> They decided it's a good idea to use ternary logic instead of binary logic

This is the one main thing that keeps me very, very skeptical regarding IOTA (and the coordinator thing, although that may be resolved someday when traffic increases so they can remove it - at least that's what they say).

It looks a bit like they had this JINN processor developed which appears to be ternary in nature, but it went nowhere, and then they pivoted that into IOTA and stubbornly kept the ternary stuff because that might revive this JINN thing again later on, if IOTA catches on. At least that's how it looks to me from the outside. It's a bit hard to get reliable information about this JINN thing, as the company that apparently developed it, Jinn Labs, is nowhere to be seen publicly, with the excuse of being in "stealth mode". As far as I can see, this entire thing could just as well be vaporware as it could be revolutionary tech developed under strict NDAs.

I'll definitely stay very, very skeptical for now, as it's so hard to get reliable information about IOTA aside from the bling-bling on the surface. I also tried to look at the code of their reference implementation on GitHub a bit recently, since it's mostly Java and I'm mostly a Java dev for a living right now I was interested, but the code quality I can see there is pretty bad overall (like there are almost no comments, most comments you find are actually code that was commented out, formatting seems incoherent, huge amount of bloat code even by Java standards among which it's hard to actually find any real logic), so that excursion also did achieve nothing with regard of alleviating my skepticism.

So much misinformation here. The Jinn project is in active development.

Have you anything more substantial as proof than your word for it?

>  the IOTA developers had written their own hash function, _Curl_

Did...did they intentionally name it curl, so that when you search for "Is curl secure?" you will find articles saying that curl - the widely used library - is secure, in the hope that people will confuse the two?

I know that you shouldn't assume malice when it can be explained with incompetence, but combined with some of the other points here, I can't help but feel that this was intentional.

To be very honest, when I first heard about IOTA, there was an article that had the phrase "the inventor of curl", I incorrectly assumed for a while that Daniel Stenberg was involved in the project.

I think the insecurity of their function was a mistake. I doubt they would deliberately code something that threw collisions and have a cover up plan that is basically name it curl and hope people get it mixed up with the real curl.

fwiw, was bored and exploring for the first time the (my God) toxic social timelines of these crypto coins and happened to end with IOTA and the story of the home grown hash function. Just posting what I just read about this topic and not an +/- either way on the topic at hand:

Sergey Ivancheglo (@Come-from-Beyond) claims the collisions were intentional and for the purposes of IOTA the hash merely had to be a one way function:

https://gist.github.com/Come-from-Beyond/a84ab8615aac13a4543...

He addresses his base-3 number system design decision as well (by appealing to authority no less :)

> name it curl

In above he references Curl-P and the final letter (see his last gist) also explicitly asks that the MIT security blogger use the full name which she declined.

The other hash function they use is called Kerl a homonym of Curl.

It gets better... their "improved" version of "curl"? Kerl.

You can't make this shit up.

kerl comes from the fact it wraps & extends keccack hashing library. So the name is actually quite fitting.

Kerl is Keccak I.E. SHA-3, the international NSA standard. They called it Kerl for fun in homage of Curl, which is still under active development with the absolute world-leading cryptographers of lightweight cryptography. Curl had to be invented to push LIGHTWEIGHT cryptography which is necessary for the Internet of Things. It's quite astonishing how much misinformation is spread around.

https://blog.iota.org/iota-foundation-hires-cybercrypt-615d2...

SHA-3 is not an NSA standard. It was invented by Guido Bertoni, Joan Daemen, Michaël Peeters and Gilles Van Assche who are researchers at various companies/universities and are from Italy and Belgium.

Your curl function is not even listed as one of the major lightweight crypto primitive on Lux' zoo: https://www.cryptolux.org/index.php/Lightweight_Hash_Functio...

Your post is full of shit.

I am very curious about what 'LIGHTWEIGHT' cryptography is defined as. I am also dubious about anyone that claims to have 'absolute world-leading cryptographers' since many strong cryptographers are quietly employed by intelligence agencies and most others are academics.

Also, since we are being pedantic, SHA-3/Keccak is an NIST standard, which is a federal agency of the United States.

I'm not an expert in the field, but there certainly have been efforts here and there to make "lightweight" crypto that needs little computational resources (and therefore battery power). One example would be KASUMI[0].

[0]: https://en.wikipedia.org/wiki/KASUMI

This kind of post doesn't work when cryptographers read it.