[TUHS] Earliest Non-English UNIX?

[TUHS] Earliest Non-English UNIX?

[segaloco via TUHS]

I recently got a Japanese copy of the SVR3 Programmer's Reference Manual as a reference on learning technical translation between Japanese and English.  It got me wondering on the subject of Japanese UNIX again, and I recall some prior discussion about codepages and eventual UTF-8 use for this and other regional porting efforts, but UNIX use in non-English settings certainly predates UTF-8.  From what I know of the Japanese situation, at least AT&T, Sony, and Sun were prominent players in the Japanese UNIX market early on.

This has me curious more generally, are there good records on how UNIX became cosmopolitan, what non-English (and especially non-Latin) languages it was first translated into, and how involved, if at all, Bell Labs and UCB (as opposed to post divestiture ATTIS/USL) were in early efforts to localize UNIX outside of the anglophone world.

Possibly related, the typo(I) page from V3 makes a quip about disabling the "English" mode to instead target other languages such as Urdu.  Urdu is written in an Arabic-derived script rather than Latin.  Was this just a "hey this is possible" quip or was that to suggest there was experience already by V3 of playing with non-Latin scripts in UNIX?

- Matt G.

[TUHS] Re: Earliest Non-English UNIX?

[Alan Coopersmith via TUHS]

On 1/27/26 15:46, segaloco via TUHS wrote:
> but UNIX use in non-English settings certainly predates UTF-8

For SunOS/Solaris, before Unicode, it was primarily ISO-8859-* for the
European languages, and Extended Unix Code (EUC) for the CJK set.

See https://en.wikipedia.org/wiki/Extended_Unix_Code and the EUC section
of _Creating Worldwide Software: Solaris International Developer's Guide_
(available at https://archive.org/details/creatingworldwid0000tuth )
for details on EUC.

While we've deprecated those locales now in favor of the UTF-8 locales
in Solaris 11.0 and later, they're still around, and the character sets
supported in things like iconv to allow translating documents, and we
know there are still some customers using the EUC locales today.

-- 
         -Alan Coopersmith-                 alan.coopersmith@oracle.com
          Oracle Solaris Engineering - https://blogs.oracle.com/solaris

[TUHS] Re: Earliest Non-English UNIX?

[Edouard Klein via TUHS]

For reference the earlier thread was

https://www.tuhs.org/pipermail/tuhs/2023-March/028243.html
"[TUHS] Bell Foreign-Language UNIX Efforts"

in which I posted a history of France's Bull efforts to internationalize
UNIX in French, and also in Russian (!) for siberian gas fields:

https://www.tuhs.org/pipermail/tuhs/2023-March/028248.html

But that was quite a bit later (late 80s, early 90s) than what Matt is
asking here.

The whole thread is worth a read. Looking forward to reading this one as
well to see if more comes up.

Cheers,

Edouard.

segaloco via TUHS <tuhs@tuhs.org> writes:

> I recently got a Japanese copy of the SVR3 Programmer's Reference Manual as a
> reference on learning technical translation between Japanese and English. It got
> me wondering on the subject of Japanese UNIX again, and I recall some prior
> discussion about codepages and eventual UTF-8 use for this and other regional
> porting efforts, but UNIX use in non-English settings certainly predates UTF-8.
> From what I know of the Japanese situation, at least AT&T, Sony, and Sun were
> prominent players in the Japanese UNIX market early on.
>
> This has me curious more generally, are there good records on how UNIX became
> cosmopolitan, what non-English (and especially non-Latin) languages it was first
> translated into, and how involved, if at all, Bell Labs and UCB (as opposed to
> post divestiture ATTIS/USL) were in early efforts to localize UNIX outside of
> the anglophone world.
>
> Possibly related, the typo(I) page from V3 makes a quip about disabling the
> "English" mode to instead target other languages such as Urdu. Urdu is written
> in an Arabic-derived script rather than Latin. Was this just a "hey this is
> possible" quip or was that to suggest there was experience already by V3 of
> playing with non-Latin scripts in UNIX?
>
> - Matt G.

[TUHS] Re: Earliest Non-English UNIX?

[Martin via TUHS]

Hello,
I would like to mention here not a UNIX per-se but heavily inspired by 
it, polish Operating System called CROOK written entirely in 
K-202/MERA-400 assembly language during 1970s and 1980s in Gdansk 
University of Technology: https://mera400.pl/Galeria
CROOK design was dictated by the same universal needs (multitasking, 
multiuser support, a hierarchical file system, etc) that shaped UNIX. It 
was build to work with both older peripherals (like tape readers) and 
newer devices produced for the MERA-400 by companies like Amepol. 
Development even influenced the design of the MX-16 processor upgrade 
for the MERA-400.

A major digital archaeology project successfully recovered CROOK and 
other software from original magnetic tapes, i.e:
https://hackaday.com/2017/03/03/raiders-of-the-lost-os-reclaiming-a-piece-of-polish-it-history/#more-245582 

Team led by Jakub Filipowicz ( https://github.com/jakubfi ) preserved 
this key piece of Polish IT history and actively documenting now all the 
archaeology efforts related with hardware/software restoration of MERA400:
https://www.youtube.com/@MERA400
The recovered disk images of CROOK-5 are now available for use in the 
EM400 emulator, i.e.:
https://mera400.pl/CROOK-5_-_obraz_dysku

Some basic specs of the machine:
- 16-bit asynchronous CPU, up to 1 million operations per second
- 48-bit hardware floating point
- 121 instructions
- 1Mword (2MB) total address space
- memory segmentation and paging: dedicated memory block for OS + 15 
blocks for user programs
- 32 interrupts (maskable, 10 priority levels)


Kind Regards!
Martin