Proposed change to greek-ibycus4 input-method

Proposed change to greek-ibycus4 input-method

[Peter Heslin]

[-- Attachment #1: Type: text/plain, Size: 1417 bytes --]

I would like to propose a change in the way the greek-ibycus4
input-method for ancient Greek handles capital letters with iota
subscript (ypogegrammeni), as a result of having been bitten by its (to
me) very surprising behavior.

Currently, when you type a capital letter followed by a normal iota, the
input method arbitrarily decides that this must actually be a subscript
iota.  But in the vast majority of cases, this is not what the user
intends at all -- he or she wants a normal iota after the capital
letter.

The reason for this ambiguity is that the ibycus4 encoding for LaTeX
does not actually support subscript iotas under capital letters (it
expects you to write them adscript, as if they were normal iotas).  So
there is no pre-existing standard to appeal to, but it seems logical to
use the | character after the vowel, just as for lower case vowels.

In other words, with the current code:

    )Ai => ᾈ 

which has two problems: (1) it is very surprising, and (2) there is no
straight-forward way to type the common sequence of Greek characters Ἀι
‐‐ you have to separate the vowels with a space and then go back and
delete the space between them.

With my proposal:

    )A| => ᾈ 
    )Ai => Ἀι

Now there is an easy way to type both sequences and the default behavior
is much less surprising.  It's also consistent with the behavior of the
greek-babel input-method.  Patch attached.


[-- Attachment #2: greek.diff --]
[-- Type: text/plain, Size: 1383 bytes --]

--- greek.el.orig	2006-07-11 21:20:34.000000000 +0100
+++ greek.el	2006-07-11 21:43:19.000000000 +0100
@@ -794,8 +794,8 @@
 
 (quail-define-rules
 
- ("{((}" ?\() ("((" ?\() ; #x0028
- ("{))}" ?\)) ("))" ?\)) ; #x0029
+ ("{((}" ?\() ("((" ?\() ("{(}" ?\() ; #x0028
+ ("{))}" ?\)) ("))" ?\)) ("{)}" ?\)) ; #x0029
  ("<<" ?«) ; #x00ab
  (">>" ?») ; #x00bb
 
@@ -1016,14 +1016,14 @@
  ("a)=|" ?ᾆ)
  ("a(=|" ?ᾇ)
 
- (")Ai" ?ᾈ)
- ("(Ai" ?ᾉ)
- (")`Ai" ?ᾊ)
- ("(`Ai" ?ᾋ)
- (")'Ai" ?ᾌ)
- ("('Ai" ?ᾍ)
- (")=Ai" ?ᾎ)
- ("(=Ai" ?ᾏ)
+ (")A|" ?ᾈ)
+ ("(A|" ?ᾉ)
+ (")`A|" ?ᾊ)
+ ("(`A|" ?ᾋ)
+ (")'A|" ?ᾌ)
+ ("('A|" ?ᾍ)
+ (")=A|" ?ᾎ)
+ ("(=A|" ?ᾏ)
 
  ("h)|" ?ᾐ)
  ("h(|" ?ᾑ)
@@ -1034,14 +1034,14 @@
  ("h)=|" ?ᾖ)
  ("h(=|" ?ᾗ)
 
- (")Hi" ?ᾘ)
- ("(Hi" ?ᾙ)
- (")`Hi" ?ᾚ)
- ("(`Hi" ?ᾛ)
- (")'Hi" ?ᾜ)
- ("('Hi" ?ᾝ)
- (")=Hi" ?ᾞ)
- ("(=Hi" ?ᾟ)
+ (")H|" ?ᾘ)
+ ("(H|" ?ᾙ)
+ (")`H|" ?ᾚ)
+ ("(`H|" ?ᾛ)
+ (")'H|" ?ᾜ)
+ ("('H|" ?ᾝ)
+ (")=H|" ?ᾞ)
+ ("(=H|" ?ᾟ)
 
  ("w)|" ?ᾠ)
  ("w(|" ?ᾡ)
@@ -1052,14 +1052,14 @@
  ("w)=|" ?ᾦ)
  ("w(=|" ?ᾧ)
 
- (")Wi" ?ᾨ)
- ("(Wi" ?ᾩ)
- (")`Wi" ?ᾪ)
- ("(`Wi" ?ᾫ)
- (")'Wi" ?ᾬ)
- ("('Wi" ?ᾭ)
- (")=Wi" ?ᾮ)
- ("(=Wi" ?ᾯ)
+ (")W|" ?ᾨ)
+ ("(W|" ?ᾩ)
+ (")`W|" ?ᾪ)
+ ("(`W|" ?ᾫ)
+ (")'W|" ?ᾬ)
+ ("('W|" ?ᾭ)
+ (")=W|" ?ᾮ)
+ ("(=W|" ?ᾯ)
 
  ("a`|" ?ᾲ)
  ("a|" ?ᾳ)

[-- Attachment #3: Type: text/plain, Size: 53 bytes --]


-- 
Peter Heslin (http://www.dur.ac.uk/p.j.heslin)


[-- Attachment #4: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: Proposed change to greek-ibycus4 input-method

[David Kastrup]

Peter Heslin <pj@heslin.eclipse.co.uk> writes:

> I would like to propose a change in the way the greek-ibycus4
> input-method for ancient Greek handles capital letters with iota
> subscript (ypogegrammeni), as a result of having been bitten by its (to
> me) very surprising behavior.
>
> Currently, when you type a capital letter followed by a normal iota, the
> input method arbitrarily decides that this must actually be a subscript
> iota.  But in the vast majority of cases, this is not what the user
> intends at all -- he or she wants a normal iota after the capital
> letter.
>
> The reason for this ambiguity is that the ibycus4 encoding for LaTeX
> does not actually support subscript iotas under capital letters (it
> expects you to write them adscript, as if they were normal iotas).  So
> there is no pre-existing standard to appeal to, but it seems logical to
> use the | character after the vowel, just as for lower case vowels.
>
> In other words, with the current code:
>
>     )Ai => ᾈ 
>
> which has two problems: (1) it is very surprising, and (2) there is no
> straight-forward way to type the common sequence of Greek characters
> Ἀι

I think there is a bit of code point vs reprentation problem here,
too.  I think that ᾈ is actually the same `character' as Ἀι or even Αι
(though Αι would also be short for Ἁι, like in ἐν Ἁιδõυ, and so needs
a different code point just in case someone wants to use a font with
spirited capitals), only with a different writing convention.  I would
think it likely that you'll never encounter both ᾈ as well as Ἀι in
the same text, but you might find consistently either one or the other
(or the third) everywhere, depending on the printer's taste.

This is in contrast to Αἰ or Ἀϊ which are actually different character
combinations.

> ‐‐ you have to separate the vowels with a space and then go back and
> delete the space between them.
>
> With my proposal:
>
>     )A| => ᾈ 
>     )Ai => Ἀι
>
> Now there is an easy way to type both sequences and the default behavior
> is much less surprising.  It's also consistent with the behavior of the
> greek-babel input-method.  Patch attached.

I think your idea sounds reasonable.  I just have no clue whether
the Ibycus rules would actually suggest one or the other.

For convenient typing in Emacs, your suggestion is likely the best and
most intuitive way, anyhow.  But if we want to convert "input just
like a quail method" into Unicode, like recently discussed on this
list, it might start making a difference.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: Proposed change to greek-ibycus4 input-method

[David Kastrup]

[-- Attachment #1: Type: text/plain, Size: 257 bytes --]

David Kastrup <dak@gnu.org> writes:

> (though Αι would also be short for Ἁι, like in ἐν Ἁιδõυ, [...]

Cough, cough.  I mean ἐν ᾍδου or ἐν Ἅιδου, of course.

Sorry.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

[-- Attachment #2: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: Proposed change to greek-ibycus4 input-method

[Peter Heslin]

David Kastrup <dak@gnu.org> writes:

> This is in contrast to Αἰ or Ἀϊ which are actually different character
> combinations.

Yes, thank you -- that is exactly what I should have said, for it's the
best argument of all.  In an Ibycus TeX file, Αἰ is written as Ai) but
when you type Ai) using the greek-ibycus4 input-method, what you
currently get is ᾼ᾿ -- which is very wrong.

> I think your idea sounds reasonable.  I just have no clue whether
> the Ibycus rules would actually suggest one or the other.

I think the above example demonstrates that the Ibycus encoding is quite
at odds with the current behavior of the input-method.  The only
question, then, is what to use when one actually does want an iota
subscript with a capital vowel (given that Ibycus itself doesn't support
that).  Using | after the vowel is the obvious suggestion.

> For convenient typing in Emacs, your suggestion is likely the best and
> most intuitive way, anyhow.  But if we want to convert "input just
> like a quail method" into Unicode, like recently discussed on this
> list, it might start making a difference.

In fact, my awareness of this problem arose out of writing and using the
conversion code you are referring to.  This change will not affect the
translation of existing Ibycus text, except to fix the class of bug
described above.

-- 
Peter Heslin (http://www.dur.ac.uk/p.j.heslin)

Re: Proposed change to greek-ibycus4 input-method

[David Kastrup]

Peter Heslin <pj@heslin.eclipse.co.uk> writes:

> David Kastrup <dak@gnu.org> writes:
>
>> This is in contrast to Αἰ or Ἀϊ which are actually different
>> character combinations.
>
> Yes, thank you -- that is exactly what I should have said, for it's
> the best argument of all.  In an Ibycus TeX file, Αἰ is written as
> Ai) but when you type Ai) using the greek-ibycus4 input-method, what
> you currently get is ᾼ᾿ -- which is very wrong.

Agreed.

> In fact, my awareness of this problem arose out of writing and using
> the conversion code you are referring to.  This change will not
> affect the translation of existing Ibycus text, except to fix the
> class of bug described above.

I can't find a copyright assignment on file for you.  This change,
however, is a purely mechanical one without creative content of its
own.  I'll just recreate it manually and you can check afterwards, if
you want to, whether the diffs agree.  I should be surprised if not.
I have no idea whether you expect to contribute to Emacs in future.
If yes, it might be a good idea to ask for a copyright assignment form
for work on Emacs.

At the moment, there are probably not many people with an assignment
on file who would be dabbling with Polytoniko Greek input methods.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: Proposed change to greek-ibycus4 input-method

[David Kastrup]

David Kastrup <dak@gnu.org> writes:

> Peter Heslin <pj@heslin.eclipse.co.uk> writes:
>
>> David Kastrup <dak@gnu.org> writes:
>>
>>> This is in contrast to Αἰ or Ἀϊ which are actually different
>>> character combinations.
>>
>> Yes, thank you -- that is exactly what I should have said, for it's
>> the best argument of all.  In an Ibycus TeX file, Αἰ is written as
>> Ai) but when you type Ai) using the greek-ibycus4 input-method, what
>> you currently get is ᾼ᾿ -- which is very wrong.
>
> Agreed.
>
>> In fact, my awareness of this problem arose out of writing and using
>> the conversion code you are referring to.  This change will not
>> affect the translation of existing Ibycus text, except to fix the
>> class of bug described above.
>
> I can't find a copyright assignment on file for you.  This change,
> however, is a purely mechanical one without creative content of its
> own.  I'll just recreate it manually and you can check afterwards, if
> you want to, whether the diffs agree.  I should be surprised if not.
> I have no idea whether you expect to contribute to Emacs in future.
> If yes, it might be a good idea to ask for a copyright assignment form
> for work on Emacs.
>
> At the moment, there are probably not many people with an assignment
> on file who would be dabbling with Polytoniko Greek input methods.

Done.  I got surprised by the diff being quite larger than expected,
which was due to unify-8859-on-... being active, so I switched those
off, reloaded and reapplied my changes in order to be on the safe
side.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: Proposed change to greek-ibycus4 input-method

[Peter Heslin]

David Kastrup <dak@gnu.org> writes:

> I can't find a copyright assignment on file for you.  This change,
> however, is a purely mechanical one without creative content of its
> own.  I'll just recreate it manually and you can check afterwards, if
> you want to, whether the diffs agree.  I should be surprised if not.

I've looked at the updated version of the file, and you have made the
changes I thought necessary.  Thanks for doing this.

> I have no idea whether you expect to contribute to Emacs in future.
> If yes, it might be a good idea to ask for a copyright assignment form
> for work on Emacs.

As it happens, a year ago rms asked me to adapt one of my elisp packages
for possible inclusion in Emacs, but I never found time to do it.  So it
would make sense for me to sign papers, in the hope that I get around to
working on that project this summer.  I don't know whom to ask for the
papers, though.

-- 
Peter Heslin (http://www.dur.ac.uk/p.j.heslin)

Re: Proposed change to greek-ibycus4 input-method

[David Kastrup]

Peter Heslin <p.j.heslin@dur.ac.uk> writes:

> David Kastrup <dak@gnu.org> writes:
>
>> I can't find a copyright assignment on file for you.  This change,
>> however, is a purely mechanical one without creative content of its
>> own.  I'll just recreate it manually and you can check afterwards, if
>> you want to, whether the diffs agree.  I should be surprised if not.
>
> I've looked at the updated version of the file, and you have made the
> changes I thought necessary.  Thanks for doing this.
>
>> I have no idea whether you expect to contribute to Emacs in future.
>> If yes, it might be a good idea to ask for a copyright assignment form
>> for work on Emacs.
>
> As it happens, a year ago rms asked me to adapt one of my elisp packages
> for possible inclusion in Emacs, but I never found time to do it.  So it
> would make sense for me to sign papers, in the hope that I get around to
> working on that project this summer.  I don't know whom to ask for the
> papers, though.

I'll send you the form off-list.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum