Re: [Dragora-members] Relation of Qi and Qire

[Matias Fonzo]

El 2020-06-12 13:18, Michael Siegel escribió:
> Am 12.06.20 um 02:06 schrieb Matias Fonzo:
> > El 2020-06-11 03:48, Michael Siegel escribió:
> > > Okay, but Qi could still simply run Qire as an external command when
> > > necessary and then work with what Qire returns, right?
> >
> > I guess this could be confusing for users.  Since these are separate
> > commands, it's best to invoke them separately (if we keep the current
> > model of two separated commands).
> >
> > > I think this would probably the best way to go about this.
> >
> > I think it's better to make use of the current interface that Qi 
> > already
> > provides, add whatever needs to be added to the Makefile, and invoke 
> > the
> > "extensions".  The interface would do the following, if Qi has
> > "upgrade", it is complemented with "update" of the Qire extension, and
> > so on...
> >
> > (This was not done before because Qire had to be developed.).
> >
> > Apart from all this, whether it's done or not. I think it will be more
> > practical for users to have to invoke a single command here for 
> > package
> > managing, and the finer grain in Graft.
>
> I'm afraid I don't really understand what you mean. It kind of seems as
> if we were saying the same thing, but then it kind of doesn't.
>
> So, we agree on it being the best for the user to only have to use a
> single utility for package management. Now, I'm not sure whether your
> comments above imply that this utility should be Qi, or if they imply
> that it should be Qire because of how things work together.

I understood that you want Qire's ability in Qi, but I also understood 
that for a better understanding of both things we should unify or extend 
with appropriate languages since the package managing is something 
complex, as Qi currently works well, maybe there is no need to rewrite 
Qi now (and I hope not :-).

The point of having only one command leads to think that Qi could invoke 
the "extensions" of Qire, by this I meant part of Qire, not the whole 
interface, taking into account everything Qi already has, including the 
documentation format.

This has the advantage that parts of "Qi" will be in Fennel/Scheme/Lips 
or whatever, while in the future parts or modes of Qi could be replaced 
in the same language (if needed).

On the other hand, we still don't know how users will interact using qi 
and qire, it might be positive (I hope).