Ask HN: Is there a less general purpose Ethereum style platform?

1 points by sandbags ↗ HN
I'm no expert so open to being corrected but a concern I have about platforms such as Ethereum is that they seek to be general purpose. From what I have seen about how smart contracts are written they are often using languages akin to Javascript & C that compile down to a general purpose VM.

This concerns me because something we've learned about software development in the last 70 years or so is that there are maybe a handful of bug free programs of any significance. Software you can't fix is, therefore, a recipe for trouble. And an Ethereum smart contract is, as I understand it, "software you can't fix".

The recent exploit of Ethereum wallet emptying would appear to bear this out but, as I say, I am still learning and perhaps I operate under some misunderstandings that says these problem are less general than I think?

Recently I have been looking into behaviour trees for implementing both game AI and also decision support systems. Something I like about behaviour trees is that they can be sophisticated but, in general, have a smaller conceptual surface space and seem easier to reason about than general purpose programming languages.

(For calibration purposes my experience of building software since the 1980's has been: assembler, BASIC, C, pascal, C++, Perl, Ruby, Erlang, Clojure).

Once you have mastered selectors, sequences, and the notions of success & failure the rest is conditions and actions whose mastery is predicated on domain knowledge.

Editing of behaviour trees can be made very intuitive through outliner style editing and I wonder if they might be easier to automatically verify.

It occurred to me that Ethereum's beginnings (arising out of MasterCoin) were a lot less ambitious and general purpose. Is there a platform that uses something closer to behaviour trees for implementing smart contracts?

0 comments

[ 4.7 ms ] story [ 8.1 ms ] thread

No comments yet.