[PRL] [plt-internal] Video of Clojure talk

[Shriram Krishnamurthi]

>>> Well, yes, but not just that.  He restricts mutation to in
>>> transactions
>>
>> Doesn't every STM have caveats about this?
>
> Caveats, yes.  But not all of them have such a strong restriction.

What gave you the impression that he restricts it?

He said it needs to be a pure function.  He did not say there was any
enforcement.  Indeed, I believe (I may be mixing up two different
parts of the talk here...) someone asked him about enforcement and he
said it was a "dynamic language" (meaning, no).  Someone seemed to say
something that suggested you could enforce even in a dynamic language,
and he agreed.  He then mumbled something about horrendously
complicated type systems (a phrase that is the last refuge of
scoundrels and Lisp programmers) but said he was going to look into a
type system for this over the next few years.

In short, I didn't see any enforcement even of purity, much less
actual commutativity.

Clarification appreciated.

Shriram