Today I finished a file with 200 K words of which 95 % were repetitions. There were some pros and cons when working with CT:
This really helped me:
You wouldn't achieve much with splitting of the documents apart from complicating and risking the export. For example, just think of splitting sdlppx project into smaller files. That would be art for art's sake. Nothing practical. It does not matter with the save of, say, 10 smaller pieces instead of one larger piece. You might increase the automatic saving intervals in Edit > Preferences > Workflow tab with larger projects.
> Filter for repeated and propagated segments with only one occurrence per repetition (to go through once and to automate any further steps)
> Option to filter all translated and propagated files (don't we have this kind of cumulative filter already?)
Yes, we have it. Please turn on "Multiple filter" in Edit > Find before applying the filters. Note that this option is canceled when you turn the filter off.
> Option to have propagated segments automatically checked when one of the other segments is checked.
Apply Task > Set checked status for the target segments on the filtered propagated segments.
> Optionally more colours to mark propagated segments (or segments with any other state). The P is fine, but when you scroll the grid with hundreds and dozens of propagated segments to find a translated, but not propageted segment a visually even clearer indication would be welcome.
Maybe it would look nice but the idea of hundreds of options, that is, for each and every "pixel" of the interface is a bit scary. The interface for options will become more complex than the main working area. The user will spend more time fiddling with options than translation. And we might reach the absurd of suboptions for options.
Indeed this is how I did it, but if other propagated segments are already pre-filled / translated (or in case of proofreading), the task becomes cumbersome (with propagated segments not following systematically each other – AAABBBCCC or ABCABCABC – but being spread around – AABBACBCCB).
> Maybe it would look nice but the idea of hundreds of options, that is, for each and every "pixel" of the interface is a bit scary. The interface for options will become more complex than the main working area. The user will spend more time fiddling with options than translation. And we might reach the absurd of suboptions for options.
A basic and a philosophical question (not speaking of this single question here). A CAT tool is complex sui generis. I assume no Studio or memoQ user ever worked through all of the hundreds of options both programs offer. Some years ago Kevin Lossner wrote that most translators use only a few percent of their programs' possibilities. This might be true, but there can always come a situation where you are very glad that option x or suboption y exists (eh, if you are pointed to it in a forum or in the, eh, documentation). Or let it say otherwise: You can use option x or suboption y, but you are not obliged to (even not to take notice of it if you do not need it). CafeTran offers in comparison to Studio or memoQ rather few options, but that cover 98 % or even more of a translator's needs (not to forget its productivity boosters). But sometimes it would be nice to have the option to adjust only some screws more – to have the liberty of choice.