[CakeML] cakeML wish list

Konrad Slind konrad.slind at gmail.com
Sun Feb 2 17:49:04 UTC 2014


For arrays: a minimal subset of what the ML standard says.
Thus array, sub,update, and length are a decent starting set of
operations. Being able to create and initialize an array in a single
step is nice, which might mean supporting array literals.

Then vectors are essentially arrays with just initialization but no update.
Strings are just an instance of vectors, but vectors of what is the
question.
I don't mind having strings as byte or int vectors, with, as you say,
support
for literals and concatenation.

Konrad.




On Sun, Feb 2, 2014 at 9:24 AM, Ramana Kumar <Ramana.Kumar at cl.cam.ac.uk>wrote:

> Hi Konrad, Freek,
>
> Do you know which operations you really need supported on arrays and
> strings? I imagine initially for arrays we might just do array, sub,
> length, and update.
>
> For strings, are characters also required as a primitive type? String
> comparisons? Should they be Unicode strings? I imagine an initial
> version might implement strings as int arrays (or int vectors) and
> support literals, size, maybe concat.
>
> We're still thinking about the implementation, but I started a new
> branch called "strings" on which the parser handles string literals.
> Probably it makes more sense to do arrays first, though.
>
> Cheers,
> Ramana
>
> On Tue, Jan 28, 2014 at 9:26 AM, Freek Wiedijk <freek at cs.ru.nl> wrote:
> > Konrad:
> >
> >>Anyway, I realize that the cakeML folk are writing furiously for
> >>ITP. But after the dust clears from that, I would like to lobby for
> >>arrays.
> >
> > Me, I really would like strings.
> >
> > And if someone wants me to proofread an ITP paper, I wouldn't
> > say no :-)
> >
> > Freek
> >
> > _______________________________________________
> > Users mailing list
> > Users at cakeml.org
> > https://lists.cakeml.org/listinfo/users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cakeml.org/pipermail/users/attachments/20140202/cc836bc2/attachment.html>


More information about the Users mailing list