Re: Berber/Tifinagh (was: Swahili Banthu)

[John Cowan]

Alexander Savenkov scripsit:

 I'm not sure I'm not taking your words out of the context, Michael.

You are.  Michael is complaining not about transliteration as such,
but about instant transliteration by font substitution.

-- 
John Cowan  [EMAIL PROTECTED]  www.ccil.org/~cowan  www.reutershealth.com
I must confess that I have very little notion of what [s. 4 of the British
Trade Marks Act, 1938] is intended to convey, and particularly the sentence
of 253 words, as I make them, which constitutes sub-section 1.  I doubt if
the entire statute book could be successfully searched for a sentence of
equal length which is of more fuliginous obscurity. --MacKinnon LJ, 1940

Re: Berber/Tifinagh (was: Swahili Banthu)

[Patrick Andries]

- Message d'origine - 

Philippe Verdy a écrit :

I was concerned recently by some people who wanted to better write the
Tifinagh languages

Stricto sensu, they are no tifinagh languages, but languages (or dialects of
the Berber language) written with the tifinagh script.

(such as Berber) with the Latin script (notably for North
Africa, but also in Europe due to the important North African community,
notably in France).

I'm not sure why you want to privilege the Latin script. A large part in
chosing a script for tifinagh is based on subjective reasons. Tifinagh is
seen as a neutral (neither Western, neither Arabic), indigeneous and deeply
historically-rooted script. These sentiments must be respected and
communities may choose whatever script they feel best suited for their own
needs (affirming one's identity should not be disregarded here).


 When the Tifinagh script will be standardized, it would then be
interesting
to allow it to be rendered correctly with Latin letters and diacritic
glyphs
on a user font preference, as it corresponds more to the now modern use of
the script...

This makes no sense : the modern use of the Tifinagh script cannot be
another script... You may have meant the modern day script used for the
berber language. This is highly disputable (Morocco just started teaching
Tifinagh in its schools and they are many Berber sites in Tifinagh and
Arabic scripts). It also does not follow that being able to switch fonts
between two scripts used for a language should take precedence over coding
properly the different scripts. What if there is no one-to-one mapping
between the characters of the different scripts ?  What about a mapping
between the Arabic and Tifinagh transcriptions of  Berber ? Should we be
able to switch between the Arabic and Latin transcriptions by a simple font
adjustment ? What is one then to do with the « emphatiques conditionnées »
noted by some scholars in Latin and Tifinagh but for which no Arabic
character seems to be available ?

Patrick Andries

Re: Berber/Tifinagh (was: Swahili Banthu)

[Patrick Andries]

De: Patrick Andries [EMAIL PROTECTED]



 - Message d'origine - 

 Philippe Verdy a écrit :
 
 I was concerned recently by some people who wanted to better write the
 Tifinagh languages

 Stricto sensu, they are no tifinagh languages, but languages (or dialects
of
 the Berber language) written with the tifinagh script.

 (such as Berber) with the Latin script (notably for North
 Africa, but also in Europe due to the important North African community,
 notably in France).

 I'm not sure why you want to privilege the Latin script. A large part in
 chosing a script for tifinagh

Recte : for Berber ;-)

Patrick Andries
- o - O - o -
L'arroseur arrosé

Re: Berber/Tifinagh (was: Swahili Banthu)

[Don Osborn]

Patrick's message on this topic gets to the heart of the issue of why to
encode Tifinagh (as Tifinagh) in the first place.  But I think that
Philippe's sentiment is not misplaced, if one approaches transliteration on
the character and not the glyph level, as John and others put it.  But
correspondences between characters or groups of characters (not
necessarily - and perhaps not possibly in some cases - one-to-one) among
different scripts are a linguistic issue not a coding one.

The advantage to clarifying the correspondences among alternative
transcriptions is indicated in part by Philippe's original concern - people
who learn Berber (more properly Tamazight, Tamashek, etc.) in one
transcription but not another.  How to facilitate their encounter with the
diverse renditions of the language?  But again it's not a font issue.

I've thought for instance about the small number of schools here in Niger
that teach in Tamajak, using the Latin based script and how easy it will or
will not be for the students to make the connections with the Tifinagh that
is traditionally used.  It would be tidy to have one-to-one correspondences,
but even if not, some fairly consistent rules would help.  I'm not sure the
extent to which people working on Tamajak in Latin orthography (presumably
from the spoken language) make reference to traditions of spelling with
Tifinagh, but it would seem essential.

And it is perhaps imporant too to take this on a broader scale to understand
the traditional ways of writing Berber in Tifinagh (and Arabic) and to
harmonize the Latin transliterations in the region.  It may well be that
aside from whatever complexities there may be on the character-to-character
level, that there may be different conventions arising in the language(s) as
written in different scripts - not to mention that there may still be work
yet to do in standardizing spellings in either script within any given
variety of Berber.  Dealing with such issues would not be served by treating
Tifinagh as anything less than a script in its own right.

And as Berber is increasingly used on computers  the internet, it will
become important to be able to do reliable conversions of text in any of the
three transcriptions - work towards which may bring up its own set of
questions.  And these would be best dealt with on the character level.

Don Osborn
Bisharat.net

- Original Message - 
From: Patrick Andries [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, November 10, 2003 8:33 PM
Subject: Re: Berber/Tifinagh (was: Swahili  Banthu)



 - Message d'origine - 

 Philippe Verdy a écrit :
 
 I was concerned recently by some people who wanted to better write the
 Tifinagh languages

 Stricto sensu, they are no tifinagh languages, but languages (or dialects
of
 the Berber language) written with the tifinagh script.

 (such as Berber) with the Latin script (notably for North
 Africa, but also in Europe due to the important North African community,
 notably in France).

 I'm not sure why you want to privilege the Latin script. A large part in
 chosing a script for tifinagh [Berber] is based on subjective reasons.
Tifinagh is
 seen as a neutral (neither Western, neither Arabic), indigeneous and
deeply
 historically-rooted script. These sentiments must be respected and
 communities may choose whatever script they feel best suited for their own
 needs (affirming one's identity should not be disregarded here).


  When the Tifinagh script will be standardized, it would then be
 interesting
 to allow it to be rendered correctly with Latin letters and diacritic
 glyphs
 on a user font preference, as it corresponds more to the now modern use
of
 the script...

 This makes no sense : the modern use of the Tifinagh script cannot be
 another script... You may have meant the modern day script used for the
 berber language. This is highly disputable (Morocco just started teaching
 Tifinagh in its schools and they are many Berber sites in Tifinagh and
 Arabic scripts). It also does not follow that being able to switch fonts
 between two scripts used for a language should take precedence over coding
 properly the different scripts. What if there is no one-to-one mapping
 between the characters of the different scripts ?  What about a mapping
 between the Arabic and Tifinagh transcriptions of  Berber ? Should we be
 able to switch between the Arabic and Latin transcriptions by a simple
font
 adjustment ? What is one then to do with the « emphatiques conditionnées »
 noted by some scholars in Latin and Tifinagh but for which no Arabic
 character seems to be available ?

 Patrick Andries

Re: Berber/Tifinagh (was: Swahili Banthu)

[Philippe Verdy]

From: Patrick Andries [EMAIL PROTECTED]
 This makes no sense : the modern use of the Tifinagh script cannot be
 another script... You may have meant the modern day script used for the
 berber language. This is highly disputable (Morocco just started teaching
 Tifinagh in its schools and they are many Berber sites in Tifinagh and
 Arabic scripts). It also does not follow that being able to switch fonts
 between two scripts used for a language should take precedence over coding
 properly the different scripts. What if there is no one-to-one mapping
 between the characters of the different scripts ?  What about a mapping
 between the Arabic and Tifinagh transcriptions of  Berber ? Should we be
 able to switch between the Arabic and Latin transcriptions by a simple
font
 adjustment ? What is one then to do with the « emphatiques conditionnées »
 noted by some scholars in Latin and Tifinagh but for which no Arabic
 character seems to be available ?

This is the role of diacritics and symbols added to the target script, so
that no information from the text written in the source script is lost.

You seem to forget that Tifinagh is not a unified script, but a set of
separate
scripts where the same glyphs are used with distinct semantic functions.
Byt itself, ignoring all other transliteration to Latin and Arabic, the
Tifinagh
scripts are already cyphers of another variant of Tifinagh script.

And I think it is the major issue which requires to choose a policy for its
encoding. If characters are encoded by their names (as they should in
Unicode)
then we are unable to produce an accurate chart showing representative
glyphs, as no variant of the script covers the whole abstract character
set,
and so this would require several charts, i.e. multiple glyphs for the same
abstract character. In this condition, why couldn't Latin glyphs be among
these, when they already have the merit of covering the whole abstract
character set covered by all scripts in the Tifinagh family?

Re: Berber/Tifinagh (was: Swahili Banthu)

[Patrick Andries]

- Message d'origine - 
De: Don Osborn [EMAIL PROTECTED]


 I've thought for instance about the small number of schools here in Niger
 that teach in Tamajak, using the Latin based script and how easy it will
or
 will not be for the students to make the connections with the Tifinagh
that
 is traditionally used.  It would be tidy to have one-to-one
correspondences,
 but even if not, some fairly consistent rules would help.  I'm not sure
the
 extent to which people working on Tamajak in Latin orthography (presumably
 from the spoken language) make reference to traditions of spelling with
 Tifinagh, but it would seem essential.

The Touareg (like the new Volkswagen) traditionnal usage of Tifinagh is very
defective if recall properly Hanoteau (in its Tamachek' Grammar this from
memory, not being at home) : geminates not noted, short vowels usually not
written, etc. I'm not too sure how you can pass from traditional Touareg
Tifinagh to the Latin-based script.

 And it is perhaps imporant too to take this on a broader scale to
understand
 the traditional ways of writing Berber in Tifinagh (and Arabic) and to
 harmonize the Latin transliterations in the region.  It may well be that
 aside from whatever complexities there may be on the
character-to-character
 level, that there may be different conventions arising in the language(s)
as
 written in different scripts - not to mention that there may still be work
 yet to do in standardizing spellings in either script within any given
 variety of Berber.  Dealing with such issues would not be served by
treating
 Tifinagh as anything less than a script in its own right.

I also believe it but I'm also not sure that Tifinagh -- still not
encoded -- is best served by us waiting until Latin transliterations of
Berber (Chleuh, Kabyle,Tamachek') are harmonized among populations scattered
in many countries.

P. A.

Re: Berber/Tifinagh (was: Swahili Banthu)

[Patrick Andries]

- Original Message - 
De: Philippe Verdy [EMAIL PROTECTED]

 This is the role of diacritics and symbols added to the target script, so
 that no information from the text written in the source script is lost.

Yes,  I know this but you cannot go from Berber written in Arabic to
Tifinagh or Latin Scripts without losing some information. At least,
according to some scholars' system. Do you also plan to modify the Arabic
script as used by the current user of Berbere before coding the Tifinagh
script ?


 You seem to forget that Tifinagh is not a unified script,

(sigh)

 but a set of separate scripts where the same glyphs are used with distinct
semantic functions.

Phonemic, I suspect. In any case, what is the problem with several
characters having the same representative glyph in the code chart (they
also have a name) ? I believe this is already the case for other characters.
Some characters don't even have no glyph representation.

 And I think it is the major issue which requires to choose a policy
for its encoding. If characters are encoded by their names (as they should
in
 Unicode)

Maybe (their phonemic value seems a good candidate to me, since letters may
have different names in different areas).

 then we are unable to produce an accurate chart showing representative
 glyphs, as no variant of the script covers the whole abstract character
 set, and so this would require several charts, i.e. multiple glyphs for
the same
 abstract character.

I don't see the problem, isn't this the whole idea in distinguishing between
glyph and characters ?

 In this condition, why couldn't Latin glyphs be among
 these, when they already have the merit of covering the whole abstract
 character set covered by all scripts in the Tifinagh family?

Because it is best to use Tifinagh glyphs as representative glyphs of the
Tifinagh script?  Because the character names also specify the value of
characters? And because, even if the script is not unified and some sign are
not used everywhere, many signs are. I counted 16 common glyphs between the
Hoggar variety and the Académie berbère's one, out a total of 28 glyphs
found in these two varieties. Nine characters use different glyphs in the
two varieties and 3 characters exists in Académie berbère's proposal but
have no mapping (even ambiguous ones) in the Hoggar's variety. Only in the
case of vowels, where the Hoggar variety which does not distinguish between
vowels and semi-vowels (has 2 signs instead of 4), would it be ambiguous to
use the Hoggar signs since the Académie Berbère uses different signs for the
semivowels /w/ and /j/. But I agree that chosing the representative glyphs
may become a sensitive issue if the Tifinagh script is to be unified, each
school might feel offended that its preferred glyphs were not chosen in
ISO/IEC 10646. This does not necessarily mean that Tifinagh should be
encoded with an easy Latin mapping in mind.

P. A.

Re: Berber/Tifinagh (was: Swahili Banthu)

[Kenneth Whistler]

Philippe Verdy wrote:

 You seem to forget that Tifinagh is not a unified script, but a set of
 separate
 scripts where the same glyphs are used with distinct semantic functions.

I think Philippe is running off the rails here.

Tifinagh is a script. It comes in a number of local varieties,
adapted to different languages and with local variations in
glyph preferences. It will be encoded as a *single* script in
Unicode, since encoding all the local orthographic varieties
as distinct scripts would really not be a service to anyone
who wants this script encoded for enabling IT processing of
Berber textual data.

The situation for Tifinagh can be profitably compared with the
Unified Canadian Aboriginal Syllabics, I think. (U+1401..U+1676)
The Cree syllabics were adapted all across Canada, and in that
case across major language family boundaries, from Algonquian
languages into Athabaskan and Inuit languages. During that process,
there were many extensions, reuses, adaptations, and inconsistencies
in the symbolic usage. Rather than encode a half dozen different
scripts for this, one for each local orthographic tradition, the
entire script was carefully unified to enable representation of
any of the local varieties accurately with the overall script
encoding. I suspect that a similar approach will be required to
finish the encoding of Tifinagh.

 Byt itself, ignoring all other transliteration to Latin and Arabic, the
 Tifinagh
 scripts are already cyphers of another variant of Tifinagh script.

This is a complete misuse of the term cypher.

There are differences in glyph usage, and there may be differences
in character usage (depending on the solution chosen for the
encoding) between local, historic varieties of Tifinagh. These
differences can be mapped against each other, but such a mapping
does not constitute a cypher, any more than local Runic traditions
(Nordic, Germanic, Anglo-Friesian) are ciphers of each other.

 
 And I think it is the major issue which requires to choose a policy for its
 encoding. If characters are encoded by their names (as they should in
 Unicode)

They are not, nor should they be.

 then we are unable to produce an accurate chart showing representative
 glyphs, as no variant of the script covers the whole abstract character
 set,
 and so this would require several charts, i.e. multiple glyphs for the same
 abstract character. 

There is nothing new about this. As the discussion in the Unicode
Standard makes clear, this is already the case for the Latin
script and many other scripts. It is patently obvious for the
Han script, among others. This is just basics of the Character-Glyph
Model.

 In this condition, why couldn't Latin glyphs be among
 these,

Because, as Doug already pointed out, Latin characters are not
Tifinagh characters, so substituting Latin glyphs for Tifinagh
glyphs constitutes lying about the identity of the characters.
Arbitrary substitutions like this outside the context of
glyph variation inherent to a character identified as a unit
of a script are *not* allowed by the Unicode Standard's
conformance clauses. It runs afoul of the basic clause regarding
character identity and interpretation, C7.

You can always shift textual data between scripts, of course,
but that is a knowing transformation of characters, known
as *transliteration*. It is *not* merely adding more glyphs
to the allowable range of glyph variation for a character. 

 when they already have the merit of covering the whole abstract
 character set covered by all scripts in the Tifinagh family?

You could say the same about any script whatsoever, as I
suspect that *every* script in Unicode has been transliterated
into the Latin script at one point or another. Why not just
map them *all* to Latin and save the messy task of having to
deal with data represented in its own script? (== That was
a rhetorical question, in case it wasn't obvious to all readers.)

--Ken

Re: Berber/Tifinagh (was: Swahili Banthu)

[Michael Everson]

Don,

Berber is often written in Tifinagh without vowels. And sometimes 
with vowels. Andd the same in Arabic. There is no point worrying 
(without it even being encoded) about Latin transliteration standards 
for it at this point.
--
Michael Everson * * Everson Typography *  * http://www.evertype.com

Re: Berber/Tifinagh (was: Swahili Banthu)

[Michael Everson]

At 23:33 +0100 2003-11-10, Philippe Verdy wrote:

You seem to forget that Tifinagh is not a unified script, but a set 
of separate scripts
What?

where the same glyphs are used with distinct semantic functions.
We haven't decided what kind of unification is appropriate for 
Tifinagh entities yet.

Byt itself, ignoring all other transliteration to Latin and Arabic, the
Tifinagh scripts are already cyphers of another variant of Tifinagh script.
I think this is rather muddled.

If characters are encoded by their names (as they should in Unicode)
What?

then we are unable to produce an accurate chart showing 
representative glyphs, as no variant of the script covers the 
whole abstract character set,
and so this would require several charts, i.e. multiple glyphs for the same
abstract character.
I made a chart. It was a start.

In this condition, why couldn't Latin glyphs be among
these, when they already have the merit of covering the whole abstract
character set covered by all scripts in the Tifinagh family?
Gosh. Because Latin is a different script.
--
Michael Everson * * Everson Typography *  * http://www.evertype.com

Re: Berber/Tifinagh (was: Swahili Banthu)

[Philippe Verdy]

From: Kenneth Whistler [EMAIL PROTECTED]
 Rather than encode a half dozen different
 scripts for this, one for each local orthographic tradition, the
 entire script was carefully unified to enable representation of
 any of the local varieties accurately with the overall script
 encoding. I suspect that a similar approach will be required to
 finish the encoding of Tifinagh.

At least! That's what I wanted to hear: that the script will be
encoded on a glyphic approach, with no intented association
with the actual phoneme they represent.

The ordering of these glyphs and their naming will possibly
be consistent with one culture, but not with another.

And each culture will use only a subset of the encoded glyphs...

And so UCA default collation will work reasonnably well for
one culture and not the others that will require tailoring.

Now, how will we define foldings, and text boundaries for
the other cultures? I don't know...

How will we perform intercultural semantic analysys for
texts that share the same words and radicals? Difficult
to answer... It will even be hard to define any
transliteration scheme between them (it would require
identification of the cultural convention, and not only
of the language, and texts written for one culture will
look as completely undecipherable in another one).

Such problems do not exist with the same level between
Serbian Latin and Serbian Cyrillic, as there's no ambiguity
in the transliteration to use...

Re: Berber/Tifinagh (was: Swahili Banthu)

[Patrick Andries]

- Message d'origine - 
De: Philippe Verdy [EMAIL PROTECTED]



 From: Patrick Andries [EMAIL PROTECTED]
   In this condition, why couldn't Latin glyphs be among
   these, when they already have the merit of covering the whole abstract
   character set covered by all scripts in the Tifinagh family?
 
  Because it is best to use Tifinagh glyphs as representative glyphs of
the
  Tifinagh script?

 No: the simple reason is the choice of the representative glyph, which
 will probably be accurate for one cultural convention but completely wrong
with
 another, as that glyph represent another phoneme coded at a different
 place where another representative glyph is used, which may also be
 wrong.

I understood your objection. I answered this is rare and names help.

 Look at the phonemic meaning of the glyph that looks like two triangles,
 pointing top and bottom to each other. Look at the glyph which looks like
a
 moon crescent (open on right side) with a dot in the middle... Which
 phonemic value do they have? This depends on cultural conventions, and it
 really looks as if there was not _one_ but several distinct Tifinagh
scripts
 using the same glyphs but with incompatible phonemic values...

1) Could you tell me what the values of those characters are for you and in
which varieties of the script ? I know of this sand-glass character but I
have a single suspicious source giving me this as having a different value
from the traditional one. I'll happily collect a second source (en privé si
tu veux) to list ambiguities. Could you also be specific for the
half-crescent?
2) As mentioned before, characters have a name and it can be used to
uniquely identify the character if need be (i.e. the referential glyph is
only a help and may not be sufficient to uniquely identify a character).

  But I agree that chosing the representative glyphs
  may become a sensitive issue if the Tifinagh script is to be unified,
each
  school might feel offended that its preferred glyphs were not chosen in
  ISO/IEC 10646. This does not necessarily mean that Tifinagh should be
  encoded with an easy Latin mapping in mind.

 I'm just suggesting that if the phonemic encoding model is used,

I don't know if this is what will be chosen (note that this is different
from your name-based scheme you suggested a few messages ago).

 the choice
 of representative glyphs will create confusion, as it will privilegiate
 one interpretation of the glyphs and not the other one. Polemics are
already
 present on the Internet because of the choice of interpretation that has
 been made by Morocco, which excludes other interpretations.

Could I have a pointer ? Tu peux me l'envoyer sous courriel privé.

P. A.

Re: Berber/Tifinagh (was: Swahili Banthu)

[Michael Everson]

At 02:04 +0100 2003-11-11, Philippe Verdy wrote:
From: Kenneth Whistler [EMAIL PROTECTED]
 Rather than encode a half dozen different
 scripts for this, one for each local orthographic tradition, the
 entire script was carefully unified to enable representation of
 any of the local varieties accurately with the overall script
 encoding. I suspect that a similar approach will be required to
 finish the encoding of Tifinagh.
At least! That's what I wanted to hear: that the script will be
encoded on a glyphic approach, with no intented association
with the actual phoneme they represent.
We have not decided what we will do with Tifinagh.
--
Michael Everson * * Everson Typography *  * http://www.evertype.com

Re: Berber/Tifinagh (was: Swahili Banthu)

[Mark E. Shoulson]

Kenneth Whistler wrote:

Philippe Verdy wrote:

 

You seem to forget that Tifinagh is not a unified script, but a set of
separate
scripts where the same glyphs are used with distinct semantic functions.
   

I think Philippe is running off the rails here.

Tifinagh is a script. It comes in a number of local varieties,
adapted to different languages and with local variations in
glyph preferences. It will be encoded as a *single* script in
Unicode, since encoding all the local orthographic varieties
as distinct scripts would really not be a service to anyone
who wants this script encoded for enabling IT processing of
Berber textual data.
Compare tengwar; much the same situation.  I realize tengwar isn't yet 
encoded, but I think there's no question of how it should be encoded, if 
it is.

when they already have the merit of covering the whole abstract
character set covered by all scripts in the Tifinagh family?
   

You could say the same about any script whatsoever, as I
suspect that *every* script in Unicode has been transliterated
into the Latin script at one point or another. Why not just
map them *all* to Latin and save the messy task of having to
deal with data represented in its own script? (== That was
a rhetorical question, in case it wasn't obvious to all readers.)
 

All of Unicode is just a cypher to strings of [0123456789ABCDEF], 
really.  Or to strings from [01], come to that.

~mark

Re: Berber/Tifinagh (was: Swahili Banthu)

[John Jenkins]

On Nov 10, 2003, at 5:46 PM, Philippe Verdy wrote:

Look at the phonemic meaning of the glyph that looks like two 
triangles,
pointing top and bottom to each other. Look at the glyph which looks 
like a
moon crescent (open on right side) with a dot in the middle... Which
phonemic value do they have? This depends on cultural conventions, and 
it
really looks as if there was not _one_ but several distinct Tifinagh 
scripts
using the same glyphs but with incompatible phonemic values...

You mean, like Latin?  Or Han?


John H. Jenkins
[EMAIL PROTECTED]
[EMAIL PROTECTED]
http://homepage..mac.com/jhjenkins/

Re: Berber/Tifinagh (was: Swahili Banthu)

[Philippe Verdy]

From: Don Osborn [EMAIL PROTECTED]

 As for other African scripts, they are most notable in the western
 and northern parts of the continent.  Tifinagh and N'ko are in the
 process of being encoded.  I just had a conversation with someone
 the other day who recounted seeing a letter written in Tifinagh
 script in a rural part of northern Niger written by someone to a
 local chief - quietly this script continues to be used.

I was concerned recently by some people who wanted to better write the
Tifinagh languages (such as Berber) with the Latin script (notably for North
Africa, but also in Europe due to the important North African community,
notably in France).

The current situation of the Berber language can no longer be maintained:
there's a real need to support the language with a unique encoding system,
even if it involves variants for glyphs (there are differences between the
official Moroccoan initiative and the desires of Berber people in other
African or European countries).

One of the most acute problem comes with the representation of Ayin, and
problems related to the case conversions of this letter (what is the correct
way to represent the uppercase Ayain? There are divergences as some will
want to borrow a Greek Sigma glyph, others will prefer the mirrored 3-shaped
glyph).

Same problem with the dot below consonnants: is it really a dot-below
diacritic or shouldn't there be a separate encoding for the dotted
consonnants, which would allow them to be rendered as dot below or above or
on the right or with an asterisk glyph, possibly with a compatibility
mapping to the sequence Latin consonnant, combining dot-below

When the Tifinagh script will be standardized, it would then be interesting
to allow it to be rendered correctly with Latin letters and diacritic glyphs
on a user font preference, as it corresponds more to the now modern use of
the script...

It would have the benefit of allowing interchanges of dictionnaries and
texts even if they are rendered differently. It could be possible if the
transliteration between the historic Tifinagh script and the Latin script
obeys to precise presentation rules, and also possible because there does
not seem to exist for now a precise orthograph of Tifinagh-based languages
when they are written with the Latin script (and this does not facilitate
the exchange of information between people sharing the same language but
distinct conventions for the written language).

Re: Berber/Tifinagh (was: Swahili Banthu)

[Michael Everson]

At 15:53 +0100 2003-11-09, Philippe Verdy wrote:

I was concerned recently by some people who wanted to better write the
Tifinagh languages (such as Berber) with the Latin script (notably for North
Africa, but also in Europe due to the important North African community,
notably in France).
Why? People do what they want. The Maltese have no trouble with the 
Latin script.

When the Tifinagh script will be standardized, it would then be interesting
to allow it to be rendered correctly with Latin letters and diacritic glyphs
on a user font preference, as it corresponds more to the now modern use of
the script...
What? The Tifinagh script is not the Latin script.

It would have the benefit of allowing interchanges of dictionnaries and
texts even if they are rendered differently. It could be possible if the
transliteration between the historic Tifinagh script and the Latin script
obeys to precise presentation rules, and also possible because there does
not seem to exist for now a precise orthograph of Tifinagh-based languages
when they are written with the Latin script (and this does not facilitate
the exchange of information between people sharing the same language but
distinct conventions for the written language).
When we encode Tifinagh we will encode Tifinagh. We will not 
meta-encode it for ease of transliteration to other scripts.
--
Michael Everson * * Everson Typography *  * http://www.evertype.com

Re: Berber/Tifinagh (was: Swahili Banthu)

[Philippe Verdy]

From: Michael Everson [EMAIL PROTECTED]

 When we encode Tifinagh we will encode Tifinagh. We will not
 meta-encode it for ease of transliteration to other scripts.

Yes that was the intent of my suggestion, I don't say that this must be
done. But what would be wrong if a font was created for the Tifinagh script
that would display Latin-based glyphs with diacritics rather than historic
glyphs?

Could it help to allow exchanges between use of the historic script in
Morocco, and those that criticize Morocco and now want to support their
language with Latin-based glyphs?

Of course this would exhibit differences of orthograph in the visual
rendering between a text written in Tifinagh then displayed with Latin
glyphs, and the same text written normally with the Latin script.

But it could help exhibit these cultural differences, as well as it could
help studying the historic texts by people which are not used to read the
Tifinagh glyphs, and would greatly appreciate to be able to interpret them
with Latin glyphs but with their native orthograph (later they could laern
to read the historic glyphs of this script with exactly the same texts).

This is not really a proposal for a transliteration, as it does not want to
map Tifinagh letters with their possible Latin equivalents, taking into
account the history of this transliteration, which may have modified the
orthograph to be easier to read in the Latin script (for example if letters
have then been added or changed in the Latin script for the modern form of
the language).

This problem is similar between Latin and Cyrillic version of
Serbo-Croatian, which is still considered as a single language (even if
there are lexical preferences now between Serbian and Croatian, only Serbian
being written today with both the Latin and Cyrillic alphabet).

I'm sorry if this seems stupid for you. The intent is not to deprecate the
historic script by writing it systematically (with a bijective mapping) with
Latin glyphs: such system will look possibly ugly in some cases for users
which have learned to write their language with the Latin script which
includes now its own orthograph.

Re: Berber/Tifinagh (was: Swahili Banthu)

[Michael Everson]

At 17:54 +0100 2003-11-09, Philippe Verdy wrote:
From: Michael Everson [EMAIL PROTECTED]

 When we encode Tifinagh we will encode Tifinagh. We will not
 meta-encode it for ease of transliteration to other scripts.
Yes that was the intent of my suggestion, I don't say that this must 
be done. But what would be wrong if a font was created for the 
Tifinagh script that would display Latin-based glyphs with 
diacritics rather than historic glyphs?
I think you should go back and learn about the characater/glyph model 
if you think this is a good idea.

Could it help to allow exchanges between use of the historic script 
in Morocco, and those that criticize Morocco and now want to support 
their language with Latin-based glyphs?
If you want to write Tifinagh in Latin, you should write it in Latin.
--
Michael Everson * * Everson Typography *  * http://www.evertype.com

Re: Berber/Tifinagh (was: Swahili Banthu)

[Philippe Verdy]

From: Michael Everson [EMAIL PROTECTED]
 At 17:54 +0100 2003-11-09, Philippe Verdy wrote:
 From: Michael Everson [EMAIL PROTECTED]
   When we encode Tifinagh we will encode Tifinagh. We will not
   meta-encode it for ease of transliteration to other scripts.
 
 Yes that was the intent of my suggestion, I don't say that this must
 be done. But what would be wrong if a font was created for the
 Tifinagh script that would display Latin-based glyphs with
 diacritics rather than historic glyphs?

 I think you should go back and learn about the characater/glyph model
 if you think this is a good idea.

I think we are bothering each other, simply because of mutual
misunderstanding. There's nothing in my question above that breaks that
model: the abstract character coded in the Tifinagh script is kept unchanged
and separated from the actual glyph it uses when rendering.

So my question is, once again: would a font that would display pointed Latin
glyphs from Tifinagh script code points really break the Unicode model? If
not, then we have a convenient way to define Tifinagh keyboards / input
methods based on this _apparent_ transliteration. This still requires a
specific mapping to codepoints in the keyboard driver when set to input for
Tifinagh, allows editing with a set of glyphs that a user can read, and then
render it transparently with a real Tifinagh font.

This would be a great tool for example on web site: the same text or web
page could be displayed with the Latin glyphs or with the historic Tifinagh
glyphs, by only selecting a distinct font. So a page in Berber would be
composed once, and readable from both communities that can read either the
Latin script or the Tifinagh script, and there would be now no need to use
one of the many kludge conventions that exist for that language when
presenting contents in Berber (notably for those pages that are coded with a
mix of Latin letters, Greek letters, or symbols like dollar and asterisk...

The purpose of this suggestion (which does not affect the separation between
Unicode abstract characters and glyphs) is _better accessibility_. Nothing
else.

Re: Berber/Tifinagh (was: Swahili Banthu)

[Michael Everson]

At 19:30 +0100 2003-11-09, Philippe Verdy wrote:

So my question is, once again: would a font that would display pointed Latin
glyphs from Tifinagh script code points really break the Unicode model?
Yes, Philippe. It is the same thing as mapping Cyrillic to ASCII 
letters. It is a hack. It is to be avoided. It is the Wrong Thing To 
Do.

If not, then we have a convenient way to define Tifinagh keyboards / 
input methods based on this _apparent_ transliteration.
This has nothing to do with encoding. You are harkening back to the 
hideous world of 8-bit font hacks of twenty years ago.

This still requires a specific mapping to codepoints in the keyboard 
driver when set to input for Tifinagh, allows editing with a set of 
glyphs that a user can read, and then render it transparently with a 
real Tifinagh font.
You are confusing character encoding, font, and input method.

This would be a great tool for example on web site: the same text or 
web page could be displayed with the Latin glyphs or with the 
historic Tifinagh glyphs, by only selecting a distinct font.
Of course. We don't need the Georgian or Armenian alphabets any more 
either. Just make everything a glyph representation of Latin.

So a page in Berber would be composed once, and readable from both 
communities that can read either the Latin script or the Tifinagh 
script, and there would be now no need to use one of the many kludge 
conventions that exist for that language when presenting contents in 
Berber (notably for those pages that are coded with a mix of Latin 
letters, Greek letters, or symbols like dollar and asterisk...
I am appalled. I thought you understood something about Unicode, Philippe.
--
Michael Everson * * Everson Typography *  * http://www.evertype.com

Re: Berber/Tifinagh (was: Swahili Banthu)

[Don Osborn]

Philippe, I thought I understood the intent of your first letter, but now
I'm not sure.  So let me back up and go over some basics as I understand
them:

1) The Berber languages as we know are written with three scripts, Tifinagh,
Arabic, and Latin.  I've been given to understand that the characters
available in the Latin and Arabic ranges cover all uses for these languages
(even alternatives in Latin transcription, see #7 below).
2) I'm not sure why one would want to encode Tifinagh as anything else,
since one has the option of using the other scripts as they are.  Tifinagh
is challenging enough as is, given its variations, history, and current
uses.  I doubt, though, that it is significantly different in difficulty
from other old scripts, or that any benefit would accrue from linking
Tifinagh encoding to non-Tifinagh scripts.
3) There may be a need to harmonize transcriptions within each script among
the different countries.  Such was the approach in sub-Saharan Africa
several decades ago re transcription of numerous cross-border languages.
This is not an encoding issue, but might be helpful as a language policy
one.
4) There is also a point in working out correspondences among the three
scripts at some point, but it's not as far as I understand an encoding issue
either (concurring here with Michael) - it could be something for scholars,
perhaps at the same time harmonization of transcriptions is considered.  In
fact, I understand that an agency in Morocco concerned with Tifinagh
recently worked up a chart of equivalent characters between Arabic and
Tifinagh alpabets.  In the longer run, having clear rules for alternating
among transcriptions would seem to be helpful (for routines that could
transliterate among the transcriptions).
5) There was a considerable discussion on Unicode-Afrique in Sept.-Oct. on
various topics related to Berber languages and transcriptions. Among these I
would mention a font, ZU-fonts (Unicode for Tamazight)
http://site.voila.fr/aghlan/ZUFonts.htm which is Latin.
6) The latter opted for a sigma - and corrected that from a Greek to a Latin
sigma.
7) The dot under issue is one similar to what is encountered in Igbo and
Yoruba (the latter alternatively a small vertical line under).  The main
issue here is standardization.  And to that end could be something
undertaken at the same time as nos. 34 above?

Hope this helps...

Don

- Original Message - 
From: Philippe Verdy [EMAIL PROTECTED]
To: Michael Everson [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Sent: Sunday, November 09, 2003 5:54 PM
Subject: Re: Berber/Tifinagh (was: Swahili  Banthu)


 From: Michael Everson [EMAIL PROTECTED]

  When we encode Tifinagh we will encode Tifinagh. We will not
  meta-encode it for ease of transliteration to other scripts.

 Yes that was the intent of my suggestion, I don't say that this must be
 done. But what would be wrong if a font was created for the Tifinagh
script
 that would display Latin-based glyphs with diacritics rather than historic
 glyphs?

 Could it help to allow exchanges between use of the historic script in
 Morocco, and those that criticize Morocco and now want to support their
 language with Latin-based glyphs?

 Of course this would exhibit differences of orthograph in the visual
 rendering between a text written in Tifinagh then displayed with Latin
 glyphs, and the same text written normally with the Latin script.

 But it could help exhibit these cultural differences, as well as it could
 help studying the historic texts by people which are not used to read the
 Tifinagh glyphs, and would greatly appreciate to be able to interpret them
 with Latin glyphs but with their native orthograph (later they could laern
 to read the historic glyphs of this script with exactly the same texts).

 This is not really a proposal for a transliteration, as it does not want
to
 map Tifinagh letters with their possible Latin equivalents, taking into
 account the history of this transliteration, which may have modified the
 orthograph to be easier to read in the Latin script (for example if
letters
 have then been added or changed in the Latin script for the modern form of
 the language).

 This problem is similar between Latin and Cyrillic version of
 Serbo-Croatian, which is still considered as a single language (even if
 there are lexical preferences now between Serbian and Croatian, only
Serbian
 being written today with both the Latin and Cyrillic alphabet).

 I'm sorry if this seems stupid for you. The intent is not to deprecate the
 historic script by writing it systematically (with a bijective mapping)
with
 Latin glyphs: such system will look possibly ugly in some cases for users
 which have learned to write their language with the Latin script which
 includes now its own orthograph.

Re: Berber/Tifinagh (was: Swahili Banthu)

[Philippe Verdy]

From: Michael Everson [EMAIL PROTECTED]
 At 19:30 +0100 2003-11-09, Philippe Verdy wrote:

 So my question is, once again: would a font that would display pointed
Latin
 glyphs from Tifinagh script code points really break the Unicode model?

 Yes, Philippe. It is the same thing as mapping Cyrillic to ASCII
 letters. It is a hack. It is to be avoided. It is the Wrong Thing To
 Do.

 If not, then we have a convenient way to define Tifinagh keyboards /
 input methods based on this _apparent_ transliteration.

 This has nothing to do with encoding. You are harkening back to the
 hideous world of 8-bit font hacks of twenty years ago.

Hey! I did not say that the Tifinagh script should be encoded with the same
code points as the Latin script. We clearly need a separate encoding of the
Tifinagh script. I don't want to reuse the same code points (so this is
completely opposite to the old 8-bit hack where the same code positions were
used with distinct meanings).

In fact that's exactly the opposite which may be possible: the Tifinagh code
point could have two graphical representations: the classical one with its
exclusive glyphs, or the Latin one. Such tolerance is not allowed when
representing code points assigned to Latin letters, so Berber texts that are
already coded with Latin code points (using many more or less successful but
incompatible conventions) will need an explicit transliterator to convert
them in the new common script...

But the new script would gain an immediate representation with existing
Latin fonts (for example a Times New Roman or Arial font, updated to support
the Tifinagh code points), until new fonts that support the classical glyphs
become available to users. Of course, to not break this evolution, there's a
need of some consensus about which alternate Latin (or Greek) glyph can
safely represent the Tifinagh code point. Then it's up to the user of that
script to select which font best represents the script and it becomes a
matter of style...

Re: Berber/Tifinagh (was: Swahili Banthu)

[Doug Ewell]

Michael Everson everson at evertype dot com wrote:

 This has nothing to do with encoding. You are harkening back to the
 hideous world of 8-bit font hacks of twenty years ago.

and Philippe Verdy verdy underscore p at wanadoo dot fr responded:

 In fact that's exactly the opposite which may be possible: the
 Tifinagh code point could have two graphical representations: the
 classical one with its exclusive glyphs, or the Latin one. Such
 tolerance is not allowed when representing code points assigned to
 Latin letters, so Berber texts that are already coded with Latin code
 points (using many more or less successful but incompatible
 conventions) will need an explicit transliterator to convert them in
 the new common script...

Such tolerance would not be allowed for Tifinagh code points any more
than it is for Latin code points.  Michael is right: the character/glyph
model says that you are not supposed to render characters in such a way
as to alter their basic identity.  You can render LATIN SMALL LETTER A
using the one-loop (Comic Sans MS) or two-loop (Times New Roman) glyph,
as a wedding-script or Fraktur, or Stterlin letter, but you cannot
render it as a b or an Arabic or Tifinagh letter, or you have
essentially lied about what character it really is.

Likewise, let's suppose for illustration that TIFINAGH LETTER YEB ends
up being encoded at U+08A1, which seems reasonable given the current
roadmap and discussion document.  This letter apparently has at least
three reasonable glyph variants, which look roughly like U+25EB, U+29B6,
and U+2296 respectively.  (Not that I am suggesting unifying this letter
with any of these shapes!)  If we assume that all of these glyphs turn
out to be acceptable for TIFINAGH LETTER YEB, then any of them (or all,
if the OpenType table has such a mechanism) could be used in a font and
mapped to U+08A1, because they would all reflect the basic identity of
the character.  But it would not be appropriate to render Tifinagh yeb
as a Latin b or a Greek beta or a Cyrillic be, because these are four
different letters, and to render them otherwise is to lie about
character identities.

 But the new script would gain an immediate representation with
 existing Latin fonts (for example a Times New Roman or Arial font,
 updated to support the Tifinagh code points), until new fonts that
 support the classical glyphs become available to users. Of course, to
 not break this evolution, there's a need of some consensus about which
 alternate Latin (or Greek) glyph can safely represent the Tifinagh
 code point. Then it's up to the user of that script to select which
 font best represents the script and it becomes a matter of style...

Again, Michael is right, and so is John Hudson.  If you want to
transcribe between Tifinagh and Latin representations of the same text,
the font is not the place to do it.  You don't have to wait for Tifinagh
to be encoded to try this sort of experiment; you could do it now with
(e.g.) Katakana, mapping U+30CE to both the correct Katakana glyph and
the Latin letters no.  But that's all it should be, an experiment, not
a standard practice.  Create two Web pages for Berber text, one in Tfng
and one in Latn, and give your users a choice that way.

-Doug Ewell
 Fullerton, California
 http://users.adelphia.net/~dewell/