On 20/12/2021 18:19, Mikael Pettersson wrote:
> On Mon, Dec 20, 2021 at 6:12 PM Fred Youhanaie <
f...@anydata.co.uk> wrote:
>>
>> Hi Dmytro
>>
>> Many thanks for the explanation and the links.
>>
>> So, we save memory when using an improper list instead of proper list of 2 elements.
>>
>> How about 2-element tuples? It seems to be similar to the improper list, but I haven't read text properly yet!
>
> A 2-element tuple consumes 2+1 = 3 words, so it's not as compact as a
> single list cell. It's also faster to determine the shape of that
> single list cell than of the 2-element tuple (for the list cell we
> only need to inspect the handle, while for the 2-element tuple we also
> have to fetch its header and inspect that).
For the digraph, I didn't care about the representation until logger crashed in me!