- From: Dominique Hazael-Massieux <dom@w3.org>
- Date: Fri, 9 Nov 2018 10:30:49 +0100
- To: Jens Oliver Meiert <jens@meiert.com>
- Cc: W3C Translators <w3c-translators@w3.org>
Hi Jens, Le 30/10/2018 à 22:16, Jens Oliver Meiert a écrit : >> I believe there are some options. For a better overview and an option >> to collaborate I’ve sketched them in >> https://docs.google.com/document/d/1OqVOxwbuhFVwVVXpJcENIvI22BU_ZL1u6_VKO_dvk5M/edit. > > Is this a model that could work, too? I guess my main overall comment is that approach #1-#3 is basically what we had before and had enough issues with that it led to us stopping the program. More specifically: * we could require all we want from "unreviewed" translations, unless we make this checked on a regular basis, the marks of "unofficial and unreviewed" could be removed very easily at any point; and more systematically, unless we have a reliable system to track changes, something that was reviewed as OK could easily be replaced by something that is not. Your notes suggest a tool (or a workflow) could help reduce that risk, but it's not entirely clear what such a tool would do that could not be easily defeated by someone intent to abuse the system (and a manual workflow, given the number of translations, seems hard to keep affordable) * the rel-unfollow deterrent was obviously not enough of a deterrent for many of the abuse we identified over the years Regarding your "†" note "Reviews could happen on a voluntary basis" - that's definitely what my thinking of the new program is, in case that wasn't clear. Dom
Received on Friday, 9 November 2018 09:30:53 UTC