[programnature]

To be fair, fast immutable hash tables have only been around as a technology since around 2005. (Yes, Mathematica is mostly based on immutability) .Other than internal R&D, Wolfram never add experimental methods to Mathematica.

Of course, its always had lists of rules as a far more general (though much slower) form of the hash table concept.

[gohrt]

I assume you mean persistent, not immutable.

    a[x] = 1
    b = a
    a[y] = 2  <---- mutates a
    a[x] = 3  <-- is this legal?
    b[x]    <--- persistent- uh, what is Mathematica's semantics here?

[taliesinb]

Yes, Associations use the same basic algorithm as Scala and Clojure's hashmaps, and are indeed persistent.

And b[x] there gives 1 as you would expect.

[davorak]

Do you know when that change was put into place? I have been using Mathematica sense version 5 and my intuition was that `b[x] == 3`. Testing in mathematica 8 which I have on hand confirms this.

[taliesinb]

Mathematica 8 doesn't have associations, so you must be testing symbol assignments.

Write a = <||> before running the above code to test this on associations; associations are persistent, the symbol table is not.

[programnature]

No, I meant lists of rules like {a->1,b->2}, which capture the same kind of patterns as in the previous example but in an immutable (and copy-on-write) way.

[anaphor]

just to clarify, you mean fast immutable hash tables that support insertion right? (Which I agree is hard) Otherwise you can just use a BST or something and it's fast enough for lookups.