Observations on Unicode, RunRev and Operating Systems

Richmond richmondmathewson at gmail.com
Fri Sep 3 16:10:09 EDT 2010


[This message is neither funny nor abusive - sorry to disappoint some 
people.]

Over the last few days something faintly ugly has come to light
that probably won't concern many RunRev developers, but might
affect anybody planning to develop multi-platform stuff with Unicode.

If one opens a Unicode font with a font development program what the
user sees are lots and lots of glyphs; what most people don't see are 
all sorts of
rules as to how they should behave when the end-user types something 
using that
font, possibly also using a text-encoding algorithm built into their 
operating system.

Why should we care?

Because, while Windows Vista and '7', and Linux works wonderfully with 
Unicode
fonts giving those rules cognisance, Mac OS and Windows XP don't . . .

------------------------------------------------------------------------------------------------------------------

Case study "Richmond thinks he knows Sanskrit".

When Sanskrit is written using the Devanagari writing system (it can be 
written
using other scripts) there is a "funny situation" where a short 'i' that 
is sounded
after a consonant is actually written prior to the consonant.

So, I learnt to write Sanskrit using Devanagari about 27 years ago with 
a calligraphic pen,
when I had not even heard of the second ASCII table, and my programming 
stretched as far
as DBAS9 on a Research Machine 280Z, and Fortran 4 on Hollerith cards 
popped in the
post.

So, obviously, I wrote my short 'i' prior to the consonant after which 
it was to be sounded
[wow, that stretched my grammar].

So, incredibly naively, I assumed that that is the order a computer 
program [erm . . . my 'Devawriter']
would enter the short 'i' and then whatever consonant: and, 'wonder of 
wonders', that worked a treat on
several of my Macs, and on a Windows XP test drive.

So, when I got my paws on the new DP of RunRev 4.5 I thought that it was 
playing "silly bu**ers"
by requiring me to input the short 'i' AFTER the consonant, but, in 
fact, this is exactly how the
rules are set up in the font.

---------------------------------------------------------------------------------------------------------------------

SO:

Macintosh and Windows XP do not "play fair" with most Unicode fonts.

--------------------------------------------------------------------------------------------------------------------

This means that types like myself, getting far too clever for their own 
good or sanity, have
to set up 2 rather different sets of algorithms for "eccentric" writing 
rules if they want to
produce standalones that fool around with Unicode fonts cross-platform.

--------------------------------------------------------------------------------------------------------------------

I should point out that this is a problem concerning the underlying 
operating systems
and nothing at all to do with RunTime Revolution beyond the fact that it 
is unable
(unsurprisingly) to over-ride certain aspects of how an operating system 
behaves.

sincerely, Richmond Mathewson.



More information about the use-livecode mailing list