Hmmh... interesting way of including comments.

@kirma At first I thought they might be automatically inserted translations for the built-in functions, not user comments. Then Google Translate didn't turn up anything useful and the characters under E seemed too numerous, so I thought maybe you are right they are user comments. Then the repetition under the two Pis made me think maybe they are automatic translations again.

I'm up too late though. Good night

@MichaelHale Hmm... probably just an example of future commenting system.

@kirma Maybe, but it does seem peculiar there is basically one comment per built-in function.

@MichaelHale Sure. That image is from SXSW talk... it didn't really became that clear to me how it's supposed to work.

Yeah, I just found it. His wording does lead me to think it will be an automatic annotation feature. Like for when you send your code to someone who doesn't speak English.

@MichaelHale We'll see... not that I care much of that feature.

Eh. FindIntegerNullVector[Log@{10, 2, 5, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97, 101}] succeeds, but FindIntegerNullVector[Log@{2, 5, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 10, 89, 97, 101}] fails. The only difference is in ordering of the list, which should not matter.

(I wrote a bug report of this already.)

@kirma If you add N@ in the beginning both works, but the first gives a different result than it does without N@. Is this expected?

@Pickett Second numerical result is rubbish. Both of above are effectively factorings of 10 over set of primes.

I suspect second instance is suffering from some sort of internal numerical instability also in the case of non-N variant.

@kirma Why do you think it's rubbish? Log[a].b is ~10^(-14), same as for the first one.

Well, numerically it's "ok", but original infinite-precision statement doesn't hold any more.

And in this case deviation is quite a big one.

Also, it doesn't actually any more use 10 in the calculation at all.

ok

The above is smallest example of this problem I found for prime factoring with FindIntegerNullVector. Of course this is not sensible use case on its' own...

@kirma in WTC2014 they showed this. They want to make the language more intuitive in another idioms. So you can turn on these explanations. They are not user comments.

@Murta "In another idioms"?

If those are just static predetermined annotations... well, maybe there's market for it in China or something.

@kirma yes. I think the idea it exactly that.

OK...

Fields medalist explains how science works in practice... :)