After you two's discussion, I have some understanding about lexical scope and dynamic scope. I will add lexical binding if my code use it.

[stardiviner]           <Hack this world!>      GPG key ID: 47C32433
IRC(freeenode): stardiviner                     Twitter:  @numbchild
Key fingerprint = 9BAA 92BC CDDD B9EF 3B36  CB99 B8C4 B8E5 47C3 2433
Blog: http://stardiviner.github.io/

On Mon, Jun 19, 2017 at 6:37 PM, Bastien Guerry <bzg@gnu.org> wrote:
Hi Nicolas,

I'm all for lexical-binding, and it's good to have it in Org's core.

The author of ob-sclang.el used "2011-2017" for the copyright years,
which was obviously a typo and tells that the header was simply copied
from another file (which is 100% fine btw).

From that, I inferred that the "lexical-binding:t" was also copied
without further thinking, especially since there is no binding at all
in this file.

I think Stardiviner is the one who should make the decision, but I
don't see what "lexical-binding:t" would add to his actual code.

For the more general concern: again, I'm all for lexical binding and
I'm well aware of its numerous advantages, but I don't think we should
rule dynamic binding from contributed Org code.  Dynamic binding has
it's limitations, but when used carefully, it also has the advantage
of being easier to grok for beginners.  We want to welcome beginner's
contributions.  So I simply recommand lexical binding for Org's core,
and what fits developers best for Org's contributions.

And I agree we can move on to something else :)

--
 Bastien