John H. Jenkins wrote:
>> The Symbol Encoding subsection (U+F000-U+F0FF) is reserved for
>> symbol/dingbat (pictoral)-type fonts (fonts consisting mainly of
>> simplified pictures).
>
> This is blatantly untrue. There is *no* "subsection" of the PUA.
> By its very nature, it's a free-for-all.
Unicode does not define a Symbol Encoding subsection -- that is a Microsoft convention, as noted -- but it certainly does describe "subareas" of the PUA, as explained in TUS 3.0, Section 13.5, "Private Use Area" (p. 323):
"Encoding Structure. By convention, the Private Use Area is divided into a Corporate Use subarea, starting at U+F8FF and extending downward in values, and an End User subarea, starting at U+E000 and extending upward."
The remaining paragraphs describe these two subareas in detail. There is a comment that "This convention is for the convenience of system vendors and software developers." This is also where you find the reference to the lack of a mechanism to avoid a "stack-heap collision" between the two areas (a reference that I understood perfectly, but which I've always thought might be totally lost on a non-programmer).
The paragraph on promoting PUA characters to full Unicode status makes a strange reference to these candidate characters being encoded by vendors in the Corporate Use subarea. I had never noticed this before. It seems inappropriate to reserve this scenario exclusively for characters defined by vendors, or in the Corporate Use subarea. In fact, Deseret and Shavian (the latter proposed for 4.0) were both originally encoded in the ConScript Unicode Registry, perhaps the best-known instance of scripts being promoted from the PUA; but they were *not* encoded by a vendor, and *not* in the Corporate Use subarea. Is there any reason for this passage to survive the 4.0 revision?
-Doug Ewell
Fullerton, California
Amateur Unicode Exegete
This archive was generated by hypermail 2.1.2 : Sat Apr 13 2002 - 15:25:32 EDT