Mark S. Miller

I know I am without context beyond my E understanding. But from the title this sounds like something I should understand. Could you explain, starting from nothing but knowledge of...

Was this ever implemented in any E (attn @kpreid ). I do not think it was implemented in my E-on-Java. In any case, I am in favor. What would the...

Does Monte have the same distinction between % and %% that E does? IIRC, the E modPow happened when using *\* with %%, not %. Is this right?

Weak references are a source of non-determinism that is actually worse than most regarding side and covert channels. Thus, they need to be closely held. See https://github.com/tc39/proposal-weakrefs/blob/master/specs/weakrefs.md#portability-and-observable-garbage-collection by @dtribble for...

WeakRefs are necessary to do distributed objects with reasonable garbage collection. WeakMaps are necessary for good gc across local membranes. Neither mechanism subsumes the other. Sorry to speak in brief...

Sorry for continuing to speak in riddles for now. But doing so leaves a record of what I need to clarify when I have time, so... It would be ok...

E did indeed use only WeakKeyTables (under some name) based on weak references, which were weaker than ephemerons. As a result, E membranes had unpluggable leaks I did not notice...

Interesting. I had nor heard of quote notation before. However, [the Wikipedia page](https://en.wikipedia.org/wiki/Quote_notation) says: > While storing the integer 802787 in binary requires 20 bits, the storing of the rational...

From the phrase "Weakmap key / WeakRef or FinalizationRegistry target (aka observed)" I'm concerned that this may be misunderstood. Even without WeakRefs or FinalizationRegistry, merely using them as a WeakMap...

Why? What's breaking?