[Aleph] please review

Marc klists at saphor.de
Sat Jun 2 18:17:47 CEST 2007

Benjamin BAYART wrote:
> Le Thu, May 31, 2007 at 08:25:02PM -0600, Idris Samawi Hamid:
>> Till then, let's use the aleph list to collect wish-list items, bugs with
>> aleph, etc. The developers do read this list so...
> As you perhaps know, I'm a long time user of Omega (acheology would
> probably find references back to the real beginnings of Omega, when it
> did not even run on Intel architectures).
> My recent concerns about omega/aleph are the OTPs. There is a real
> problem in that point which is not an implementation concern, but a
> design problem. It seems luaTeX (tries to) address the problem and I'm
> quite impatient to have à look at it. Mostly, in my mind, those
> questions are opened:
> - will the translation happen on source (like preprocessor) or on nodes
>   (like paragraph transformer), the main mistake of OTP was to try to do
>   both ?
> - how will that interact with opentype features ?
>   o possible to load a "normal" font, and then add it (via macros?)
>     features like the ones expect in opentype?
>   o possible to change the features after loading an opentype fond?
>     (like one can change a \fontdimen after loading a font)
> - there is a dual syntax, one from eTeX for handling high-number
>   registers and another from Omega, with behaviours hard to understand
>   and with different limits, will that be unified (I guess so) ?
>   o a brand new syntax deprecating the two previous ones ?
>   o another kludge to try to have averything backward compatible?
>   o new features enabled and that's all ?
> That project sounds like good news. I should go more often to TeX
> meetings, I juste discovered the existance of luaTeX while people
> seems to be informed :)
I got the same surprise --- the roadmap would indeed address a number of
key issues ahead, especially a better handling of truetype / open type
fonts (their handling with Aleph is a pure pain --- rastering them with
ttf2afm and all the stuff. font_install_unicode.pl soothes some of the
pain, but even then)

As promised, I'd like to add a few issues that I experienced when
working with Aleph and related tools (I'm writing this largely out of
the top of my head, so the list is likely incomplete):

 * a solution to recurrent catcode issues, making the use of the
verbatim environment essentially impossible
 * the intransparency of OTPs (I'd like to see a true preprocessor which
allows to inspect the results of a given action, or possibly nothing at
all with the actual preprocessing being the user's responsibility.
Virtual fonts make up for a good part of what would be lost)
 * rare instabilities of Aleph
 * sometimes difficult to understand right to left typesetting semantics
 * numerous errors in existing OTPs, e.g. in the Greek part of omlgc and
family (inherited from Omega, but extremely dangerous, as, of course, no
warning informs about the wrong character being printed)
 * almost unusable ovp2ofm (extremely unstable, often reacts to small,
syntactically correct changes in the virtual font with core dump)
 * problems with ttf2afm (based on a very old version of free-type,
difficult to even install, but hopefully soon superfluous)

Especially the poor quality of many related, but essential tools such as
ovp2ofm made the whole system often brittle.

On the positive side, the virtual font support of Aleph itself is a key

As to my specific requirements, I've used the following scripts:

 * Latin, including phonetics and many specific transcriptions e.g. for
Old Near Eastern scripts
 * Greek, again including many unusual combinations of Ancient Greek
with diacritics (unsure epigraphic reading of characters etc.)
 * Hebrew including vocalization
 * Ethiopic
 * Cyrillic
 * Egyptian hieroglyphs

as well as a couple of words in Arabic and Chinese scripts and a few
dozen glyphs from a ttf font of my own making (some hieroglyphs, some
cuneiform glyphs,  etc.).

(I've sometimes considered drafting article on the production cycle
starting with TEI, but couldn't convince myself that somebody would be
interested in this)

As to samples: The publisher Niemeyer may post something on the search
inside feature on Amazon
but that isn't decided yet. Alternatively, I could send a specimen copy
to the team also as a token of gratitude towards Aleph.

Best regards,


More information about the Aleph mailing list