Conference 7.286::digital

    re .0: YOu'll probably soon have more pointers than you ever knew
    existed... ;-)
    
    Typically, our systems have been able to handle those for a long time,
    so if you stay within Notes/VAXmail(ALL-IN-1 etc. you shouldn't have
    much problems. It's mostly the UNIX world that has been lacking in this
    area... (I guess I'd better put my asbestos suit on now).
    
    P.S. When I moved here (Munich, Germany), DEC was able to provide video
    terminals, printers etc. with all the standard Umlauts. Siemens wasn't.
    

    Check the compose character key section of any recent VT manual.

    The technology and support for this has been around for about a decade
    in any OpenVMS based environment. 
    
    The interface on my spiffy new Windows 95 system stinks by comparison.
    (Look in help for subjects on the "Character Map".)

My keyboard (came with avanti running NT) doesn't have a compose
key, but I access notes from a VMS system and displaying back the
graphical interface to notes via eXcursion (trying to avoid 
PATHWORKS and DECnet).

As someone married to a person with such letters in his name, I've
given up hope that many people (US) grasp the difference.  When talking
to Americans, we drop the proper letters in favor of a's and o's.
When writing to family and registering paperwork (births/marriage/etc),
we use the proper letters.  Having someone desire to spell the name
properly is a mind boggling improvement.

    
    Keyboard without a compose key, can simulate the compose by
    pressing ALT-SPACE.  This is courtesy of the Unix folks from
    MIT who redefined what the key meant.
    
    Of course, this will only work under something that understands
    the X11 hack.  I could not even make a guess how to do it from
    NT or Windows 95.  You probably have to double click on something ;-)
    
    

    On standard Windows PCs, Alt + combinations of keypad numbers will 
    give special characters and characters with diacritical marks.  For 
    example, Alt + 130 = �
    	     Alt + 133 = �
    
    etc.
    
    M

    re .6: How very user friendly. Compose-a-" for � is of course totally
    incomprehensible, because it was invented 15 years ago?
    
    Best regards from Mr. J�rvinen (my username is utterly incorrect).

Digital systems were able to display diacritical marks more or less correctly
since when the VT220 came out.  A minor change was needed at some later point
to be compliant with the ISO-Latin-1 standard.  But as Mr. J�rvinen points out,
things like usernames or file names never caught up. 

PCs use some other standard for what the high-bit-set characters should be, so
there are two competing standards, PCs and most everyone else. You see this
often on Usenet, someone attempting to use umlauts or whatever in a word
appears as garbage in the other system. 

    From a PC, if you use eXcursion, DECterm (dxterm) recognizes
    ALT-space as Compose. Similarly, if you use VT320e (aka VTstar) for
    a Telnet emulator, ALT-space sort of works. It's order-sensitive,
    inconsistent, and incomplete (e.g. "a returns �, but a" just beeps;
    c, returns � but  ,c just beeps, and neither /u nor u/ returns �).
    
    KEA! (another Windows-based terminal emulator) will display an LK201
    window on which you can click Compose (I don't know if ALT-space
    works with KEA!).
    

    I got this Compose Key aide from Larry a few years ago, I don't think
    he'll mind sharing it now.
    
    -- Kenny House
    
    
from Larry Seiler (6 jun 86):
    +----+----+----+----+----+----+----+----+----+----+----+----+----+------+
    | ~  | !  | @  | #  | $  | %  | ^  | &  | *  | (  | )  | _  | +  |Delete|
    | `  | 1  | 2  | 3  | 4  | 5  | 6  | 7  | 8  | 9  | 0  | -  | =  |      |
    +----+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+------+
    | Tab  | Q  | W  | E  | R  | T  | Y  | U  | I  | O  | P  | {  | }  |      |
    |      | q  | w  | e  | r  | t  | y  | u  | i  | o  | p  | [  | ]  |Return|
 +--+-+----+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+    |
 |Ctrl| Caps | A  | S  | D  | F  | G  | H  | J  | K  | L  | :  | "  | |  |    |
 |    | Lock | a  | s  | d  | f  | g  | h  | j  | k  | l  | ;  | '  | \  |    |
 +----+---+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+-+--+----++---+
 | Shift  | >  | Z  | X  | C  | V  | B  | N  | M  | ,  | .  | ?  | Shift  |
 |        | <  | z  | x  | c  | v  | b  | n  | m  | ,  | .  | /  |        |
 +--------+----+-+--+----+----+----+----+----+----+----+----+----+--------+

  " "   # ++  ' '   @ aa  [ ((  \ //  ] ))  ^ ^   ' '   { (-  | /^  } )-  ~ ~ 
  � !!  � c/  � L-  � Y-  � SO  � OX  � CO  � a_  � <<  � +-  � ^2  � ^3  � /u
  � P!  � ^.  � ^1  � o_  � >>  � 14  � 12  � ??  � A`  � A'  � A^  � A~  � A"
  � A*  � AE  � C,  � E`  � E'  � E^  � E"  � I`  � I'  � I^  � I"  � N~  � O`
  � O'  � O^  � O~  � o"  � OE  � O/  � U`  � U'  � U^  � U"  � U"  � ss  � a`
  � a'  � a^  � a~  � a"  � a*  � ae  � c,  � e`  � e'  � e^  � e"  � i`  � i'
  � i^  � i"  � n~  � o`  � o'  � o^  � o~  � o"  � oe  � o/  � u`  � u'  � u^
  � u"  � u"

    Under Windoze, it depends...
    
    In Word95, you can type control+' then a to get �.  There's a bunch of
    similar things -- see Insert Symbol, then seleft the (normal text)
    font, click on the character you want and look at the shortcut key.  If
    desired, you can change the shortcut key to be more to your liking.  If
    the ALT+0165 stuff is all that is currently defined as a shortcut key,
    it will at least show it to you (ALT+0165 is �) so you don't *have* to
    carry around a cheat sheet.
    
    In other applications, ALT+4 digits is all you get.  Note that the 4
    digits myst be typed ON THE KEYPAD, and not using the keys at the top
    of your keyboard.  Also note that any leading 0s are MANDATORY where
    indicated.
    
    This might be better in Windoze95 / Windoze NT 4.0, but I doubt it.
    
    It's not even consistent within Office95 (Word OK, difficult for the
    rest of the suite). 
    
    Steveg

	On Windows95 and Windows NT, go to the Control Panel, select
	Keyboard, and for the US, select United States-International.

	Now for letters like ����� and the like, you type things like
	'e 'a `e `a "o ^e  Note that it doesn't work in my terminal
	emulator because it takes over the keyboard with its own
	mappings. Also, hitting characters like " will cause the
	output to pause while it waits to see if you will hit another
	letter, like o. Hitting " then SPACE gave me " right away. It's
	taken a few days to get used to it, but it's worth in in being
	able to spell names properly.

						mike a.k.a. mich�al
						(pronounced "Me-haul")

RE .12 - Cool! Thanks for the tip.

-- Ni�a :-)

    Cute. Doesn't seem to offer �, though, unless there's some
    combination I missed. (Not that I ever use it; it's the principal of
    the thing...)

    � is just compose c ,
    
    at least on this keyboard alt space then the combination works too.
    
    Dave�
    

    Nobody mentioned the 7-bit national character sets. When my niece sends
    me e-mail from Germany she creates Umlauts on a keyboard with 7-bit
    codes. The "|" is labeled "�". When I receive it I see the "|". 
    Other countries have other standard national 7-bit codes. 
    
    I've seen even some e-mail transports trying to fix that on the fly
    sometimes producing largely encrypted text.
    
    Guenther

    re .16: yes. re .8: Even the VT100 and VT52 were able to use NRC's.
    You'd have to get a kit from CSS with a few new keycaps and character
    ROM. A real joy for C programmers to write something like
    
    foo()
    �
    	char bar�10�;
        ......    
    �

    Hey, you want to talk 7-bit?  How about 6-bit?
    
    My first project in Digital was doing the initial design work on French
    and Danish versions of WPS-8.  I presented the result to the product
    line manager, who reamed me out in front of 8 other suits (me a fresh
    new hire trembling at my overheads).  The product line manager found it
    absolutely incomprehensible that it might pose some difficulties to
    expand the alphabet -- this on a machine that wanted to store 
    characters two to a (12-bit) word.
    
    I got out of that project as fast as I could, heading for the oasis of
    the 16-bit PDP 11.   But they did in fact proceed to build it, and it
    was delivered in something like 1980.

I think I'm missing something. I don't have in my W95 PC a United States 
(International) keyboard. there are several English (country) but no 
International.

The way I deal with this matters is to have two keyboards habilitated 
(English (United States) and Spanish (Mexican), and switch between keyboards 
with a "left Alt+Shift" combination, now I can type all spanish characters. 
The only problem is that unless you have a diagram of the keyboard layout or 
know the key positions by heart, is a little of trial and error, since now 
many signs changed keys assignments.

/OLD

    re .19: I don't have my W95 handy, but in NT 4.0 it's control panel ->
    keyboard -> input locales, then select one of them (e.g. "English
    (United States)" and click properties. You should get a dialog for
    keyboard layout and find "US International" there.

Note that one irritation is that double and single quotes lead into
sequences, so you don't get access to the international characters
with no changes in standard keys (unfortunately).

    
    Re: all.
    
    While it is true that we have been able to spell our native languages
    correctly when composing TEXT documents, many things in VMS and UNIX
    are broken when seen from a local perspective.
    
    For instance, my username cannot be Webj�rn, a file cannot be named
    
    �vers�ttning_fr�n_utl�ndskt_spr�k.txt =
    
    (translation_from_foreign_language.txt)
    
    What's worse, is that while M-Soft operating systems and products allow
    filenames with the full ISO latin character sets onto shares on our
    legacy servers, those names when displayed on the native system
    translates to gibberish. (try it on pathworks, you'll know what I mean)
    
    The perception is naturally that our 'legacy' UNIX and VMS systems
    appear even older and more dated when compared to the M-Soft stuff.
    
    
    The world wide web has introduced umlauts en masse, since much
    information is graphic, and noone has yet invented a scheme to filter
    away the dots and asterisks from pictures.
    
    I really don't want to restart the lobbying I did in the mid 80's,
    suffice to say that developers interested in international markets 
    pay attention to appearance and perception.....
    
    Gullik Webj�rn
    
    

    The Web may have introduced many umlauts, but I still can't have
    Ora.J�[email protected] or some such as my mail address.

    
    VMS will be fixing a lot of the problems with file naming between
    PCs/NT and VMS with HFS-1.
    
    

Well, with eXcursion's decterm ALT and SPACE doesn't give me a
compose character.  But I can switch in the eXcursion control panel to
the Finnish keyboard to get � and � (and � should I want it).

e.g. my daughter's name is Johanna Maarit M�nnist� 

but the non character keys are all different, switching back
and forth is not friendly, maybe a later version of eXcursion
would help.

	Re: .22

>       What's worse, is that while M-Soft operating systems and products allow
>   filenames with the full ISO latin character sets onto shares on our
>   legacy servers, those names when displayed on the native system
>   translates to gibberish. (try it on pathworks, you'll know what I mean)

	Hi, Gullik.  This will be fixed in the next release of
	Pathworks for Digital UNIX.  I am the engineer working on the
	code.

	Maybe you can help me.  I am trying to decide what MS-DOS
	character sets to support.  I would like to cover all of them,
	but some folks think that code pages 437 (Latin US) and 850
	(Latin 1) should be sufficient.  Do you have any customer or
	personal experience that can help the decision?  For example,
	do you know if code page 865 (Nordic) is widely used in
	Sweden?

	Thanks,

	Tom Woodburn

|    The Web may have introduced many umlauts, but I still can't have
|    Ora.J�[email protected] or some such as my mail address.

	If I am not mistaken, you can have quoted printable encoding
	of your address (if you use MIME) along the lines of:

		Ora J=E2rvinen <[email protected]>

	My father's name comes into my mail system as a coded 8bit
	string and then appears in my inbox with 8bit text.

	>Per

|		Ora J=E2rvinen <[email protected]>

   Just checked. If your name appeared in the from field it
   would look something like this:

	From: =?ISO-8859-1?Q?Ora J=E4rvinen_?=<[email protected]>

   >Per

    
    For fairness, let us mime encode ALL wowels in all languages, ...
    
    but then we get infinite recursion, don't we??? ( =Ax and =Ey )
    
    Gullik
    

     >>Ora.J�[email protected]
    
    Jetzt wollen die noch das wir uns mit den bloeden Umlauten
    auseinandersetzen. Die Sprache ist schon schwer genug, Danke.
    
    Wir haben uns an
    
    BHAJEE::JAERVINEN "Ora, the Old Rural Amateur"
    
    gewoehnt. You kannst es auch, okay ?
    
    ..Kevin..
    
    

    Well, I consider my name as part of my personality... (some of my
    friends might even insist it's all the personality I have ;-) but I
    have to put up with the missing Unlauts - even here in Germany, most
    bills I receive have "Jaervinen" on them even though it's not my name.
    Maybe I should deny payment... ;-)
    
    Note: in Finnish, you _can't_ replace the umlauts with 'ae' etc. - the
    �'s, �'s, �'s etc. are letters in their own. It's just because I live
    in Germany my username happens to be 'JAERVINEN' - lacking Umlauts, a
    Finnish system mangler would just drop the dots and use 'JARVINEN' as
    the username. (And .25's daughter's account would be "MANNISTO", not
    "MAENNISTOE").
    

    Yes I know the problem
    
    Kevin V. Mannerings .NES. Kevin VON Mannerings
    
    Stadtwerke Frankfurt, I think the computer billing system  did it
    automatically.
    
    ..Kevin..

    �Stadtwerke Frankfurt, I think the computer billing system  did it
    �automatically.
    
    Yes, your excellency... ;-)
    
    Getting a bit off topic (nothing unusual in this notesfile...), I
    recently received a letter from a S/W vendor addressed as
    
    Ora Jarvinen
    ...street...
    Solomon, AZ 85551
    Germany
    
    85551 happens to be the Postleitzahl (zip code) for Kirchheim near
    Munich where I live... their computer apparently was just fed with the
    zip code, and the city was added automagically.
    

There are two problems being mixed up here.
Getting characters with diacritical marks that don't existing atomically 
on keyboards is a mechanical problem, one that can vary from system
to system and possibly even among programs on a system.

Not being allowed to use those characters in things like user names
is a largely separate problem.
User name restrictions evolve from the linking of these name with the
file structure of the operating system.
My account name on VMS is POWERS because that allows my home directory
to be called [POWERS] and pointed to in a file called 000000.POWERS.
Because of operating system restrictions, you can't use the character
code for ']' in a user name, it would complicate the OS's token parsing.
But in certain national use character sets, the numeric code for ']' 
is an attributed vowel or consonant.

Character codes in computers are fundamentally numeric (with some data
structure overlays for multibyte character sets and the like).
There has been great progress in unifying the numeric codes for character
sets during the past 15 years.  We are probably at the point where
a common choice of character encoding could allow definite delineation
of alphabetic characters that would allow OS token recognition to reserve
special characters for delimiters (punctuation) and still allow
all langauges that shared the same common superset alphabet to use
their entire character repertoire.

Of course, some advanced systems like Unix and the MacOS are already
well ahead of the crowd by not limiting the character set for filenames,
thereby loosening the restrictions on what people's names can be
on thise computers.

- tom]

    
    From a strictly Swedish point of view, a new age begun when MicroSoft
    quietly introduced ISO Latin-1 as the default encoding in windows.
    
    By magic text documents were interchangeable with preserved content.
    
    The 'naive' users never noticed this milestone, the perception was
    that it always worked properly. We had been using latin-1 for years
    on our legacy systems.
    
    Now, some years later, national characters (encoded as latin-1) are
    everywhere in the served PC environment. Progress on the more slowly
    evolving platforms have been abysmal...
    
    The big difference between the traditional computer companies (like us)
    and MicroSoft is that they see $$$ in not annoying the customer, and
    treating him/her polite.
    
    I know internationalization is a difficult issue if you consider all
    languages, encodings, rules and heritage. It takes a business desicion
    to decide if this is indeed important, and if you have decided it is
    then whatever needs to be done should be, disregarding techie
    objections.
    
    On the other hand, if you decide not to solve language problems,
    don't expect an international community to love you...
    
    It's more than 10 years since I heard the CEO of Televerket
    complaining not beeing able to get mail correctly adressed to him...
    
    Credit to MicroSoft who understand that is the end user is the
    customer.....
    
    Gullik
    

    re: .22
    
    $ touch �vers�ttning_fr�n_utl�ndskt_spr�k.txt
    $ ls -la �vers�ttning_fr�n_utl�ndskt_spr�k.txt
    -rw-------   1 smith    users          0 Jan 31 12:08 �vers�ttning_fr�n_utl�ndskt_spr�k.txt
    
    Was a part of UNIX long before Microsoft came along, if I'm not
    mistaken.
    
    -Tom

> $ touch �vers�ttning_fr�n_utl�ndskt_spr�k.txt
> $ ls -la �vers�ttning_fr�n_utl�ndskt_spr�k.txt
> -rw-------   1 smith    users          0 Jan 31 12:08 �vers�ttning_fr�n_utl�ndskt_spr�k.txt
>     
> Was a part of UNIX long before Microsoft came along, if I'm not mistaken.

	Yup, this is quite correct.  The only thing not allowed in a
	UNIX file name is '/' because that's used as a seperator, and
	'\0' (null character) because UNIX uses it to terminate a string.

|    For fairness, let us mime encode ALL wowels in all languages, ...
|    
|    but then we get infinite recursion, don't we??? ( =Ax and =Ey )

	Not so. The '=' sign is a quote character and the two characters
	immediately following are part of the quote expression...

	>Per

    Could you imagine FORTRAN developed in Germany? They would have called
    it F�RTR�N, yuk!
    
    G{�|u|ue}nther

Re .34:

> Of course, some advanced systems like Unix and the MacOS are already
> well ahead of the crowd by not limiting the character set for filenames,
> thereby loosening the restrictions on what people's names can be
> on thise computers.

On the contrary, filenames on these "advanced" systems are limited to 8 bit 
characters (minus some for overhead), therefore < 256 characters. Try mixing 
some Cyrillic/Roman/Greek/etc characters in that filename.

Some advanced systems like Windows NT allow 16 bit Unicode characters (minus 
some for overhead), therefore < 65536 characters. (I just created a user 
called J�rvinen in our Windows NT domain.)

Some computer languages (Java, for instance) also explicitly allow for Unicode, 
thus allowing the Greek letter pi (*not* the two Roman letters "pi") as a 
perfectly good variable name, for instance.

(Totally off the thread, I once thought how amusing it would be if the Cyrillic  
alphabet was named after a guy called Cyril. I went and looked it up, and sure 
enough, it was invented by a St Cyril back in the 12th century or thereabouts. A 
head full of useless trivia can keep one amused for hours.)

PJDM

    
    Re: .40
    
    Exactly....
    
    Only a company understanding the importance of language will make a
    desicion to waste large amount of storage space, to increase the
    allowable character set range to fit most (all?) languages.
    
    This is a tradeoff. A desicion has been made that it IS important 
    enough. There has not been a debate about how to force other languages
    to fit in. Fitting computers to people. Not fitting people to
    computers. If you still are not convinced, talk to a 'naive' user.
    
    People 'driving' computers without understanding them are in MAJORITY!
    
    Gullik
    

. Some computer languages (Java, for instance) also explicitly allow for
. Unicode, thus allowing the Greek letter pi (*not* the two Roman letters
. "pi") as a perfectly good variable name, for instance.

Yet another example of Scott McNeally's arrogance! Will this megalomania
never end? What other universal constants will be declared to be variables?
Talk about introducing change for its own sake ...

PK

    re .40:
    
>(I just created a user 
>called J�rvinen in our Windows NT domain.)
    
    Great, what's my passowrd then so I can actually log in?
    
    ;-)
    

    Hi:
    .0 - Checking back in...
    (.1 - Told me I'd receive a lot of replies, but I had no idea...)
    
    				So...
    
    		Th��k y� �ll f�r th� �d���t���!!! 
    
                Now about this metric system...;-)
    
    Regards,
    bill

    >(I just created a user
    >called J�rvinen in our Windows NT domain.)
    
    >>    Great, what's my passowrd then so I can actually log in?
    
    Try  NUISANCE
    
    the little green man in there will know at once :-)

    The terminal editor I use does not let me use the alt space to
    compose characters. I want to try and assign one of the function
    keys to pass the alt and space characters. Does anyone know what
    the character code is for the alt key?. I was going to try and
    assign ???||040 to a function key to see if I could get my emulator
    to pass the sequence to VMS.