cet1@cl.cam.ac.uk (C.E. Thompson) (05/01/91)
This is a query directed at suitable-expert@adobe. The new PostScript LRM says (p.284) that `when you copy a font directory for the purposes of creating a modified font, you should _not_ copy UniqueID' (except in the case when only FontName, FontInfo, FontMatrix and Encoding have been modified). It doesn't explicitly say the same thing about XUID, but surely the same applies to this value? Existing PostScript code will only omit UniqueID in such situations (or alter UniqueID, which is sufficient in the case of Type 1 fonts, as a mismatch with the UniqueID value in the Private directory effectively causes it to be ignored: see p.17 of the Black-and-White book). However, the new LRM says (p.285) that `PostScript interpreters that recognize the XUID array ignore UniqueID whenever an XUID is present'. Doesn't this create a serious compatibility problem? Using a new font (with XUID as well as UniqueID) with a new interpreter, but old PostScript code (correct when it was designed), the XUID will be left unaltered, but it is the one thing that ought to have been changed. Even if we all now start using code that omits XUID as well as UniqueID (and such omissions can be done more conveniently in level 2, because one can copy everything and then use 'undef'), this won't help with existing applications. Chris Thompson JANET: cet1@uk.ac.cam.phx Internet: cet1%phx.cam.ac.uk@nsfnet-relay.ac.uk