This SRFI uses the "it is an error" language from R7RS-small. It means the same as "it is undefined behaviour". Thus, this SRFI does not preclude future SRFIs or standards that decide to specify what happens in some cases of this undefined behaviour.
This SRFI does not include limitations that aren't already in R7RS-small. Am Mo., 25. Sept. 2023 um 12:20 Uhr schrieb Dr. Arne Babenhauserheide < arne_...@web.de>: > > Amirouche <amirou...@hyper.dev> writes: > > That breaks the rule I dub 'the scanning rule': > > > > For people used to reading from top-bottom, from left-to-right; > > > > It is better to read code from top-bottom, from left to right. > > > > The point of the scanning rule is to make the code a bare minimum > > predictable to limit the cognitive effort required to read the code, > > if the scanning rule is applied, the user can focus on things > > that matters. > > There are two main approaches to explain a problem: > > - bottom to top, and > - top to bottom. > > For complex problems top to bottom can be more effective, and not being > able to reference things defined later is a typical stumbling point for > new developers. I stumbled over that when I started, and it was actually > wrong on the slides in the lecture I took at University, so many people > stumble over that. > > But that’s not the point I make: the point is that the SRFI limits what > it does not change. > > If this SRFI would include limitations on definitions and a later Scheme > standard would allow re-ordering non-side-effecting defines, this SRFI > would forbid code that’s allowed in the later standard (and without > need, because the limitation on definitions is just duplication from > R7RS). > > Best wishes, > Arne > -- > Unpolitisch sein > heißt politisch sein, > ohne es zu merken. > draketo.de >
