In a message dated 12/12/01 3:59:49 AM Pacific Standard Time,
aettlinger_at_worldnet.att.net writes:
> <<The subject of "are" is "same-suit 'N-1', both 'N-2's of the opposite
> color, and the other same-color 'N-3'.">>
> That seems to be correct. I'll concede grammatical correctness. You
> tricked me with a plural subject. (Shlomi, I don't consider this
> "flaming".)
Questioning grammar may not be flaming, but accusing me of trickery MUST be a
FLAME!! ;>)) (how the heck do I make a winking smile?)
Did I accuse you of trickery with your ace in a freecell when the only
problem was my lack of careful analysis? %)) (another poor try)
> <<Try this experiment: >>
> I'll take your word for it. It's plausible enough. Just wanted to
> understand the rationale.
PLEASE don't take my word for it. I'm trying to get you out of your MS/FcPro
chains.
> Now, then, if this isn't probing too deeply, what is your approach to
> that other class of moves called "supermoves". By my defintion, a
supermove
> is a singe source-destination expression that permits a human player to
move
> more cards from point to point than is allowed by M/S Freecell.
Your definition conflicts with my understanding, which is based upon
MKeller's discussion of the printed solution to 617. There, he uses the term
for a 4-card column-to-column move using one empty freecell and one empty
column, a move allowed by MS.
> For that
> matter, what is your approach to that subset of multiple-card moves which
> are implemented in M/S Freecell? Does it gain advantage for a solver to
> choose such moves?
I do not know exactly which multiple-card moves are implemented in MS. There
is an explanation on MK's FAQ page, but I haven't tried to reproduce all
possible variants. I just know that MS is limited and wrong. However, a good
solver must be able to produce printed MS solutions for everyday FC users,
while also being able to print robust solutions for the solver group. To that
end, the solver community must learn which moves MS (dis)allows. I'm happy
with MK's explanation if it is complete and can be confirmed.
In a constrained way, one can claim that a solver limited to MS moves has an
advantage in that fewer branches on the game tree produce a smaller
transposition table in many cases. The tradeoff is that a solver using robust
moves may find a solution sooner, which therefore has built a smaller TT.
> Oh, yes, relative to the MD5 question. When you say MD5 is slow, are
you
> referring just to the processing time of computing the hash value? Or is
> there something else inherent to it that makes it slow? Is part of its
> definition the nature of the secondary search? Or is the problem that it
> results in a time-consuming secondary search?
Most of the hash schemes I tried were designed for strings, use a single scan
over the string, and produce indistinguishable results in terms of time and
collisions. My judgement of MD5 was based upon the extra processing and the
size of the 128-bit key. If MD5 has a very low collision rate, then I may be
very wrong. All gracious flames invited.
BR
Received on Wed Dec 12 2001 - 09:12:28 IST