Re: [Chicken-users] Eggology

[Arto Bendiken]

On 8/23/07, felix winkelmann <address@hidden> wrote:
> On 8/22/07, Sunnan <address@hidden> wrote:
> > Benedikt Rosenau wrote:
> > > Anyway, I propose the following: please keep dependencies between
> > > eggs small.
> > I disagree; sometimes, it seems better to split common code to libraries
> > than to have duplication. Dependencies can be hell, but so can duplication.
> >
> > >  Further, no mutual dependencies (A needs B, and B needs
> > > A) should be created.
> > I guess I can agree with that. If that should happen, here are two
> > solutions to that:
> > Either:
> > 1)  create a third package, C, with the stuff from A that B depends on,
> > so that A depends on B and C, and B depends on C, or:
> > 2) join the two packages.
> >
>
> I agree with Benedikt that dependencies should be kept at a minimum.
> It starts with simple sharing of code but quickly everything ends up in
> a tangle of dependencies that no one can comprehend. "tool" is a good
> example:
>
> tool -> srfi-37, args-doc
> args-doc -> srfi-37, srfi-95
> srfi-95 -> array-lib
> array-lib -> srfi-42, miscmacros, misc-extn
> srfi-42 -> syntax-case
>
> This is insane.

I've also scratched my head on the seemingly overkill dependencies of
some eggs, but OTOH, as has been mentioned, it's not very useful if
every egg has to reinvent solutions to already solved lower-level
problems just to avoid introducing a dependency.

I believe this is simply a case of a situation where more of the
generally useful stuff should be pushed down to Chicken's standard
library, whence all eggs could rely on that functionality always being
available (unless compiled out e.g. when installing a very minimal
Chicken).

Of course, I've no idea if this really is what Felix wants to do with
Chicken. But it's at least an opportune time to be asking these kinds
of questions, given the R6RS situation.

-- 
Arto Bendiken | http://bendiken.net/