Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: anton@mips.complang.tuwien.ac.at (Anton Ertl)
Newsgroups: comp.lang.forth
Subject: Re: QUIT and ABORT
Date: Sat, 24 May 2025 15:12:16 GMT
Organization: Institut fuer Computersprachen, Technische Universitaet Wien
Lines: 50
Message-ID: <2025May24.171216@mips.complang.tuwien.ac.at>
References: <87bjtn2hct.fsf@gmail.com> <1009s09$c9ih$1@dont-email.me> <4481f037e7bf4a8d7faa59640a9dcfac01c42226@i2pn2.org> <100b23g$g4a6$1@dont-email.me> <5ddd51973a3f163595605432f497ac65a5ed1336@i2pn2.org> <100c51f$sfr3$1@dont-email.me> <732e53d829317e715612f6941a69b62e7d38ac15@i2pn2.org> <100f2hm$1itg8$1@dont-email.me> <5b4885b76e5d683a9a55c51f6905d50aa86192ac@i2pn2.org> <100s31p$jh5t$1@dont-email.me> <956c844426013a9d8039bebd7f7c0a31@www.novabbs.com>
Injection-Date: Sat, 24 May 2025 17:30:46 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="bd3a0369fd21f4be5426cc9af645fe5f";
	logging-data="822135"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+28r+kskWnp/JpeXxHy2FY"
Cancel-Lock: sha1:/C6A8kkYlZPbaomsOSMlzQ2QVdg=
X-newsreader: xrn 10.11

mhx@iae.nl (mhx) writes:
>On Sat, 24 May 2025 9:20:57 +0000, Ruvim wrote:
>
>[..]
>> In a standard system, any behavior that is not explicitly allowed by the
>> standard (and can be detected by a standard program) is prohibited.
>
>This can't be falsified - there is no list which enumerates all possible
>behavior.

If you throw a standard program at the system and it behaves
differently from what the standard specifies, the system is not a
standard system.  This looks like a falsification of the claim "This
system is a standard system" to me.

There is a test suite (maintained by Gerry Jackson) that tests a
system for standards conformance, but as Dijkstra said, tests can only
prove the presence of bugs, not their absence.

>Also, the effect of e.g. a random memory (hardware) error is
>not
>something that can be prohibited.

It's probably pretty hard to implement a standard system on broken
hardware.  It's probably better to use working hardware instead.

In any case, if the system deviates from what the standard specifies,
it does not matter whether that's because of a bug in the system
software or breakage in the hardware.

>> In a
>> standard program, any behavior that is not explicitly prohibited by the
>> standard is allowed.
>
>This can easily be falsified with a extensive test suite.

I fail to see how.  Instead, Jonah E. Thomas once had the project of a
Forth system (written in Forth) that would do its utmost to only
execute standard programs and report deviations; but that would catch
only cases that happen in the particular execution of the program.
Quite a bit later the C and C++ world introduced "sanitizers" which
follow the same principle (and have the same shortcoming).

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: https://forth-standard.org/
EuroForth 2023 proceedings: http://www.euroforth.org/ef23/papers/
EuroForth 2024 proceedings: http://www.euroforth.org/ef24/papers/