Actually, it's worse than that. The entity set labelled "ISO" (which is 
the one I was proposing to retire) actually gives you declarations like this

<!ENTITY rceil  SDATA "[rceil ]"--/rceil C: right ceiling-->

I refuse to believe that these are any use to anyone, other than those 
maintaining legacy SGML systems.

I also find it hard to believe that retiring/updating the other set 
would cause many problems (even for died in the wool P4 addicts like 
Syd) ==  but the purpose of my email was to see if anyone could make a 
plausible case for retaining it.


David Carlisle wrote:
> On 12 April 2010 10:22, Lou <[log in to unmask]> wrote:
> 
>> Leaving aside the vagueness as to what "the TEI web sites" means -- there
>> are indeed sets of declarations, apparently derived from the same sources as
>> those mentioned by David, via Docbook v4 , at
>> http://www.tei-c.org/Entity_Sets/Unicode/ which give the Unicode mappings,
> 
> I'd agree with Syd Bauman that moving or removing entity files is
> always problematic and best avoided, however I just had a look at the
> ones cited and for example it has
> 
> <!ENTITY scnE	"& #xE2B5;"> <!--  -->
> 
> using private use area code points (from the original STIX submission)
> that sort of made sense at the time, as Unicode didn't have any
> suitable characters but this codepoint is unlikely to have the desired
> effect  in current software, and is explicitly non-portable.
> 
> The W3C version of isoamsn.ent has
> 
> <!ENTITY scnE             "& #x02AB6;" ><!--SUCCEEDS ABOVE NOT EQUAL TO -->
> 
> using a character added at Unicode 3.2 which does appear at that
> location in unicode math fonts such as stix beta or cambria math.
> 
> Thus while you probably want to keep the entity files at their
> existing location, a warning that updated ones are somewhere else
> might be in order.
> 
> David
>