Florian,

Using the attribute as hint to the translator is a clever idea. In fact, I think we should consider adding an attribute for that express purpose, like 'title' or 'tooltip'. Since lex-replace (or by any other name) will be used for word counting, I think it may cause problems to overload it with the descriptive text.

For example,

<t-u maxwidth="40" units="chars"> <source><ph lex-replace="count">%2d</ph> characters found in <ph lex-replace="Name of the file with a long descriptive text">%12s</ph>.</source>

becomes the following after lex parsing:

"count characters found in Name of the file with a long descriptive text."

Concerns:

1. 'count' is more than 2 digits and may exceed the maxwidth constraint causing the translator extra work to reduce the size.

2. Although the file name should be considered one word, the long descriptive text artificially inflates the word count. (any may exceed the maxwidth constraint)

In XLIFF 1.1 (without the lex-replace attribute), inline tags are not counted, so the parsed text would be:

" characters found in ."

which isn't exactly right either.

I suppose the preferred lexical equivalent would be:

"00 characters found in ssssssssssss."

which could be represented in XLIFF as:

<source><ph lex-replace="00" title="number of characters">%2d</ph> characters found in <ph lex-replace="ssssssssssss" title="file name">%12s</ph>.</source>

Regards,

Ektron, Inc. +1 603 594-0249 x212 http://www.ektron.com