On Tuesday, 10 August 2021 at 18:53:50 UTC-7 Travis Scrimshaw wrote:
> I am -1 on hashability-on-demand for the suggested reason: it could lead
> to hard-to-track-down bugs in code where we end up calling a hash of
> something that could be *very* implicit, input to a UniqueRepresentation or
> a @cached_method/function. As Nils said, the user has no business hashing a
> mutable object, and we should explicitly tell the user that rather than
> silently changing it. It violates the principle of least surprise.
>
At the moment, it's just a conjecture that such scenarios *might* be
possible. I think we need to see some real-world example where this problem
arises before we further assess how much we need to mitigate against this.
But that brings us back to the problem that I did not ASK for a mutable
object! I wanted an immutable one!
V=VectorSpace(GF(2),3)
v = V.gen(1)
w = V.gen(2)
v.set_immutable() #this is annoying, but OK
w.set_immutable() #this is again annoying
D={}
D[v]=1
D[w]=2
D[v+w]=3 #the error here is uncalled for!
If we're going to insist on people expressing intent on whether vectors
will need to be mutable or immutable, then in parallel with all other
arithmetic structures, vectors should be by default immutable. If someone
wants a mutable vector, they should ask for a mutable explicitly. If we'd
start from scratch I'm pretty sure that's what we would arrive at. So if
we're going for purity of design, I think we'd need to take the leap and go
with that.
That's going to
> Imagine a bank that gave 50% of all your deposits to charity because it
> thought that is what you wanted to do because you checked out a charity
> website, but you only would find out about this when you checked your
> balance. I would not be happy with this bank.
>
and I don't see the parallel with the topic here, so let's leave the straw
men out of it.
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-devel/9a8aab3c-6b68-4b44-be63-c44bab94e12an%40googlegroups.com.
