você está querendo enganar quem sugerindo que o objetivo do open source também seja libertar o ciberespaço eliminando todo o software opressivo que subtrai as liberdades que todo usuário merece ter? esse é o objetivo do software livre. nunca foi o do open source. se não almejam ao mesmo objetivo, se não têm a mesma motivação, querer dizer que são a mesma coisa só porque um copiou a definição do outro é ficar no pensamento muito muito raso, é não entender o que realmente importa, nem onde queremos chegar. típico...

Alexandre Oliva started following Vinícius Peixoto.

daqui a pouquinho vou bater um papinho com os presentes no auditório do Instituto de Computação da Unicamp - prédio IC3, sobre (atendendo a pedidos) como ganhar dinheiro com #SoftwareLivre.

then don't panic, don't pay, then sue them for fraud when they fail to live up to their part of the deal

how would you test for NaN without such apparent conventional weirdness a x != x?
a hash compare function can be fixed:
compare(x,y) returns (x = y or (x != x and y != y))
it's not quite efficient as a simple compare, but it shows NaNs are expected and taken into account in some way that presumably makes sense for the application. but maybe it doesn't, and the hash function itself (rather than the compare function) should reject NaNs by raising an exception or using some other out-of-band means to reject them

that view was about real numbers, not FP, FWIW
division is the reverse operation of multiplication
1 x 0 = 0, 5 x 0 = 0, n x 0 = 0
reverse it and you get
0 / 0 = 1, 0 / 0 = 5, 0 / 0 = n
if 0/0 = 0/0, it would follow that
1 = 0 / 0 = 0 / 0 = 5
which would be problematic IMHO
of course, the following is also problematic
1 = 0 / 0 != 0 / 0 = 1
see, the problem is that 0/0 is indeterminate, we can't even tell which set of values it corresponds to, to tell whether they're the same
boolean can't express that
floating point compare flags can, but are there any languages that grant access to them by operators?

Alexandre Oliva started following Cibero.