Re: Terminology question re ASCII

From: Philippe Verdy <verdy_p_at_wanadoo.fr>
Date: Tue, 29 Oct 2013 18:59:22 +0100

2013/10/29 Shawn Steele <Shawn.Steele_at_microsoft.com>
>
> I would concur. When I hear “8 bit ASCII” the context is usually
confusing the term with any of what we call “ANSI Code Pages” in Windows.
 (or similar ideas on other systems).

Of course not just Windows (or MSDOS). This was seen as well in vrious
early OSes for personnal computers from various brands nd various countries
(not just US like Atari, but as well from Japan, France, Germany, UK,
Sweden and certainly others, where neither the US-only "ASCII" or "ANSI"
were standard). We've also seen these documents speaking bout "US-ASCII"
when they actually meant an 8-bit encoding whose lower 7-bit part matched
ISO 646 for US (i.e. the real "ASCII" standard from ANSI).

Due to Windows however (also in IBM OS/2, IBM DOS, and other derived OSes
by Digital Research for example, and also in some brands of Unix, CPM,
VMS... as well as in early development/porting for Linux), the ambiguity
rose when people started to speak about "ANSI" as an encoding when it
actully standard body developing various standards (including for other
encodings), and later this was corrected (not in Windows which uses the
incorrect terms "ANSI codepage" when none of them were actaully coming from
ANSI but from Microsoft, IBM, or some other national bodies, and later
modified by Microsoft !) by simply using "ASCII" instead of "ANSI", when
they should have just spoken of **some** range of 8-bit encodings supported
by the underlying OS whose lower 7-bit part was more or less based on some
national version of ISO 646 (or sometimes only in its "invariant" part,
excluding significant parts reserved for C0 controls but tweaked to encode
printable characters, e.g. for in VISCII or in IBM PC codepages for DOS).

7-bit and 8-bit encodings have always been a mess to reference, with
frequently ambiguous or wrong names, and many aliases being developed when
trying to disambiguate them (e.g. the IBM and Microsoft numeric codepages,
later aliased again on other systems !). This lead to the creation of an
international registry for encoding identifiers to fix the recommended
idenfiers for interchange and deprecate the other aliases (but Microsoft
never used it directly, it continued using its own numeric codepages, and
just accepted a few named aliases, sometimes incorrectly, for example when
Microsoft Frontpage confused and aliased ISO-8859-1 and windows-1252,
changing them in incomptible ways, forcing now HTML5 do declare that
"ISO-8859-1" is no longer this standard but "windows-1252").
Received on Tue Oct 29 2013 - 13:01:54 CDT

This archive was generated by hypermail 2.2.0 : Tue Oct 29 2013 - 13:01:56 CDT