I hope I 'm not throwing a cat among the pigeons, but I use the different statuses for a different purpose in a complex translation assignment for publication, arbitrarily assigning them the following meanings:

I use "Translation rejected" to earmark segments in which material needs research before validation (quotations to be sourced, validation required by the client from sources held exclusively in manuscript form by the client);

"Translation approved" denotes segments that have been previously translated in the series, and can be passed on unchanged;

I mark as "Signed off" segments that have been updated from previously-translated segments in the same series;

"Sign-off rejected" marks segments that are not to be translated: in this assignment, these are mainly instructions to the proofreader and page setter.

I should state that I operate solo, in direct liaison with the end client, who does not work with CAT software.

Point taken. But again, this is always possible as you wish. The idea here is to destroy the possibility to have the steps review and sign-off as separate processes. What you want, is always possible, provided you do not mix statuses from different processes. If you use the statuses from translation process, the file will remain "in translation" until all segments are finally set to "translated".

So please enlighten me, where the problem really is. If you do not want to have segments stored in your TM, use "project TM". You can store there as many "crap" segments as you wish. After having reviewed your translation and confirm all segments, you update your main TM with it.

Taking it in detail:

"Review mode – This mode looks different from the translation mode, which many translators find off-putting. It also requires re-confirming every segment."
This is not true. You do not need to confirm all segments one by one. Most probably you never ever tried to open a file in review or sign-off mode. Try and you will see the difference.

"Manually reverting segments to draft – This allows revision in translation mode but adds unnecessary steps and clicks. As above, translators may not want to reconfirm every segment if only making minor tweaks here and there."
Every change applied to a confirmed segment in translation mode reverts it to draft. You do not seem to work with Trados Studio. If you do not confirm a segment because you want to possibly change it afterwards, it also remains in draft status. You can confirm all segments in one go, you can filter on status and confirm only segments with a certain status. 

There are gazillions of possible ways of working with TS. This request to change makes really no sense, as you can work as desired with the current design.

I agree with Philippe's comment above. By default, translators confirm each segment before moving to the next unconfirmed one. Only confirmed segments are stored in the TM for future reference. However, confirmation does not mean the translation is final.

Currently, self-revision works in two ways (if you want the progress bar to be meaningful):

• Review mode – This mode looks different from the translation mode, which many translators find off-putting. It also requires re-confirming every segment.
• Manually reverting segments to draft – This allows revision in translation mode but adds unnecessary steps and clicks. As above, translators may not want to reconfirm every segment if only making minor tweaks here and there.

This process is inefficient. Many colleagues prefer revising directly in translation mode because changing workflows can reduce efficiency. In a large translation department, even small inefficiencies add up.

As long as you use not translated, draft of translated for your segment statuses, the file will remain "In translation". As soon you change ANY of the segments to translation rejected or translation approved, the file will go to the status "In review". And this is a very desired behaviour. What you describe is mixing processes. Changing this would render the review or sign-off process unusable. File in translation: given statuses are not translated, draft, translated. File in review: given statuses are translation rejected and translation confirmed. For sign-off these are sign-off rejected and signed-off. If you are ready with your translation, open the file for review and check how this works. Translation is the first step, review the next. Why should we as translators not do a "review" on our own file? And if you insist on the file being "in translation", do not use the statuses for the review or sign-off process. This is really as simple as that.

I understand the logic, but it does not match the way we work in practice. In my opinion, as long as I have not delivered the project and it is under my responsibility of translator, it is In translation. In the work chain, "In review" is supposed to be when the revisor receives it. As long as I am in charge of the document as a translator, it is never In review.  

As many people, I confirm the segments as the translation progresses, but it does not mean I approve the translation, because the QA remains to be done.

That means that, in practice, I often have to bring back manually the document status to In Translation, because it automatically switched to In review. I think that is the idea behind that suggestion.