I am not arguing to exclude TinyScheme and SIOD from get.scheme.org <http://scheme.org/> . My point is simply to assign a strict meaning to the label “implementation of Scheme”. The get.scheme.org <http://get.scheme.org/> page should be organized as a table that gives for each implementation of Scheme (and variants) which features of the RnRS they support (and perhaps to what extent). I remember seeing a page for this (is it somewhere on scheme.org <http://scheme.org/>?). The current get.scheme.org <http://get.scheme.org/> page is somewhat like that already with labels for each RnRS supported, so some implementations could have no label, or perhaps a “low” score in the column corresponding to the closest standard. There could also be two sections. The first is for RnRS conformant Scheme implementations and the other for “partial implementations of Scheme”. That way, people new to the site will have a rough idea of what they are getting into when they get a specific implementation. Marc > On May 14, 2024, at 3:09 PM, Stephen De Gabrielle <xxxxxx@gmail.com> wrote: > > I feel we are poorer - and doing potential schemers a disservice - if we continue to exclude TinyScheme and SIOD from https://get.scheme.org because, TinyScheme only does a subset of R5RS, and SIOD only Implements the scheme from the Lambda Papers. > > Are we lawyers/philosophers/pedants arguing over definitions? > or are we scheme enthusiasts who would like others to get the benefits we get when using(or implementing) a scheme for our applications? > > Best regards > Stephen > > PS. In case I was not clear - the standards and SRFI information is essential - I am not suggesting we 'throw the baby out with the bathwater'. > > On Tue, May 14, 2024 at 1:48 PM Marc Feeley <xxxxxx@iro.umontreal.ca> wrote: > I’m somewhat of a hardliner on the subject of what is required for a system to qualify as a “Scheme”, which for me means an implementation of the Scheme language. > > A “Scheme” should be defined as an implementation that conforms to one of the RnRS reports. Others should qualify as “Scheme subset”, or “Scheme like” implementations. I think we all agree that small deviations from an RnRS report are OK. But some features are essential, such as lexical scoping, proper tail-calls, and continuations. Just having parentheses and being simple does not qualify as “Scheme”. > > By the way, Ribbit would qualify as a Scheme since it fully conforms to R4RS. > > Marc > > > > > On May 14, 2024, at 1:55 PM, Jakub T. Jankiewicz (via schemeorg list) <xxxxxx@srfi.schemers.org> wrote: > > > > On Tue, 14 May 2024 09:23:40 +0100 > > Stephen De Gabrielle <xxxxxx@gmail.com> wrote: > > > >> I’m very tempted to try make a case for calling ‘Snap!’ a scheme dialect: > >> https://snap.berkeley.edu (looks like another block programming clone of > >> scratch, but has scheme capabilities like first class functions) > > > > Snap! may have similar semantic to Scheme but I would not call Scheme > > a graphical/block language, it's not even lisp. The same JavaScript, R or Ruby > > have lisp similarities but calling those languages lisp would be overuse of > > the term. Some people call them lisp though. > > > >> > >> Indicating if a scheme supports RnRS or SRFI’s is obviously very useful. > >> > >> Scheme is a living (and evolving) language with a passionate community in > >> addition to implementation communities. Let’s not exclude any. > >> > > > > -- > > Jakub T. Jankiewicz, Senior Front-End Developer > > https://jcubic.pl/me > > https://lips.js.org > > https://koduj.org > > > >
