Re: Multiple private agreements (was: RE: Code pages and Unicode)

From: Philippe Verdy <>
Date: Wed, 24 Aug 2011 07:34:05 +0200

2011/8/24 Luke-Jr <>:
> On Tuesday, August 23, 2011 10:29:58 PM Philippe Verdy wrote:
>> 2011/8/24 Doug Ewell <>:
>> > (3) which contains the same PUA code point with two meanings
>> The only numbered item to sacifice is number (3) here. that's the case
>> where separate PUA agreements are still coordinated so that they don't
>> use the same PUA assignments. This is the case of PUA greements in the
>> Conscript registry.
> Too bad the Conscript registry is censoring assignments the maintainer doesn't
> like for unspecified personal reasons, increasing the chances of an overlap.

It's their choice, their private decision. Nobody is required to
accept the conditions of CSUR. In fact other groups could be created
to coordinate other choices compatible with each other.

Even the UTC could create its own PUA registry, probably coordinating
it with WG2, and with the IRG, for experimenting new encodings, or
working on proposals, helping document the needed features or
difficulties, and cooperate better with non-technical people that have
good cultural knowledge, or that have access to rare texts or corpus
for which there still does not exist any numerisation (scans), or
whose numerisation is restricted or not financed, and for which it is
also impossible to create OCR versions.

In order to get financements, some of those projects would need to
exhibit only some fragments, explaining what is found in the rest of
the corpus, using significant samples, but also new creating didactic
documents, for which PUAs will be needed if they want to interchange
with something else than handwritten papers, and photocopies or scans
(which are not easy to handle via emails or in HTML pages, or that are
to reproduce).

Such PUA registry is not required to be stable for extensive periods.
Its content will evolve so that the encoded documents will be valid
for a limited time. This also means that the necessary fonts required
to keep those texts in a legible way (and possible future reencoding,
to new PUAs or to standard assignments in the UCS) would have to be
kept with those PUA texts. Those fonts should be clearly versioned,
containing an expected lifetime for which the PUA registry may
warranty some stability (example: the PUA registry will make
assignments only by early leases that will need to be renewed by
interested people).

Note that I clearly want that PUA fonts contain explicitly the
character properties needed for proper rendering. Simply because it is
expected that PUA documents will be created and interchanged for a
limited time. There will be almost no transforms of those texts, only
updates to their content via editing.

Now which font format will be the best suited for this work with PUA
texts? May be OpenType is not the best fit (tools to create them are
too complex for most users, and often are too costly, probably a
consequence of this complexity that destinate those tools only to very
few specialists), when there are simpler formats that are easily
editable from more tools (SVG fonts look promising, even if their
typographic capabilities are not very advanced for now; I just hope
that someday there will be support for this format in more renderers,
even if those fonts are larger in size for less glyphs inside; but
this SVG format can be easily zipped into a SVGZ format also
recognized automatically).

But some OSes or applications are offering simple accesory tools to
create PUA glyphs stored in personal fonts that can be reedited,
embedded, or uploaded to the recipients of a document needing these
glyphs. This may be used as an extension to input method editors,
notably for entering custom sinograms). Those tools won't let you
create glyphs with perfect metrics, or fonts with ligatures/GSUB
features, or advanced GPOS'itioning. Drawing tools are minimized to
reproduce how we draw basic shapes with the circle head of a pen, of
the elliptic head of a pencil, or the thin linear head of some
highliting pens. Some other tools just let you use a scan and produce
basic shapes.
Received on Wed Aug 24 2011 - 00:39:59 CDT

This archive was generated by hypermail 2.2.0 : Wed Aug 24 2011 - 00:40:02 CDT