Phil,
> On 26 Mar 2019, at 17:49, Phil Clayton <[email protected]> wrote: > > Rob, > > [A somewhat belated response now. I only noticed the delivery failure from > December 2016 today! Trying again...] Better late than never! > > On 27/11/16 14:21, Rob Arthan wrote: >> Dear All, >> Roger Jones and I are doing some more work on Unicode and UTF-8 support in >> ProofPower. >> We are currently considering two changes to the Unicode mapping as currently >> defined at: >> http://www.lemma-one.com/ProofPower/unicode/pp-unicode.html >> 1) Supporting Unicode/UTF-8 in xpp and the various document-processing >> scripts would be >> much simpler, if each character in the ProofPower extended character set >> mapped to a single >> Unicode code point. Currently the only exception is the symbol for >> distributed concatenation >> which has to be translated into two code points (a frown followed by a >> slash). Since frown-slash >> is already accepted as a synonym for the single character for distributed >> concatenation we >> would like to withdraw the single character mark-up from the ProofPower Z >> library. > > I think this is a change for the better anyway. Thanks for the endorsement. > > >> 2) We would like to use the MathML XML entity set as a standard set of >> human-readable >> names for Unicode code points. However, there a few slight discrepancies >> that need to be resolved: >> At Phil Clayton’s (nice) suggestion, the ProofPower mapping currently maps >> the greek >> letters to the corresponding code points for Mathematical Alphanumeric >> Symbols in the range >> 1D400–1D7FF. The MathML entities use the code points for Greek in the range >> 0370–03FF. >> The Mathematical Greek Symbols do look nice, but to gain compatibility with >> MathML, >> we would like to revert to using the code points for Greek in the range ç. >> This will also allow you to enter Greek by switching to a standard Greek >> keyboard mapping. >> In passing we will also use the right symbol for φ (the current mapping uses >> what LaTeX >> calls \varphi rather than \phi). >> Any comments on this would be appreciated. > > I have no objection to this change but it doesn't appear that the > mathematical Greek symbols are incompatible with MathML. Looking at the > MathML 3.0 standard, in particular section 7.5, doesn't MathML support the > Mathematical Alphanumeric Symbols of Unicode using the mathvariant attribute? > That is, characters in the range 1D400–1D7FF are variants of corresponding > unstyled characters in the BMP (plane 0). Section 7.5 also says: > > A MathML processor must treat a mathematical alphanumeric character > (when it appears) as identical to the corresponding combination of > the unstyled character and mathvariant attribute value. > We aren't defining a MathML processor. [As an aside: does MathML really achieve this? If so, it presumably arranges for maths bold, fraktur and ordinary Roman letters to be treated differently somehow.] > Have I misunderstood? > > Are you defining MathML markup for Z? We aren't defining MathML markup for Z. What we are doing is closer to the Z Standard's identify markup. However we want to use MathML's entity set as the basis for the naming scheme for the symbols, since the UCS/Unicode names are unintuitive and unwieldy. We do need to define what Unicode code points the entity names map to (as, without a great deal of work, ProofPower will need to fix on one code point for each entity name for use on output). The code points in the 0370-03FF range have the advantage of being supported in a great number of fonts. Regards, Rob. _______________________________________________ Proofpower mailing list [email protected] http://lemma-one.com/mailman/listinfo/proofpower_lemma-one.com
