Four experiments [was Re: [Aleph] feature request :-))]

[ishamid]

Hi Giuseppe and our fellow conspirators,

>===== Original Message From Discussions concerning Aleph <aleph at ntg.nl> =====

>The problem, and the *proper* way to do it, is to set up
>different priorities for the various OCPs! The magic number
>after the \add<...>ocplist *should* be different. If you give
>priority 1 to the first, priority 2 to the second and priority
>3 to the last it *will* work, *regardless of the order in which
>you push the ocplists!

Ok, we r going to need Hans' help here:

\definefiltersequence[][]

now needs keys to determine the relative priority of the \add<...>ocplist 
lines, e.g., something like

\definefiltersequence
  [Sequence]
  [SegmentOne,SegmentTwo,Segment3]
  [1,2,3]

for

\ocplist\Sequence=
\addbeforeocplist 1 SegmentOne
\addbeforeocplist 2 SegmentOne
\addbeforeocplist 3 SegmentOne
\nullocplist

Now what about more complicated cases, where there are meta-sequences?

\def\MetaSequence{%
\clearocplists
\usefiltersequence[SequenceOne]
\usefiltersequence[SequenceTwo]
}

Does SequenceTwo need to use, say, priority 4, 5, and 6, or can it use 1, 2, 
and 3 like SequenceOne? This can get really messy for long concatenations of 
Sequences!

I agree with an earlier comment you made that an abstraction list based on 
complete \ocplist segments is more elegant than one based on 
\addbeforeocplist <#> \<foo>
segments. On the other hand, in the complete \ocplist segments approach we 
need a manageable way to handle all the relative priorities of the \ocplist's 
so that we can construct working meta-sequences.

So here is the ideal abstraction process (using ConTeXt/Gamma):

%%% low-level filter definitions

\definefiltersynonym[Name1][OCPname1]
\definefiltersynonym[Name2][OCPname2]
:

%%% ocplist segments, each for a specific task

\definefiltersequence
[SequenceOne]
[Name1,Name2]
[1,2]

\definefiltersequence
[SequenceTwo]
[Name3,Name4]
[3,4] % or [1,2]???

:

%%% next abstraction layer

\def\MyTaskOne{\usefiltersequence[SequenceOne]}
:

%% For example, in the middle of a 
%% document the user may want to redefine
%% \MyTaskOne.

%% Hans, can we have a more user friendly macro for
%% this abstraction layer?

%%% Meta-sequences %% gives everthing needed per language
%% Hans, can we have better macros than this?

\def\MetaSequenceTwo{%
\clearocplists
\MyTaskOne
\usefiltersequence[SequenceTwo]
\MyTaskThree
\MyTaskFour
:
}

Note that each \MetaSequence<> contains a \usefiltersequence[] line. This is 
the catalyst ocplist that determines the primary function of the 
\MetaSequence<>. So \MetaSequenceTwo carries all the information needed for a 
given language from whatever is predefined in the document; the only thing it 
adds is a specific task (\usefiltersequence[SequenceTwo]) that is now 
independent of all other tasks.

So if \usefiltersequence[SequenceTwo] stands for, say, full vocalization, then 
\MetaSequenceTwo will always carry out full vocalization, no matter how the 
other \MyTask<>'s are defined elsewhere in other \MetaSequence<>'s.

A similar procedure can be constructed for the \addbeforeocplist-segment 
approach. But I hope we can get the more elegant approach working.

Thnx so much for your clarifications, Giuseppe. I'm going to integrate this 
into my bigger abstraction project for Arabic and report back, probably later 
this week. And I welcome any other comments!

Thnx again
Idris

>Whaddyathinkof this?

Just smashing;-)

>
>Working example:
>
>===ocp-work===
>\ocptracelevel=1
>
>% greek ocps
>\ocp\GrTexUni=grpo2uni \ocp\GrUniToFont=uni2greek
>
>%arabic ocps
>\ocp\ArabUni=7arb2uni
>\ocp\UniCUni=uni2cuni
>\ocp\CUniArab=cuni2oar
>
>% abstractions for greek
>\def\GreekTranscription{\addbeforeocplist 1 \GrTexUni}
>\def\GreekFont{\addbeforeocplist 2 \GrUniToFont}
>\def\GreekFull{%
>                \addbeforeocplist 1 \GrTexUni
>                \addbeforeocplist 2 \GrUniToFont
>                }
>
>% default
>\ocplist\GreekOcp=
>                \addbeforeocplist 1 \GrTexUni
>                \addbeforeocplist 2 \GrUniToFont
>                \nullocplist
>
>% abstraction xperiment 1 % expands the `1'
>\expandafter\expandafter\expandafter\ocplist%
>\expandafter\expandafter\expandafter\MyGreekOcpOne%
>\expandafter\expandafter\expandafter=%
>\expandafter\GreekTranscription%
>\GreekFont
>\nullocplist
>
>% abstraction xperiment 2
>\expandafter\ocplist\expandafter\MyGreekOcpTwo\expandafter=%
>\GreekFull
>\nullocplist
>
>% abstraction xperiment 3
>
>\ocplist\GreekOcpSegmentOne=
>                \addbeforeocplist 1 \GrTexUni
>                \nullocplist
>
>\ocplist\GreekOcpSegmentTwo=
>                \addbeforeocplist 2 \GrUniToFont
>                \nullocplist
>
>\font\greek=omlgc at 12pt
>\greek
>
>%default
>\pushocplist\GreekOcp
>
>a b c d
>
>% xperiment 1
>\clearocplists
>\pushocplist\MyGreekOcpOne
>
>1 a b c d
>
>% xperiment 2
>\clearocplists
>\pushocplist\MyGreekOcpTwo
>
>2 a b c d
>
>% xperiment 3
>\clearocplists
>\pushocplist\GreekOcpSegmentOne
>\pushocplist\GreekOcpSegmentTwo
>
>3 a b c d
>
>% abstraction xperiment for arabic
>
>% default
>\ocplist\ArabicOCP=
>\addbeforeocplist 1 \ArabUni
>\addbeforeocplist 2 \UniCUni
>\addbeforeocplist 3 \CUniArab
>\nullocplist
>
>% abstraction xperiment 4
>\ocplist\ArabicOcpSegmentOne=
>                \addbeforeocplist 1 \ArabUni
>                \nullocplist
>\ocplist\ArabicOcpSegmentTwo=
>                \addbeforeocplist 2 \UniCUni
>                \nullocplist
>\ocplist\ArabicOcpSegmentThree=
>                \addbeforeocplist 3 \CUniArab
>                \nullocplist
>\font\arabic=omarab at 12pt
>\arabic
>\textdir TRT\pardir TRT
>
>% default
>
>\clearocplists
>\pushocplist\ArabicOCP
>
>Aal-Hamdu Aal-Hamdu Aal-Hamdu
>
>\bigskip
>
>% xperiment 4
>\clearocplists
>\pushocplist\ArabicOcpSegmentOne
>\pushocplist\ArabicOcpSegmentTwo
>\pushocplist\ArabicOcpSegmentThree
>
>4 Aal-Hamdu Aal-Hamdu Aal-Hamdu
>
>\bye
>====