Tuesday, September 13, 2016

Re: Re : vFAD Zanata statistics

Le 13/09/2016 à 17:12, jean-baptiste@holcroft.fr a écrit :
> On 13 September 2016 at 12:10, Alex Eng <aeng@redhat.com
> <mailto:aeng@redhat.com>> wrote:
> Statistics from Zanata. Let me know if you see anything wrong.
> Wow statistics are very impressive. Thanks Alex.
> Anyone interested in writing article for CommBlog about this? Last one
> here [1]
> Thanks,
> Pravin Satpute
> 1.
> https://communityblog.fedoraproject.org/fedora-translation-sprint-5-days-50-members-20-thousand-words/
oh, I haven't notice it was a mailing list discussion ;) sorry for a so
short answer!

Here are three little graph about contributors :
* We need to understand what the Brazilians did, they have really
weird stats on three users : danniel, zdenek and fredlima. It can be
normal, but having a one sentence explanation is mandatory for an article.
** I feek like it is a different way of using Zanata, french team uses
Fuzzy and Translated, my guess is pt-br uses Tranated-Review (witch is
probably a better practice but makes reviewers mandatory)
* I corrected the total as it was wrong (Alex, do you know why?)
* In key numbers : it makes 112 366 word change done by 53 translators
on 20 languages
** About translators : 15 have >= 1000 words edits and 33 have >= 100
words edits
** About languages : 12 have >= 1000 words edits

Here are two little graphs about packages
* Best package is LIVEUSB-CREATOR, followed by INITIAL-SETUP (first
welcome message).
* Surprisingly, ANACONDA is in position 24/31 with only 9 languages at
100% ! This is _very_, _very bad_, for f24 we had 21 languages at 100% !
(my source is an extraction from 1st june 2016)

=> we should accept to lose half of installation tool languages for a
new release.

Jean-Baptiste Holcroft
trans mailing list

No comments:

Post a Comment