>David Goldsmith asked...
>> I mean, how fast does Klingon have to draw?
>** KLINGON SHMINGON... THEY'RE A PACK OF WEAK-BRAINED TRIBBLE-KISSING
> I WILL ACCEPT ANY PROPOSAL ORIGINATING WITH THE KLINGON HIGH COMMAND,
>PRECISELY WHEN THEY BEAM AN ARMED DELEGATION INTO THE UTC MEETING ROOM AND
>DEMAND THAT THEIR SCRIPT BE ENCODED. UNTIL SUCH TIME, I THUMB MY NOSE AT
>SNIVELING KLINGON EMPIRE. And you may quote me on that!
>There are ACTUALLY more important things like Egyptian hieroglyphics, a few
>"recent" historical scripts, and other large historical scripts proposed
>Plane 1. It will not be "outer space" forever.
I'm willing to cobble together Deseret Alphabet support using one of the
user planes and actually do a performance measurement to see what the
cost would be. (The advantage being that I have a buncha data in the
Deseret Alphabet and so can get reliable numbers.)
Dave Opstad did a back-of-the-envelope calculation, however. He
estimates that implementing surrogates as ligatures will slow down layout
by about 1%.
One thing that I think makes a difference is that Apple's layout engine
assumes that ligature/conjunct formation *will* be taking place (even for
Latin text) and is therefore designed to make it efficient.
John H. Jenkins
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:40 EDT