The problem was first addressed by the participants of the Hilandar Conference at Columbus, Ohio, 11th-13th April, 1981 (proceedings in Polata k"nigopis'naja vol.5 (1981), pp.43-89), and a proposed transliteration for computerisation of mediæval and modern cyrillic was published ib. p.89. This formed the basis of the system published by A.J.Geurts, A.Gruijs, J.Van Krieken and W.R.Veder, ib. vol.17-81 (1987), p.28. This system has been quite extensively used. It was further modified by Carsten Grünberg in the proceedings of theBlagoevgrad conference (D.J.Birnbaum [et al.] (eds.), Computer processing of medieval Slavic manuscripts: proceedings [of the] First International Conference, Sofia, 1995), p.65.
At the same time, a somewhat different approach, resembling the transliteration systems already in use for electronic processing of ancient Greek, was adopted by the Corpus Cyrillo-Methodianum Helsingiense; this has also won some followers, and is well suited to the encoding of the canonical Old Church Slavonic glagolitic texts, but it seems doubtful to me how easily it could be extended for other forms of Slavonic texts, particularly the early printed cyrillic books which are my immediate concern.
My proposals below therefore belong to what we might term the Hilandar Conference tradition, in which the basic principle has been to use the letters j and h and the numerals 0-5 as modifiers. My proposals relate specifically to the transcription of early printed texts, for which, given that they have a limited repertory of characters, it ought in principle be possible to devise a complete transcription. I have not attempted to address the problems of MS transcription. Equally I recognise that once we do more work we shall inevitably encounter more problems that I haven't covered.
Some of my proposals relate to principle, others to practice. In order to avoid confusion, I have tried to deviate as little as possible from the scheme as it already exists.
On matters of principle,
Having got this far, it seemed a natural progression to attempt to do the same thing for modern cyrillic. Here the basic principles are the same - the use of selected ASCII characters as modifiers to expand the repertory sufficiently to provide equivalents for all the cyrillic characters - but the practice has had to be slightly different. In particular, we cannot use numerals as modifiers, because in modern texts we shall want to use them as numerals. For the same reason, other traditional transliterations such as the apostrophe for the soft sign are also unusable. The selected modifiers are h, j and x. The result looks a bit peculiar at first sight (though there are precidents for some items that might strike the modern reader as odd, e.g. for "jl" and "jn" in sixteenth-century Venetian printings), but it does seem to answer the requirements of a complete and unambiguous machine-readable system. It is presented in Table 2.
One aspect which does slightly worry me is the fact that there are occasional discrepancies between the two system (i.e., mainly, equivalent letters having different encodings). I would welcome comments on the extent to which it is desirable or possible to harmonise the two. It seems that the modern system cannot be brought into line with the old cyrillic system (because of the numerals); on the other hand, the modern system, not having been used so far, can be freely modified, whereas a considerable amount of work has already been done using the old cyrillic system as published in Polata k"nigopis'naja vol.17-81, of which the present proposals are a comparatively minor revision, or rather extension.
I have already prepared TEI-conformant WSDs for both systems, which I propose to publish in due course. Before I do so, I would like to incorporate any revisions arising from your comments on the proposed transcriptions. I shall therefore post these transcriptions on the web for a limited period, at the end of which I shall remove them. I should be grateful if, during that period, if you observe any errors and omissions, or if you have any suggestions for their improvement, you would send them to me. The definitive versions should appear shortly afterwards.
Since HTML does not yet have any adequate representation for cyrillic characters, the tables have had to be inserted as .gif files, which you will have to load if your browser doesn't do it automatically. This may make the whole page very slow, for which I apologise.
Please note also that the transcription in Table 1 is intended specifically for use with early printed cyrillic books. It can also, of course, be used for modern Church Slavonic. You should not expect it to be fully adequate for other purposes, such as the encoding of mediæval manuscripts, which entail a somewhat different set of problems.
