[Inform] mistype.h

8 views
Skip to first unread message

Cedric Knight

unread,
May 2, 2003, 7:44:27 PM5/2/03
to
mistype.h is an Inform library to automatically correct single-letter
typing errors, in a similar way to nitfol. Now you can type >KILL HTE
DRAGON to your heart's content.

It's available from

http://www.metebelis3.free-online.co.uk/mistype.h

I've tested it on Zcode & Glulx interpreters, but before uploading it to
the archive I would like some feedback, especially about whether it
should be more context-sensitive, pseudointelligent, or speed-optimized.

And BTW: a slightly belated happy 10th birthday to Inform....

CK


Roger Firth

unread,
May 7, 2003, 8:52:57 AM5/7/03
to
"Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org> wrote in message
news:NrDsa.13630$xd5.7...@stones.force9.net...

> mistype.h is an Inform library to automatically correct single-letter
> typing errors, in a similar way to nitfol. Now you can type >KILL HTE
> DRAGON to your heart's content.
>
> It's available from
>
> http://www.metebelis3.free-online.co.uk/mistype.h
>
> I've tested it on Zcode & Glulx interpreters, but before uploading it to
> the archive I would like some feedback, especially about whether it
> should be more context-sensitive, pseudointelligent, or speed-optimized.

It seems to work excellently. My only suggestion is that you
skip the 'deletions' tests for two-letter words: LOOK AR ROCK
is changed to LOOK A ROCK, rather than the more obvious
LOOK AT ROCK. Apart from that, just great.

Many thanks, Roger
--
/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
You'll find my Cloak of Darkness, Parsifal, Informary
and more at http://www.firthworks.com/roger/

Cedric Knight

unread,
May 8, 2003, 7:01:21 AM5/8/03
to
Roger Firth wrote:

>> mistype.h is an Inform library to automatically correct single-letter

>> typing errors, in a similar way to nitfol. [...]
>> http://www.metebelis3.free-online.co.uk/mistype.h


>
> It seems to work excellently. My only suggestion is that you
> skip the 'deletions' tests for two-letter words: LOOK AR ROCK
> is changed to LOOK A ROCK, rather than the more obvious
> LOOK AT ROCK. Apart from that, just great.

Thanks. I think you're right about the two-letter word problem & have
updated the file, but this now results in

LOOK AN ROCK

which is even less grammatically correct :)

I'll have to make the next version sensitive to grammar and keyboard
layout.

CK

Roger Firth

unread,
May 9, 2003, 1:15:00 AM5/9/03
to
"Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org> wrote in message
news:WSqua.16103$9C6.8...@wards.force9.net...

Rather than making substitutions ABC...Z, you could try in order
of English frequency ETAOINSHRDLU... -- it might help a little?

Cheers, Roger

Cedric Knight

unread,
May 12, 2003, 4:10:36 PM5/12/03
to
Roger Firth wrote:
> "Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org>

>> Roger Firth wrote:
>>
>>>> mistype.h is an Inform library to automatically correct
>>>> single-letter typing errors, in a similar way to nitfol.
>>>
>>> It seems to work excellently. My only suggestion is that you
>>> skip the 'deletions' tests for two-letter words: LOOK AR ROCK
>>> is changed to LOOK A ROCK, rather than the more obvious
>>> LOOK AT ROCK. Apart from that, just great.
>>
>> Thanks. I think you're right about the two-letter word problem &
>> have updated the file, but this now results in
>>
>> LOOK AN ROCK
>>
>> which is even less grammatically correct :)
>>
>> I'll have to make the next version sensitive to grammar and keyboard
>> layout.
>
> Rather than making substitutions ABC...Z, you could try in order
> of English frequency ETAOINSHRDLU... -- it might help a little?

IMHO this wouldn't be so useful in the general case. While it does
increase the likelihood of finding a word earlier in the search, it
doesn't necessarily increase the probability that the word found is the
appropriate one.

As a matter of interest, was the above ('ar') a genuine typo (or a
pirate impression)? Many substitutions are with keys nearby on the
keyboard, so ar->at looks more likely than ar->an.

So... I have added (a) keyboard sensitivity; (b) context sensitivity.
Also (c) a facility to 'likl eth gronad' (kill the dragon) which someone
not normally known for dyslexia requested. The result is that the
library has tripled in size (about 2K object code, 23K source), and gone
back to beta status.

Just in case anyone's interested it now works like this:

If you are in the same room as Bob and a box,
'x bov' is read as 'x bob', while
'x boc' is read as 'x box', due to keyboard proximity.

If Bob leaves the room, 'x bov' or 'x bor' is read as 'x box', but 'x
bob' is still not spellchecked; hence this behaves a bit like TADS in
that you can work out what words will probably occur later in the story.

The grammar is recognised in such a way that verbs substitutions are
always preferred to nouns as the first word on a line (or after a
comma). Prepositions are inferred according to declared grammar, so

'look ag rock' is read as 'look at rock', whereas
'take ag rock' is read as 'take a rock'.

((Although at the moment 'take ar rock' still seems to give 'take
at rock' becuase I decided to ignore valid determiners here - nothing's
perfect))

I'm cautiously confident the parser will now correctly replace about 90%
of errors.

Of course, the 'LOOK A ROCK' problem could also be solved by explicitly
allowing 'look <noun>' as Mike Arnautov seemed to be suggesting on rgif.
While I agree with Mike Roberts that a more interesting and widely
applicable world model has necessitated a more complex parser (and not
the other way around), I think that making the parser more tolerant
could be helpful.

CK

Steve Evans

unread,
May 12, 2003, 11:17:00 PM5/12/03
to
"Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org> wrote in message news:<XfTva.21950$xd5.9...@stones.force9.net>...


> So... I have added (a) keyboard sensitivity; (b) context sensitivity.
> Also (c) a facility to 'likl eth gronad' (kill the dragon) which someone
> not normally known for dyslexia requested. The result is that the
> library has tripled in size (about 2K object code, 23K source), and gone
> back to beta status.

Cedric, this is an excellent library.

When I was testing last year's comp game I manually added synonyms to
the game file to cover my common dyslexic typing errors in case others
playing the game were similarly afflicted.

This helped, but not when I was seriously hung-over, or some such, and
even my normal dyslexic responses had dyslexia.

One example in my game was the noun "ankh" which I would almost
invariably get wrong.

But to now find with mistype.h that I can type 'X ANKH' as 'X SNGKMH'
and still have the story understand what I'm trying to say, is
positively spooky.

Cheers,

Steve

Cedric Knight

unread,
May 13, 2003, 7:46:15 AM5/13/03
to
Steve Evans wrote:
> When I was testing last year's comp game I manually added synonyms to
> the game file to cover my common dyslexic typing errors in case others
> playing the game were similarly afflicted.

[snip]

> But to now find with mistype.h that I can type 'X ANKH' as 'X SNGKMH'
> and still have the story understand what I'm trying to say, is
> positively spooky.

Full telepathic abilities are on my 'to do' list :)

I'm relieved my comment about dyslexia wasn't misunderstood; probably
more IFfers are dyslexic than one would guess. Me, I'm just a rubbish
typist. I also didn't fully realise the full symbolism of the Ankh in
'Photograph' until just now - clever (I think).

CK

Fredrik Ramsberg

unread,
May 22, 2003, 6:02:23 AM5/22/03
to
"Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org> wrote in message news:<NrDsa.13630$xd5.7...@stones.force9.net>...

> mistype.h is an Inform library to automatically correct single-letter
> typing errors, in a similar way to nitfol. Now you can type >KILL HTE
> DRAGON to your heart's content.

Great work!

I'd like to request an extension to this module: For games in languages
that need exotic characters, like Swedish that I'm working with, there's
always the question whether you should write the game transcribing those
characters to normal characters so that everyone can play the game, or
trust the interpreter writers to add support for international characters
and trust the users to have the right keyboard to make it convenient.
This could be the solution we're all waiting for! If it could be made to
recognize certain character combinations as likely or even positive
misspellings for international characters, then the game author could
pretty much forget about this issue. There'd still be the issue of
interpreters not _displaying_ international characters correctly, but
at least for Latin-1 characters, that's a lot less common a problem.

So, for Swedish:

Level 1
"aa" is almost certainly a misspelling of "å"
"ae" is almost certainly a misspelling of "ä"
"oe" is almost certainly a misspelling of "ö"

Level 2
"a" is likely to be a misspelling of "å" or "ä"
"o" is likely to be a misspelling of "ö"

Both the player and the programmer should preferrably be able to turn this
particular character transcribation option on or off at will, since it
creates new words that may interfere with existing dictionary words.
Just like the mistype.h file works now, the player should be able to
turn off the normal mistype option entirely too, to preserve CPU speed.

Is this crazy?

Do you want to add this?

If you don't want to add it, would you mind if I gave it a go, based
on your source code?

/Fredrik

Cedric Knight

unread,
May 22, 2003, 2:23:32 PM5/22/03
to
Fredrik Ramsberg wrote:
> So, for Swedish:
>
> Level 1
> "aa" is almost certainly a misspelling of "å"
> "ae" is almost certainly a misspelling of "ä"
> "oe" is almost certainly a misspelling of "ö"

Wot no ø? Oh, that's Norwegian.

>
> Level 2
> "a" is likely to be a misspelling of "å" or "ä"
> "o" is likely to be a misspelling of "ö"
>
> Both the player and the programmer should preferrably be able to turn
> this
> particular character transcribation option on or off at will, since it
> creates new words that may interfere with existing dictionary words.
> Just like the mistype.h file works now, the player should be able to
> turn off the normal mistype option entirely too, to preserve CPU
> speed.

Oh, is it that slow? It shouldn't be.

>
> Is this crazy?
>
> Do you want to add this?

Well here's a first attempt for Swedish, with minimal changes and based
on your spec:
http://www.metebelis3.free-online.co.uk/svmistype.h

One potential problem is that most of the code is assuming *single*
errors, so wouldn't help for words with multiple accents. I see two
possibilities:
1) Copy the non-dictionary word to a different, temporary buffer, making
your Level 1 changes throughout to see if tokenise matches it to a
dictionary word.
2) Just rely on what I've labelled as 'last resort' (word distance)
checking, but remove all diacritics from each dictionary word as it is
considered.

In fact even with mistype.h as it stands, without special provision for
Swedish, 'undersok' will be corrected to 'undersök' because of this last
resort checking.

>
> If you don't want to add it, would you mind if I gave it a go, based
> on your source code?

If any one wants to use part or all of the source for the library
extension, please do - it's public domain and not copyright.

I'm not yet sure whether to try a generalised international version -
different languages have different accented letters with different
values and different keyboard layouts. (I'd use approach 2 above if I
did.)

HTH

CK

Cedric Knight

unread,
May 23, 2003, 10:45:09 AM5/23/03
to
Cedric Knight wrote:
> Fredrik Ramsberg wrote:
>> So, for Swedish:
>>
>> Level 1
>> "aa" is almost certainly a misspelling of "å"
>> "ae" is almost certainly a misspelling of "ä"
>> "oe" is almost certainly a misspelling of "ö"

Does anyone have a complete international list of these digraphs? There
are the ones above, plus ue for u-umlaut and ss for 'beta' in German; in
Spanish 'ny' can be used, and I believe the stress marks are ignored.
Portuguese? French? Norwegian/Danish?

>> If you don't want to add it, would you mind if I gave it a go, based
>> on your source code?
>
> If any one wants to use part or all of the source for the library
> extension, please do - it's public domain and not copyright.

I meant to say that you may want to tidy up the source if you do. I
indent braces strangely, which belies Pascal experience.

> I'm not yet sure whether to try a generalised international version -
> different languages have different accented letters with different
> values and different keyboard layouts.

These features will be options in the next release - a potential problem
in the current release with creme brulee is fixed.

CK


dhan

unread,
May 23, 2003, 12:58:16 PM5/23/03
to
Hi...

> Does anyone have a complete international list of these digraphs? There
> are the ones above, plus ue for u-umlaut and ss for 'beta' in German; in
> Spanish 'ny' can be used, and I believe the stress marks are ignored.
> Portuguese? French? Norwegian/Danish?

In Spanish "ny" doesn´t replace "n-tilde", only in Catalan, a
north-eastern community with his own language. By the way, what is a
stress mark? I don´t know if I ignore them :-)

dhan

Andrew Engels Rump (formerly Leif Andrew Rump)

unread,
May 27, 2003, 6:47:03 AM5/27/03
to
After drinking 3 Pan Galactic Gargle Blasters, "Cedric Knight"
<ckn...@gn.babpbc.removeallBstosend.org> mumbled in news:ZDqza.23155
$9C6.1...@wards.force9.net:

> Cedric Knight wrote:
>> Fredrik Ramsberg wrote:
>>> So, for Swedish:
>>> Level 1
>>> "aa" is almost certainly a misspelling of "å"
>>> "ae" is almost certainly a misspelling of "ä" "oe" is almost
>>> certainly a misspelling of "ö"
> Does anyone have a complete international list of these digraphs?

Hm, no, but I may add the following diagraphs for Danish:
"aa" is usually a misspelling of "å"
"ae" is usually a misspelling of "æ"
"oe" is usually a misspelling of "ø"

Andrew
--
*** The opinions expressed are not necessarily those of my employer. ***
* Software Engineer Andrew Engels Rump * BLIK og ROERarbejderforbundet *
* Immerkaer 42, 2650 Hvidovre * Tlf: +45 3638 3638, Fax: +45 3638 3639 *
Home: N55°41'38.9" E12°29'08.6" (WGS 84) Work: N55°39'50.9" E12°27'47.4"
E-mail: mailto:newa...@rump.dk WWW http://www.rump.dk/homepage/andrew/

John Colagioia

unread,
May 27, 2003, 1:10:12 PM5/27/03
to
"Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org> wrote in
news:JG8za.22940$9C6.1...@wards.force9.net: [...]

> I'm not yet sure whether to try a generalised international version
> - different languages have different accented letters with different
> values and different keyboard layouts. (I'd use approach 2 above if
> I did.)

Well, it seems to me that you don't want a completely generalized
version, since different languages will have different requirements.
However, it might be worth making it configurable by language, with
different keyboard layouts supplied (I'd offer to help on that one,
but unfortunately, I've long since left the job where we had the piles
of foreign keyboards in the basement).

And I'd probably go with your second approach, too. It's less
computationally intensive, and probably not much less accurate,
overall.

Cedric Knight

unread,
May 27, 2003, 6:31:07 PM5/27/03
to
dhan wrote:
> Hi...
>
>> Does anyone have a complete international list of these digraphs?
>> There are the ones above, plus ue for u-umlaut and ss for 'beta' in
>> German; in Spanish 'ny' can be used, and I believe the stress marks
>> are ignored. Portuguese? French? Norwegian/Danish?
>
> In Spanish "ny" doesn´t replace "n-tilde", only in Catalan, a

Oh, I thought I had seen it somewhere, maybe I was mistaken. But if you
didn't have an n-tilde on the keyboard, how would you show it?

> north-eastern community with his own language. By the way, what is a
> stress mark? I don´t know if I ignore them :-)

I meant the acute accent over a,e,i,o, or u indicating a longer or
heavier syllable, as in Per{'u}. I don't know what the proper name for
them is in English or Spanish.

Thanks for the responses from John & Andrew too.

CK

Daniel Dawson

unread,
May 28, 2003, 2:26:09 AM5/28/03
to
You pick up and read article <7XRAa.24196$9C6.1...@wards.force9.net>, written by
Cedric Knight <ckn...@gn.babpbc.removeallBstosend.org>. It says:
>dhan wrote:
[Snip]
[Forgot who wrote this originally...]

>>> There are the ones above, plus ue for u-umlaut and ss for 'beta' in
>>> German;

That's not a 'Beta', it's an 'Ess-Zett', a ligature of the letters 's' and 'z'
(in their archaic forms, in which the 's' looks like an 'f' without the
horizontal stroke, and the 'z' resembles an old-fashioned '3'). It does look
surprisingly like the Greek letter Beta, though; does anyone know if there is a
connection, or is it just a freak coincidence.

>I meant the acute accent over a,e,i,o, or u indicating a longer or
>heavier syllable, as in Per{'u}. I don't know what the proper name for
>them is in English or Spanish.

We (English-speakers) just call it an acute accent; the opposite accent (`) is
called a grave accent.

--
Email: Daniel Dawson <ddawson at icehouse.net>
(Please include my name (as above, but change ' at ' to '@') in To: field
when emailing me.)
Web: http://www.icehouse.net/ddawson/

Fredrik Ramsberg

unread,
May 28, 2003, 2:45:23 AM5/28/03
to
"Cedric Knight" <ckn...@gn.babpbc.removeallBstosend.org> wrote in message news:<JG8za.22940$9C6.1...@wards.force9.net>...

> Fredrik Ramsberg wrote:
> > So, for Swedish:
> >
> > Level 1
> > "aa" is almost certainly a misspelling of "å"
> > "ae" is almost certainly a misspelling of "ä"
> > "oe" is almost certainly a misspelling of "ö"
>
> Wot no ø? Oh, that's Norwegian.

Norwegian and Danish. It corresponds to our ö.

> > Level 2
> > "a" is likely to be a misspelling of "å" or "ä"
> > "o" is likely to be a misspelling of "ö"
> >
> > Both the player and the programmer should preferrably be able to turn
> > this
> > particular character transcribation option on or off at will, since it
> > creates new words that may interfere with existing dictionary words.
> > Just like the mistype.h file works now, the player should be able to
> > turn off the normal mistype option entirely too, to preserve CPU
> > speed.
>
> Oh, is it that slow? It shouldn't be.

I don't know how slow it is, so this is no accusation. I just
thought it might be a good idea unless it's extremely fast. There are
still some pretty slow platforms out there.

> > Do you want to add this?
>
> Well here's a first attempt for Swedish, with minimal changes and based
> on your spec:
> http://www.metebelis3.free-online.co.uk/svmistype.h

Why thanks! It works pretty well. I don't quite agree with some of
the more outrageous corrections it makes, but I could probably tune
it to suit me better.

> One potential problem is that most of the code is assuming *single*
> errors, so wouldn't help for words with multiple accents. I see two
> possibilities:
> 1) Copy the non-dictionary word to a different, temporary buffer, making
> your Level 1 changes throughout to see if tokenise matches it to a
> dictionary word.
> 2) Just rely on what I've labelled as 'last resort' (word distance)
> checking, but remove all diacritics from each dictionary word as it is
> considered.

I've implemented method 3 instead: The user has commands to
turn "transcription" (the best term I could think of) on and off.
When on, all occurances of any of the three special character
combinations are replaced by the corresponding accented
characters, unless preceded by the escape character "!". I do
this in LanguageToInformese. The coding was rather
straightforward and it gives the player full control (except
he/she can't use "!" for anything but escaping while in
transcription mode, but I don't think that's a serious problem).

> > If you don't want to add it, would you mind if I gave it a go, based
> > on your source code?
>
> If any one wants to use part or all of the source for the library
> extension, please do - it's public domain and not copyright.

I used your code for inspiration, which should be very ok then.

> I'm not yet sure whether to try a generalised international version -
> different languages have different accented letters with different
> values and different keyboard layouts. (I'd use approach 2 above if I
> did.)

Making it general enough to be easily customisable for each
language sounds like a better idea.

Should anyone want to use my source for this stuff, it will
soon be available as part of Swedish.h v1.1.

/Fredrik

Sophie Fruehling

unread,
May 28, 2003, 4:58:23 AM5/28/03
to
On 27 May 2003 23:26:09 -0700, dda...@nospam-icehouse.net (Daniel
Dawson) wrote:

>You pick up and read article <7XRAa.24196$9C6.1...@wards.force9.net>, written by
>Cedric Knight <ckn...@gn.babpbc.removeallBstosend.org>. It says:
>>dhan wrote:
>[Snip]
>[Forgot who wrote this originally...]
>>>> There are the ones above, plus ue for u-umlaut and ss for 'beta' in
>>>> German;
>
>That's not a 'Beta', it's an 'Ess-Zett', a ligature of the letters 's' and 'z'
>(in their archaic forms, in which the 's' looks like an 'f' without the
>horizontal stroke, and the 'z' resembles an old-fashioned '3'). It does look
>surprisingly like the Greek letter Beta, though; does anyone know if there is a
>connection, or is it just a freak coincidence.

It must be a freak coincidence. Why would the Germans have designed
their 's' and 'z' so that they look like a greek beta if put together?
Also, it doesn't much look like a 'beta' in the book I just looked in,
which was printed in Salzburg in MCMXLVI ;). There is an additional
line going up diagonally from the left upper end (of the 'beta' to the
right.

-- Sophie

Cedric Knight

unread,
May 28, 2003, 8:33:15 AM5/28/03
to
Daniel Dawson wrote:
> [Forgot who wrote this originally...]
(it was me)

>>>> There are the ones above, plus ue for u-umlaut and ss for 'beta' in
>>>> German;

I didn't mean to perpetuaute confusion, and assumed people would know
what I meant (yes, eszet).

>
> That's not a 'Beta', it's an 'Ess-Zett', a ligature of the letters
> 's' and 'z' (in their archaic forms, in which the 's' looks like an
> 'f' without the
> horizontal stroke, and the 'z' resembles an old-fashioned '3'). It
> does look surprisingly like the Greek letter Beta, though; does
> anyone know if there is a connection, or is it just a freak
> coincidence.

Well, having done a bit of hand typography for fun, I remember that
cases of type often had oe and ae ligatures, but not sz/ss. So I
imagine English typesetters often used a Greek beta, maybe from a
different font, because it was handier than making up type just for that
purpose. (I mentioned 'creme brulee' earlier because we had a nightmare
setting that in a poetry book - it's possible to add marks to characters
on the Ludlow machine we were using, but creating new characters would
be very difficult. The same was probably true for linotype work.)

It is also true that if you look at (say) eighteenth-century English
engravings, then 'ss' at the end of a word also looks very much like a
'beta'. English-speaking people may have thought the German character
was 'ss', not 'sz'. BTW it is now less frequent after the spelling
reforms, isn't it?

Anyway, this is not a problem for my purposes (finding a general set of
alternative character representations).

>
>> I meant the acute accent over a,e,i,o, or u indicating a longer or
>> heavier syllable, as in Per{'u}. I don't know what the proper name
>> for
>> them is in English or Spanish.
>
> We (English-speakers) just call it an acute accent; the opposite
> accent (`) is called a grave accent.

OK. I was reluctant to call it an acute accent, for the same reason I
put 'beta' in quotes, that acute to me describes something phonetic (a
rising or sharpened vowel) as well as a mark on the page. Just
'accent'/'acento' (meaning stress, like in music) would be correct,
maybe?

To rephrase my question:
outside Germanic languages and Esperanto which we have covered, are
there any conventions for representing accented characters where the
accents are not available, other than just dropping the accent?
(Polish 'l/w'? Turkish 's/sh'?)

Here's a run-down of what various available Inform translations seem to
do with accents:

Dutch: assumes no accents or 'ij' ligatures. All dictionary words in
lower ZSCII.

Esperanto: converts 'c^' (two characters) to 'cx' in
LanguageToInformese, but doesn't attempt 'ch' possibilities. Assumes
dictionary words are in 'cx'..'ux' format lower ZSCII.

French (J-L Portico), Spanish: By default, if a word is unrecognised in
LanguageToInformese the parser removes accents and tries again, but
author can choose to have accented dictionary words and require them
from the player if preferred.

German (inform-de stabil): Always converts umlauts to ae, oe, ue, eszet
to ss. Assumes lower ZSCII dictionary.

Spanish: As French.

Swedish (1.0): No conversion. Assumes accented dictionary words. (1.1)
user control for people without ability to enter accents?


Fredrik -
If the parser only contracts ae, oe, aa in *unrecognised* words, is
there still a need for the "!" notation? BTW I don't know what
mistype's more outrageous 'corrections' were, but you can always make
the last-resort checking less, um, enthusiastic by defining 'Constant
MISTYPE_FUZZ 10' before including the module.

CK

John W. Kennedy

unread,
May 28, 2003, 11:36:09 AM5/28/03
to
Cedric Knight wrote:
> Does anyone have a complete international list of these digraphs? There
> are the ones above, plus ue for u-umlaut and ss for 'beta' in German;

ß is "esszett" ("sz"), not "beta", although it resembles one. But "ss",
rather than "sz", is indeed used to replace it.

--
John W. Kennedy
"Sweet, was Christ crucified to create this chat?"
-- Charles Williams: "Judgement at Chelmsford"

dhan

unread,
May 28, 2003, 12:13:51 PM5/28/03
to
Hi Cedric,

>>> in Spanish 'ny' can be used, and I believe the stress marks
> >> are ignored. Portuguese? French? Norwegian/Danish?
> >
> > In Spanish "ny" doesn´t replace "n-tilde", only in Catalan, a
>
> Oh, I thought I had seen it somewhere, maybe I was mistaken. But if you
> didn't have an n-tilde on the keyboard, how would you show it?

Yes, I have a n-tilde on the keyboard, I only said that "ny" doesn´t
replace it. In another language spoken in Spain, Catalunya, the "ny"
replaces the n-tilde.

> > north-eastern community with his own language. By the way, what is a
> > stress mark? I don´t know if I ignore them :-)
>
> I meant the acute accent over a,e,i,o, or u indicating a longer or
> heavier syllable, as in Per{'u}. I don't know what the proper name for
> them is in English or Spanish.

Ah, you were speaking about "tildes", that is the spanish word for
stress mark. I want to test mistype.h library on spanish version of
Inform, I will tell you how it works.

Regards,
dhan

Cedric Knight

unread,
May 28, 2003, 1:38:42 PM5/28/03
to
dhan wrote:
> Hi Cedric,
>

>>> In Spanish "ny" doesn´t replace "n-tilde", only in Catalan, a
>>
>> Oh, I thought I had seen it somewhere, maybe I was mistaken. But if
>> you didn't have an n-tilde on the keyboard, how would you show it?
>
> Yes, I have a n-tilde on the keyboard, I only said that "ny" doesn´t
> replace it. In another language spoken in Spain, Catalunya, the "ny"
> replaces the n-tilde.

OK. Suppose how would you type 'little girl' in Castellano without
pressing that key? How do you represent the Spanish for 'gnu' (large
African antelope) using only 7-bit ASCII?

And if you didn't have the key (I think to the right of 'p') to produce
' tildes, would it inconvenience you? How would you say 'behind'?

>> I meant the acute accent over a,e,i,o, or u indicating a longer or
>> heavier syllable, as in Per{'u}. I don't know what the proper name
>> for them is in English or Spanish.
>
> Ah, you were speaking about "tildes", that is the spanish word for
> stress mark.

Oh, I see, how confusing!

> I want to test mistype.h library on spanish version of
> Inform, I will tell you how it works.

Thanks, that would be valuable data. Spanish Inform (InformATE) makes
its own changes in a similar way to mistype.

CK

Cedric Knight

unread,
May 28, 2003, 1:39:35 PM5/28/03
to
John W. Kennedy wrote:
> Cedric Knight wrote:
>> Does anyone have a complete international list of these digraphs?
>> There are the ones above, plus ue for u-umlaut and ss for 'beta' in
>> German;
>
> ß is "esszett" ("sz"), not "beta", although it resembles one. But
> "ss", rather than "sz", is indeed used to replace it.

In that case, why does the IBM PC character set (codes 0xE0-0xE2) go
'alpha, esszett, gamma'? :)

CK


L. Ross Raszewski

unread,
May 28, 2003, 5:49:33 PM5/28/03
to

WHat you're seeing actually is a beta. Character code 0xE1 is indeed a
greek beta, and, I'd even wager that most times you see an esszett on
your computer screen, you're actually seeing a beta. Esszett looks so
much like a beta that th' one is generally used to substitute for
th'other, especially in a character set where space counts.

Daniel Dawson

unread,
May 28, 2003, 10:13:11 PM5/28/03
to
You pick up and read article <chu8dv41js5tp28mb...@4ax.com>,

written by Sophie Fruehling <sfrue...@aon.at>. It says:

>Also, it doesn't much look like a 'beta' in the book I just looked in,
>which was printed in Salzburg in MCMXLVI ;). There is an additional
>line going up diagonally from the left upper end (of the 'beta' to the
>right.

Which direction?

Also, I have a book lying around that was printed in Stuttgart. The 's' part of
the 'ß' therein, instead of flowing into the 'z' part as it does in modern
fonts, actually curls over the 'z'; it looks almost exactly like an ('f'-like)
's' with the the 'z' stuck to it (though if you look closely, the curl of the
's' extends slightly further). Oh yeah, I forgot to mention, the typeface is a
strong Gothic.

Daniel Dawson

unread,
May 28, 2003, 10:56:50 PM5/28/03
to
You pick up and read article <242Ba.30437$xd5.1...@stones.force9.net>,

written by Cedric Knight <ckn...@gn.babpbc.removeallBstosend.org>. It says:

>It is also true that if you look at (say) eighteenth-century English
>engravings, then 'ss' at the end of a word also looks very much like a
>'beta'. English-speaking people may have thought the German character
>was 'ss', not 'sz'. BTW it is now less frequent after the spelling
>reforms, isn't it?

Yep, that's for sure. 'daß' -> 'dass', etc. Speaking of which, the Germans love
to officially revise their language, don't they?

>OK. I was reluctant to call it an acute accent, for the same reason I
>put 'beta' in quotes, that acute to me describes something phonetic (a
>rising or sharpened vowel) as well as a mark on the page. Just
>'accent'/'acento' (meaning stress, like in music) would be correct,
>maybe?

Well, the marks acute (´), grave (`), tilde (~), umlaut/diaeresis (¨), and
circumflex (^) (as well as others that don't exist in this character set) are
all called accents. However, 'accent' also refers to stressing of syllables,
whether indicated by marks or not. Isn't English great?

Sophie Fruehling

unread,
Jun 1, 2003, 7:19:55 AM6/1/03
to
On 28 May 2003 19:13:11 -0700, dda...@nospam-icehouse.net (Daniel
Dawson) wrote:

>You pick up and read article <chu8dv41js5tp28mb...@4ax.com>,
>written by Sophie Fruehling <sfrue...@aon.at>. It says:
>
>>Also, it doesn't much look like a 'beta' in the book I just looked in,
>>which was printed in Salzburg in MCMXLVI ;). There is an additional
>>line going up diagonally from the left upper end (of the 'beta' to the
>>right.
>
>Which direction?
>
>Also, I have a book lying around that was printed in Stuttgart. The 's' part of
>the 'ß' therein, instead of flowing into the 'z' part as it does in modern
>fonts, actually curls over the 'z'; it looks almost exactly like an ('f'-like)
>'s' with the the 'z' stuck to it (though if you look closely, the curl of the
>'s' extends slightly further). Oh yeah, I forgot to mention, the typeface is a
>strong Gothic.

Well, I guess, it looks pretty much like the one you describe. It
doesn't say what the typeface is called, though. (It must have been
one of the last books printed in such a font, given that it was
published in 1946.)

-- Sophie

Reply all
Reply to author
Forward
0 new messages