Conversation started Aug 21, 2012 at 8:12.
Aug 21, 2012 08:12
one caveat about monoid homomorphisms: you have to add the identity-preservation rule separately, the multiplicative property isn't enough (unlike with groups)
@DavidWheeler Ohh, right, That is a good point. Almost missed that.
Artin sends you off like a hitchhiker in a galaxy without a guide.
pretty interesting when you have people to discuss around with, but painful otherwise
however, if a semi-group homomorphism between monoids is surjective, then it does preserve identity.
Okay, nice result. just proved that too.
given any set X, you get a generic monoid, called the monoid of transformations of X, or End(X). i like to think of End(X) as: things you can do (in X). this has a sub-monoid of bijections, Sym(X) or Aut(X), which i think of as: reversible things you can do (in X).
sort of the difference between: one-way processes, and two-way processes.
you can make this more formal by considering M-sets (and their "reversible" cousin, G-sets)
a lot of things can be M-sets...for example computer programs, where you have a concatenation of instructions that operate on a data structure.
Aug 21, 2012 08:35
@DavidWheeler That is what I was thinking of :-)
Irreversible programs belong to End(X) but not to Sym(X)
and hence, it is possible to rollback those instructions if there is an error and such.
And one can then formally define which instructions are reversible and stuff, just by looking at the final output.
exactly. so it is preferable to work with Sym(X), if you can.
No, that point was different. Sheesh, I mixed up two different things. Well, it is getting confusing for me now, I should go back and practise a little myself. Will come back later for more stuff.
Btw, G-sets are just group actions right?
well, obviously sometimes you might want to discard data. that is a one-way process, but it can improve efficiency.
yes, given a G-set, you can explicitly come up with a homomorphism G-->Sym(X), and vice-versa.
equivalently (Cayley's theorem): every group is a permutation group.
@DavidWheeler every group is a subgroup of permutation group
the trouble is (for finite sets), that Sym(X) has order |X|!, which is often much bigger than |G|, so it's not an "efficient representation".
yes, i should have said "subgroup" as "permutation group" often means: a full symmetric group.
Aug 21, 2012 08:46
@DavidWheeler Can I ask you a number theory question?
@DavidWheeler hmm, I know only lang and artin's terminology so may not understand some other normal conventions.
@FortuonPaendrag you can ask, but i may not know the answer.
@DavidWheeler Thanks for the stuff, will read more into it and ask here if i come up with some difficulty.
@JayeshBadwaik lol, that's ok, those are both "classics" so i'm sure you can "get by" only knowing what's in them :)
 
Conversation ended Aug 21, 2012 at 8:47.