response 1 of 15:

Jul 15 04:49 UTC 1993

"Gaston Gonnet and Steven Benner of Zurich, wondered recently, 'What is
the longest word spelled out in the sequence of a protein in the protein
sequence database using the one-letter code for amino acids?' The
longest they found were two words containing nine letters each.
To get the answer, Gonnet and Benner matched the Oxford English
Dictionary (unabridged) against the SwissProt protein sequence database
(verson 23). Twenty-three minutes of computational time produced the
two nine-letter words, hidalgism and ensilists (plurals apparently
qualified). Hidalgism, the authors report, means the behavior of a hidalgo,
or Spanish gentleman. In proteinese, the word appears at position 247-255
on the integrase of bacteriophage lambda. Ensilists are people who (sic)
preserve crops by ensilage, or storing green fodder in pits or silos. The
word appears at positions 81-89 of the PRRB protein from Escherichia coli.
Gonnet and Benner conceded in their letter to Nature that the two words
are candidates for the most unusable pieces of information to be found
simultaneously in lexicography and biochemistry."

response 3 of 15:

Jul 15 13:38 UTC 1993

Just right for summer reading, right?

response 5 of 15:

Jul 16 01:06 UTC 1993

It *would* be interesting to see the frequency of all "word" matches
between the protein database and the unabridged. Do you think the NIH
or EFA would issue a grant to carry out the work? It would be multi-
disciplinary.

response 7 of 15:

Jul 16 04:35 UTC 1993

Yes, *you*, with your taxes, in the SETI program. But I think SETI is
much more worthwhile - there probably are intelligent beings "out there".
It would be analogous to when dinosaurs were discovered: they really
shook us up, but there is absolutely nothing we can do about them. 
Now, to turn "drift off"....we're looking for useless language facts,
(especially if they cost a lot to document?).

response 9 of 15:

Jul 16 13:16 UTC 1993

Well, they're not "facts", but you could note the attempts to reconstruct the
(hypothetical) common ancestors to the (mostly theoretical) common ancestors
of modern languages.

response 11 of 15:

Jul 19 13:49 UTC 1993

The attempts are somewhat interesting.  The idea that they have any
relationship to anything anyone ever actually spoke shows (IMO) a total
ignorance of the theory of probability.  Throw together a whole bunch of
probabilistic inferences, and even if each one is (for the sake of argument)
very likely, the combination quickly becomes very unlikely.  And in this
field the ratio of data to inference is quite low, I think.

response 13 of 15:

Jul 19 17:22 UTC 1993

Yes (I think), except that we'd also have to have no oral history (stories
about the Romans etc.) and a lot less in the way of artifacts that might
indicate cultural connections.  (I'm guessing that the oral history part
may not be very important - I think the medieval conceptions of the ancient
world were based heavily in written records, but I could be wrong - but
that archaeological remnants of high cultures could add quite a lot.)

response 15 of 15:

Jul 20 12:40 UTC 1993

I was thinking more of the rather incredible medaeval legends surrounding
(say) Vergil.  (Anyone familiar with the fantasy _The Phoenix and the Mirror_
by (I think) Avram Davidson?  It's built around Vergil as in the legends not
as in reality.)