[META] Find solution for growing number of strings extremely fast

Problem/Motivation

The number of strings to translate on ldo is growing too fast for localization teams which hardly succeed in keeping their translated strings ratio.
The consequence is that this ratio tends to drop and, as of today, there is no team with a translated ratio greater than 20% (see the Explore languages page).
These quite low and ever-dropping ratios contribute to lower the motivation of contributors and give a false idea of their work and the state of translated strings for their languages—which does not help to recognize their often underestimated work.
Some (a lot ?) of these strings are only minor— or, worse, erroneous—variations of existing strings.
Some (a lot ?) of these strings are deprecated strings from older versions of projects and will never be translated.

Proposed resolution

• Change strings count and ratios logic to only include the last version of each project.
• Code a developer-oriented module, code name "Logorrhea killer", which would parse all strings in t() and compare them to strings already present in ldo to offer the module developer the opportunity to replace his redundant strings with the ones available on ldo. For the developer, this cross-pollination catalyzer* would have several benefits: reuse—or, at least—get inspiration from existing strings vs. reinventing anything from scratch, release a module much appealing to translators, make less mistakes...

Remaining tasks

Talk about it. :)
Understand if all this is possible/desirable and how tough it would be.
Code!

User interface changes

None on ldo, "Logorrhea killer" would be an independent module.

*©SebCorbin ;)