[HarfBuzz] New Indic standard?

[Ed Trager]

On Fri, Aug 21, 2009 at 11:45 AM, Shriramana Sharma<samjnaa at gmail.com> wrote:

>
> The gold standard would be the judgment of native users, not some other
> rendering engine. Uniscribe is OK for the most part, but where you doubt,
> ask the friendly natives. I think there are Indians here who can cover most
> if not all of the Indic scripts. I am available for Devanagari (as used to
> write Sanskrit and Hindi) and Tamil (and Kannada and Telugu, if no native
> Kannadiga or Teluguvallu is here).
>

OK, so to summarize based on Behdad's and Shriramana's responses:

(1) We need to compile the INPUT data : ideally a nicely organized and
comprehensive set covering all the possible ligatures, conjuncts, etc.
for the major Indic scripts.  That's "all" that Behdad needs.

(2) We need to create a set of graphic images showing the *ideal*
desired output.  I suppose the following sort of workflow might be
possible:

       (2.1). For the "NON-NATIVES" (like myself): compare PANGO vs.
UNISCRIBE (maybe also vs. ATSUI) graphic renderings of the Unicode
strings in the test data set:  Visual discrepancies form a "short
list" of problematic entries that need to be passed on to the
"NATIVES" for adjudication.

       (2.2). The "NATIVES" like Shriramana will for the most part be
able to just "eyeball" problematic or questionable forms.

       (2.3) Hopefully some of the "NATIVES" are also picky
typographers who can also point out cases that are technically
"correct" but which suffer from poor or less-than-ideal placement of
reph, diacritic vowels, poorly-joined "half" forms, etc.

(3) Behdad will be motivated to add the necessary modifications to
PangoView (to spit out the glyph IDs and positioning data that is
needed for automated testing of the OUTPUT) once some intial
compilation of data makes it worthwhile to do the work.

(4) It would also be nice to eventually include in the test data cases
covering the issues mentioned by Peter Constable in the Unicode Public
Review Issues #37 "Proposal on Clarification and Consolidation of the
Function
of ZERO WIDTH JOINER in Indic Scripts".  Does anyone know if
Constable's proposal is now accepted as part of the standard?

- Ed