There is a few questions to answer first:
- what is the 8-bit language encoding you use (does your 8-bit encoding of the "important" characters differer form the first 256 positions of the UNICODE?)
- are the affected characters used only in the data, or are they also hardcoded within routines/methods/class definitions/...?
- are there $listbuild structures used (if you use persistent classes then the answer is "yes")?
- is it sure that the $lb structures were not used in subscripts (oh, it is silly idea isn't it :-)?
- are binary data strictly separated form the textual ones (I mean not only pictures etc. but various hashes, signatures etc.)?
- is it a production system?
- how big is the database, what is the time window to provide the conversion, how fast is your hardware?
- + bunch of minor or subsequent questions...
The answers and their combination affect heavily the conversion procedure. For example, if your answer to the first question is "no problem, the character codes do not differ", then no conversion is needed at all.
If, on the other side, you face a transfer of xTB of data on a huge production system with hundreds of concurrent users with SLA > 4nines, consider some budget and hire a Maatrix expert (me).
Currently the data in subscripts does not need any translation. Once you enable your database to UNICODE, the data in subscripts pop up as UNICODE data.
My last visit to Buchlov, a few years ago, was in the scope of a medieval philosophy conference, focused on the works of Jan Duns Scotus, doctor subtilis. I look forward comparing the atmosphere - then and now.
Log in or create a new account to continue