Specify zero-th power of zero

[pirapira]

0 ^ 0 might be zero or one. So, the specification of EXP needs to decide the result of EXP on two zeros.

[efaysal]

It looks as indeterminate. However doing some Mathematics,
it turns to be well definite with a real value which is 1.

lim x^x when x --> 0 is equivalent to lim exp(xlog(x)) when x --> 0.

As lim xlog(x) when x -->0, is 0. This is a known Mathematical fact not a convention!

lim x^x when x --> 0 is simply exp(0) = 1.

0 ^ 0 better to be evaluated as 1.

If you decide that 0 ^ 0=1, Mathematics are supporting you totally.
Now it's better to assess if this Mathematical fact is the most suitable to be coded into the software, EVM, for instance.

[chriseth]

You are giving an analytical explanation for an algebraic question. Note that the EVM does not know whether there is anything between 0 and 1 :-)

[efaysal]

I'm aware about that situation, that's why I finished my text with:

Now it's better to assess if this Mathematical fact is the most suitable to be coded into the software, EVM, for instance.

Most numerical software use the fact that 0^0 =1 , using 0^0 =0 instead of, will be a surprising convention ... hopefully it will not be exploited for any thing weird ...

[efaysal]

FYI: 0^0 is still almost always defined as 1, as mandated by the IEEE 754 floating point standard.

You can go with the convention using 0^0 =0, if you think that will change the world of floating point standard.

[axic]

From a practical standpoint, both Solidity's internal constant evaluator and ethereumjs-vm (remix) consider it as 1.

[axic]

@winsvega can you add a test to the official suite for EXP(0,0) == 1?