[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Scheme-reports] Scheme-reports Digest, Vol 10, Issue 1
- To: "Thomas Bushnell, BSG" <tb@x>, "scheme-reports@x" <scheme-reports@x>
- Subject: Re: [Scheme-reports] Scheme-reports Digest, Vol 10, Issue 1
- From: "Larry D. Lee jr." <llee@x>
- Date: Fri, 29 Oct 2010 17:05:47 -0400
- In-reply-to: <AANLkTimCa9A4fukWGzNXmKdYY2DBHt+1GKFwyvzwuBZ=@mail.gmail.com>
- References: <dnx7ef8a58fqdp0yywes5jtp.1288205285660@email.android.com> <AANLkTimteqJugzqpEQ9rEgWaf8b7M1b=WkCR0YJFO2Gf@mail.gmail.com> <1288375848.3522.13.camel@motoko> <AANLkTimCa9A4fukWGzNXmKdYY2DBHt+1GKFwyvzwuBZ=@mail.gmail.com>
I agree that Scheme-1 should not provide filesystem, networking, or IPC
support.
Scheme-1 should be a minimal, but complete, language that can be
implemented in embedded environments that lack these features.
However, many applications rely on filesystem, networking, and IPC
operations.
If we do not specify filesystem, networking, and IPC interfaces, as part
of the languages core specifications, implementations will provide
support independently. These implementations will probably diverge, and
the portability of Scheme programs will suffer.
Scheme-2 should define specifications for these interfaces; and these
features should be optional, to allow it to target environments that do
not have these features.
I'm not arguing that Scheme-2 should go beyond what's established in the
Posix standard.
The Posix interface has found wide acceptance within the broader
programming community, and it provides a relatively complete interface
to many system environments.
On Fri, 2010-10-29 at 16:35 -0400, Thomas Bushnell, BSG wrote:
> Based on long experience, I am skeptical about the ability of language
> designers to adequately model operating system facilities without
> becoming operating system designers in the process, and usually
> failing to provide adequate access to the operating systems already in
> use.
>
>
> I would be happy if Scheme-2 said "this is how we map to Posix.1
> facilities", and very unhappy if they started deciding what a good
> networking interface looks like.
>
>
> My primary concern is that Scheme-1 shouldn't have any of this. I'm
> happy to see files removed from Scheme-1, but the
> backwards-compatibility expectations for Scheme-1 might make that
> implausible, and I'm content to see them stay.
>
>
> Thomas
>
> On Fri, Oct 29, 2010 at 11:10 AM, Larry D. Lee jr. <llee@x>
> wrote:
> I talked with John Cowan and sketched out my argument that,
> for
> consistency, IPC operations should be handled the same as File
> operations.
>
> Several IPC constructs, such as named pipes, are essentially
> files or
> could be implemented as such. Any Implementation that provides
> access to
> a system's filesystem should provide IPC support to remain
> consistent.
>
> Certain environments don't have filesystems, however, and
> Scheme should
> still target those environments. Obviously Scheme's
> requirements can not
> require support for something that doesn't exist in a targeted
> environment.
>
> John Cowan, informed me that a proposal had be submitted to
> drop file
> support from tiny Scheme and move it to WG2.
>
> This proposal makes sense. It would allow Scheme to be
> implemented in
> environments that don't have filesystems, while ensuring that
> a standard
> IPC library exists for those environments that do.
>
> Tiny scheme should drop support for file operations, and
> Extended scheme
> should include support for IPC and socket operations.
>
>
> On Fri, 2010-10-29 at 13:52 -0400, Thomas Bushnell, BSG wrote:
> > Not for working group one they aren't.
> >
> >
> > Not every scheme even HAS files or a filesystem around, let
> alone a
> > network. Part of the charter for WG1 is embedded devices.
> >
> >
> > And what about non-Unix systems? Are we going to mandate
> that if the
> > OS has no sockets or pipes, it isn't Scheme?
> >
> >
> > Thomas
> >
> > On Wed, Oct 27, 2010 at 11:48 AM, Larry Lee <llee@x>
> wrote:
> > Are IPC functions under consideration? Sockets,
> pipes, etc are
> > as integral as filesystem operations.
> >
> > "scheme-reports-request@x "
> > <scheme-reports-request@x> wrote:
> >
> > >Send Scheme-reports mailing list submissions to
> > > scheme-reports@x
> > >
> > >To subscribe or unsubscribe via the World Wide Web,
> visit
> > >
> >
> http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports
> > >
> > >or, via email, send a message with subject or body
> 'help' to
> > > scheme-reports-request@x
> > >
> > >You can reach the person managing the list at
> > > scheme-reports-owner@x
> > >
> > >When replying, please edit your Subject line so it
> is more
> > specific
> > >than "Re: Contents of Scheme-reports digest..."
> > >
> > >
> > >Today's Topics:
> > >
> > > 1. New features in WG1 Scheme (John Cowan)
> > >
> > >
> >
> >----------------------------------------------------------------------
> > >
> > >Message: 1
> > >Date: Sun, 17 Oct 2010 23:42:11 -0400
> > >From: John Cowan <cowan@x>
> > >Subject: [Scheme-reports] New features in WG1
> Scheme
> > >To: scheme-reports-wg1@x,
> > > scheme-reports@x
> > >Message-ID: <20101018034210.GG373@x>
> > >Content-Type: text/plain; charset=us-ascii
> > >
> > >Here's a non-authoritative summary of the currently
> approved
> > new features
> > >in R7RS small Scheme:
> > >
> > > Lexical syntax:
> > >
> > >Inexact values +inf.0, -inf.0, and +nan.0.
> > >
> > >Escaped symbols with |...|.
> > >
> > >Named character escapes (short and long names) in
> strings.
> > >
> > >Numeric character escapes in strings and symbols.
> > >
> > >#; S-expression comments and #|...|# block
> comments.
> > >
> > > Syntax:
> > >
> > >Internal DEFINE-SYNTAX.
> > >
> > >LETREC* is added, and internal DEFINE is made to
> use it.
> > >
> > >Tail patterns, ellipsis escaping, ellipsis
> redefinition, and
> > _ as wildcard
> > >in syntax-rules.
> > >
> > > Modified procedures:
> > >
> > >MAP and FOR-EACH are allowed to take lists of
> unequal length,
> > and stop
> > >when the shortest list runs out.
> > >
> > >ASSOC and MEMBER take an optional third argument,
> the
> > predicate to use.
> > >
> > > New procedures:
> > >
> > >random-integer, random-real, current-error-port,
> delete-file,
> > >file-exists?, make-list, copy-list, list-set!,
> string-map,
> > >string-for-each, string->vector, vector->string,
> copy-vector,
> > >vector-map, vector-for-each, exact-integer?.
> > exact-integer-sqrt,
> > >finite?, nan?, error, syntax-error,
> open-input-string,
> > >open-output-string, get-output-string,
> > read-with-shared-structure,
> > >write-with-shared-structure.
> > >
> > >--
> > >Values of beeta will give rise to dom!
> John Cowan
> > >(5th/6th edition 'mv' said this if you tried
> > http://www.ccil.org/~cowan
> > >to rename '.' or '..' entries; see
> > cowan@x
> > >http://cm.bell-labs.com/cm/cs/who/dmr/odd.html)
> > >
> > >
> > >
> > >------------------------------
> > >
> > >_______________________________________________
> > >Scheme-reports mailing list
> > >Scheme-reports@x
> >
> >http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports
> > >
> > >
> > >End of Scheme-reports Digest, Vol 10, Issue 1
> > >*********************************************
> > _______________________________________________
> > Scheme-reports mailing list
> > Scheme-reports@x
> >
> http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports
> >
> >
>
>
>
>
_______________________________________________
Scheme-reports mailing list
Scheme-reports@x
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports