Re: [MIT-Scheme-users] let-fluid

[Matt Birkholz]

> From: Taylor R Campbell <address@hidden>
> Date: Tue, 25 Aug 2015 12:57:29 +0000
> 
> [...]
> This is another fluid compatibility issue -- in Git, FLUID-LET as
> temporary shallow-binding of global variables has largely been
> replaced by local deep-binding of objects called fluids, but nobody
> has made an effort to maintain any semblance of compatibility (yet).

Would a "semblance" of compatibility resemble adding new bindings
named e.g. $flonum-unparser-cutoff (a Scheme48 naming convention) and
removing the old ones (e.g. flonum-unparser-cutoff)?

We would at least get better errors -- about the FLUID-LET and not
about internal procedures stumbling over a fluid object.

More compatibility than that will be elusive.

When I proposed this incompatibility, I asked for alternatives.  All I
remember was encouragement to "bite the bullet".  Are we having second
thoughts now?

> From: Taylor R Campbell <address@hidden>
> Date: Fri, 17 Jan 2014 13:49:09 +0000
> 
>    Date: Thu, 16 Jan 2014 20:00:54 -0700
>    From: Matt Birkholz <address@hidden>
> 
>    If we want SMP(?) and don't want it in a distant fork, we might just
>    bite the bullet and replace our fluid bindings with fluid objects
>    (like e.g. s48's). [...]
> 
> That's what I have been intending to do for ages.  For large sets of
> fluids like the compiler uses, we could merge them into one fluid with
> a large data structure.

> From: Chris Hanson <address@hidden>
> Date: Fri, 17 Jan 2014 22:40:01 -0800
> 
> R7RS defines parameters, which we'll eventually need to implement
> anyway for compliance.  Converting the system's bindings over to
> parameters is probably the right thing for a variety of reasons.