I think Transifex actively keeps people without accounts from even seeing strings, and that it is broken for translators, language coordinators, managers, project maintainers, and developers.
@Matth78 You can either look up “kingu” on Transifex to see how many contributions and projects are listed, or I will gladly tell you exactly how it fails for any of the topics you are interested in.
Both options are waste of time, so here goes:
Locking strings means people finding errors can’t improve them, which is many orders of magnitude more prevalent than malice. You are then banking on either the broken non-threaded Transifex messaging system, or the hidden commenting system.
You will not get all the best translators into the reviewer role, and even if you did they are not infallible.
With Transifex you can only stay atop everything as it happens by e-mail, or see quality erode.
You can’t know if suggestions are actually deleted entries, or ones found in the translation, so every time you want to go over something it turns into a whole lot of checking, mostly in vain.
There is no language-wide, or project search. If you don’t put up with all that, you aren’t arriving at any level of consistency. The only features reminiscent of search and overview are well hidden.
It is of course way worse than that, but you can try to argue differently from that premise.
That means new translators don’t learn, which even more-so solidifies Transifex as the home of the drive-by translator. The biggest projects either have very talented and extremely diligent people, or quality suffers. The Tor project in particular is famously bad. If you want “participation” but not string-changes, you have the worst of the worst.
You can’t solve the issue with roles or subsystems, because it doesn’t scale.
For every guy like klint telling you no difference could be made, you shed a whole lot of possibly good translators by default, and they won’t say anything about it.