XML 1.0 does not accept all Unicode characters .... the legal ones are:
#x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] |
[#x10000-#x10FFFF]So if you wish to serialize a CAS to a file or to a remote service you'll have to avoid the 29 legal (but useless?) low value ones. UIMA could replace or escape them but both have possibly undesirable side-effects (lost information & non-standard XML.) At the least this restriction should be documented.
