000 ASCII characters, 6 Gb of data. Wow.

Additions to MLRDF

[Edit] : MLRDF / MaLiDeRVo has become DOAML.

Some new ideas for MLRDF (see here for the beginning) :

  • new property reads (domain: foaf:Person) in order to indicate that a Person reads the discussions on a list.
  • new property posts (domain: foaf:Person) in order to indicate that a Person posts to a list.
  • new property participates (domain: foaf:Person) in order to indicate that a Person reads and posts to a list.
  • new property moderator-of (domain: foaf:Person) in order to indicate that a Person moderates the discussions on a list.

All of the three properties have a range of MailingList, ie the description of a Mailing List according to this schema.

  • last but not least, new property description-page, that you can use to indicate the URI at which a list is described. It hasn't got any range specified, but it should be an URI or a foaf:Document (or equivalent) describing this URI.