Draft proposal for a Mersenne related query language

$Id: querylang.txt,v 1.2 1997/06/10 19:30:18 wedgingt Exp $

Form of query from client to server:

Each query contains one "choice" from each of the following
        categories:

Category A:
# of exponents to provide data on &/or reserve at once,
        maximum of 1000 (?)

Category B:
0 smallest exponent
1 next exponent after N (N specified in query)
2 previous exponent before N (N specified in query)
3 largest exponent
4 w/ smallest trial factor not checked
5 w/ smallest k not checked (in 2*k*exponent + 1)
6 w/ smallest remaining cofactor
7 w/ smallest remaining trials to prove cofactor prime
  (not that it ever will since such a cofactor would
   have already been proved prime by faster methods)

 (3-5 ignore exponents with no data at all)
 (5-7 are limited to first 100(?) exponents
  not prime nor completely factored ?)

Category C:
0 ... that is known to be prime,
1 ... that is composite and completely factored,
2 ... w/ at least one (known prime) factor,
3 ... w/ two matching LL residues,
4 ... w/ one or mismatching LL residues,
5 ... w/ no residues, only trial factor data,
6 ... w/ no data at all,
7 ... , don't care what info on the exponent already,

Category D:
0 ... return all info on the exponent and
1 ... return partial LL test data and
2 ... all (known prime) factors and
3 ... smallest trial factor not checked and

Category E:
0 ... don't reserve: query only (unaffected by reserves)
1 ... reserve: will return new work later or cancel
	reserve immediately (skips other reserves)
2 ... cancel reserve of N; here's new work, if any
	(require at least one "pass" of a trial factorer?)
3 ... continue reserve of N; here's new work (checkpoint)
  (2-3 may want to be separated out from the queries since
   they give new data to the server rather than query it)


Data on each single exponent:

 primality:
 0 unknown
 1 one or mismatching LL residue(s)
 2 at least two matching LL residues
 3 at least one (known prime) factor
 4 completely factored
 5 prime

 LL residue(s) (if known; only matching residues are reported
        to queries) (otherwise report a count of residues?)

 list of (known to be prime) factors

 smallest trial factor not checked by "bug-free" program

 (smallest trial factor not checked by "buggy" program?)

 (last two may be smaller than one or more factors if
  those factors are not from trial factorization)

 (attributions? methods? other gaps in trial factoring?)