Start a new topic

Segment comparation

I made this request sometime ago but wasn't taken into account, I just don't understand the CT results, it makes me take too much time to understand the results. Is me the only one who sees something is wrong in the way CT shows this results?

By the way, why the TM match from CT is SO low compared with the others?


image



The meme is for justification LOL.

I agree there is a much needed improvement in the readability area of the TM matches comparison, both in the TM view and in the Auto-assembling view (F1).


I have recently a suggestion in this regard: that the characters defined in the "Do not match" setting (Preferences > Memory) should be displayed in the TM comparison: https://cafetran.freshdesk.com/support/discussions/topics/6000055710 

I agree with both of you. I'm sure I spend more time trying to interprete the TM matching comparation than retranslating the whole segment again.

I was going to open a discussion on this very topic. For me, the way MemoQ and Omega (but also DVX)—oh, I didn't want to mention competitors—show matches is the way to go: highlighting new words with a coloured font (no background colour) and simply striking out old words (no background colour). 


I find that background-coloured texts tend to be an obstacle to segment readibility when trying to understand where the segment has changed.

I don't think it's bad mentioning other tools, we're not Martin Luthero, and most of us come from other tools :P.

Well, I did mention other tools, but then I didn't...

Besides the presentation (I agree with you) you can – concerning percentages – go even further.


In my humble opinion the percentage presented above is not too bad, but the devil is inside the small segments


Compare e,g.


criminal translators

criminal translator


or


bad worker

bad workers


For CT this should only be 50 % (and could so be out of the auto-propagate limit). For most other tools it will be more. CT compares on a word-to-word base only (correct me if I am wrong). This means it does not look into a word or compare the letters. Less accuracy means more speed. I wonder how much speed it would take to have a finer comparison.

The segment comparison has received a major overhaul in today's build, in case anyone hasn't noticed.

Igor, thank you for implementing also the new segment comparison feature. Since in general one single strike-through line is not well visible over Japanese characters, may I ask you to make it at least red and possibly bold in a future build?

 

Thank you

It is red in the last update. Your color settings are still from the previous update. Either reapply your current theme via the View > Themes menu or set it to red via View > Colors > No subsegment match color.

Done, thank you.

Much clearer, indeed!


Old:


image


New:


image



CafeTran Espresso 2019 uses the same font attribute (red, strikethrough) for:

  1. Text that's missing in the new source (see screenshots above).
  2. Text that's added to the new source.
  3. Text that's different in the new source.

image


image


Personally, I think that 2. and 3. should be marked separately. E.g. green underlined for inserted text, blue underlined for modified text.


1 person likes this

Seconded.


1 person likes this
Login to post a comment