On Nov 26, 2018, at 8:46 PM, Ian Lance Taylor wrote:
>
> On Mon, Nov 26, 2018 at 5:14 PM Dan Kortschak
> wrote:
>>
>>
>> Question:
>>
>> One of the areas where we do this is in the mat package for matrix
>> arithmetic and linear algebra. We use panics in lieue of a compile-time
>> type check
P.S. given that many functions return a tuple (pointer, error) where the
pointer is nil if error is not and vice versa, I actually also doubt the
usefulness of the Optional type. On the other hand, that makes behavioral
contracts (static or run-time) more relevant.
Le mar. 27 nov. 2018 à 18:05, Ma
>
> What is the zero value of a
> non-optional pointer? (If I were doing a clean-slate Go, I'd probably
> leave zero-values out, but I'm not).
>
I haven't really spent much thought about it. I suppose that would also
fall in the realm of static or run-time checks of behavioral contracts.
>
--
Y
Quoting Marko Ristin-Kaufmann (2018-11-27 10:29:15)
> if you could call template hacks really dependent types...
You really couldn't, they're entirely different beasts.
>I expect most applications to need wider family of behavioral contracts
>-- and I doubt that Go would ever support the
Hi Ian,
I'm actually not in favor of dependent types -- this might have come wrong
"over the line". I find them limiting and confusing in practice (but I have
limited experience, mostly in C++ if you could call template hacks really
dependent types...).
I expect most applications to need wider fam
Odd that the paper you chose to link to is one specifically addressing a
difficulty integrating the ideas with Java's class hierarchies -- while
the broader subject may be relevant to Go, I don't think that paper is.
Wrt Dan's question of doing static checks, I'll quote the paper:
> In principle,
@Dan: P.S. Sorry, I probably should have addressed the email to Tamas
Gulacsi.
On Tue, 27 Nov 2018 at 10:41, Dan Kortschak
> wrote:
>
>> Yes, this was raised as a possibility by Ian. But this negates a very
>> valuable aspect of matrices that I completely failed to consider in
>> writing the qu
Hi Dan,
What you are referring to is called *behavioral types* (see, *e.g., *
https://www2.ccs.neu.edu/racket/pubs/fse01-flf.pdf).
If you want to go that route, you need (usually arbitrarily complex)
behavioral contracts (not to be mistaken with the term type contracts used
here) and static checks
Yes, this was raised as a possibility by Ian. But this negates a very
valuable aspect of matrices that I completely failed to consider in
writing the question: the issue of submatrices and how these are
valuable.
On Mon, 2018-11-26 at 20:49 -0800, Tamás Gulácsi wrote:
> Just a silly idea: what if
On Mon, Nov 26, 2018 at 6:15 PM Dan Kortschak
wrote:
>
> Background:
>
> As part of the Gonum dev team, I have an interest in using Go for
> scientific computing. This stems from long experience with unsafe
> computing environments often used in, particularly, bioinformatics
> (perl, python and R
10 matches
Mail list logo
