These subscripts in CO2, SH6, CH4 etc. are a real PITA. They mess up auto-translation (formerly known as auto-assembling) and are a real pain especially when they are followed by a dash + word.
How about a filter option to replace them with the corresponding Unicode chars?
And ... allow the user to replace them in SDLXLIFF files.
Example of messing up:
I don't blame CafeTran Espresso at all, here! Just looking for a workable solution.
For SDLXLIFF this will be difficult, since CafeTran Espresso doesn't know what the content (meaning) of the IDs is:
On the other hand, the user is smarter and can tell CafeTran Espresso that SF₆ and SO₂ are meant here.
Here's a screenshot from an Ms Word document where subscript is used for indexes of chemical formulas:
Intimidating, isn't it? Dating is nice, intimate too, but intimidating? Hmmm...
Here's a screenshot of how it looks if 'one' would have bothered to use the appropriate Unicode chars:
Nice, isn't it? And note that the glossary terms are being recognised.
But an import filter option would be nicer :)
Why don't you just define those non-translatable fragments between red tags as hidden tags? Then CafeTran will merge all the tags as only one tag (e.g instead of three). You will still see the full contents in the grid. Alternatively, uncheck the Hide tags details option to see the hidden tag's contents in the editor.
Will this work in words like CH4-Gas too? I always thought that it's not possible to enter a tag in a glossary entry.
Don't enter any tags in the glossary. Just tell CafeTran to hide the non-translatable expressions behind the (purple) tag:
e.g 4^ in your last example, or |\d+^ to hide all the digits.