I wrote:

> Further, by doing so we are creating again a closed system
> using its own conventions with no links with external tools adapted
> to Unicode. I will be able to process a file and extract information
> from it with, say, Python very easily if they use a known representation
> (iso encodings or Unicode), but if we have to parse things like \japaneseai
> or similar, things become more difficult.  I think it's a lot easier
> moving information with blocks of text and not with single chars.

:-/ It seems I've given a good argument against what I'm upholding... (at
least in part)