Roberts (the programmer of the Welsh spellcheck program CySill) CLS008@ UK.AC.BANGOR.VAXC, who has been involved in this. My contact with character set problems has been with the proposed ISO 10646 where Johann W. van Wingen of Leiden (BUTPAA@EARN.HLERUL2) has been very supportive of all the minority languages. He was responsible for getting the Netherlands to reject the original draft DIS 10646 on the grounds that it did not support Welsh. The UK's ISO representatives are less than helpful in this respect. The problem with restricted character sets (like 8859) is that Welsh really needs 28 accented character combinations (i.e. acute, grave, circumflex, and diaeresis on any of a, e, i, o, u, w, and y - and, yes, they all do seem to occur, although some are extremely rare). a-u is no problem, and y/Y+acute, and y/Y+diaeresis occur in other standards, leaving the 8 w- symbols, plus the remaining 4 y- symbols. Acorn planned including support for Welsh in RISCOS, but gave up when they discovered how many extra character positions were required. In the Dictionary's software (which we have written ourselves) we have simply used an ad hoc solution to fonts and character sets, which does the job but which we would not expect anyone else to use. CySill may be able to help as they are bound to come up against the problem. Some Welsh fonts have been produced by MEU Cymru (Micro Electronics Unit of the Welsh Joint Education Committee - sorry I don't have an e-mail address), for use in schools, but I don't know what the character encoding is or how they are intended to be used. Andrew Hawke / Geiriadur Prifysgol Cymru / Aberystwyth