Reports are confusing

[emmapeel2]

Describe the issue

Contributor Reports are very confusing

I already tried

• I've read and searched the documentation.
• I've searched for similar filed issues in this repository.

Steps to reproduce the behavior

I am having issues to use and understand the contributor reports.

This was particularly unsettling on our last report round, we needed to verify how many 'reviews' and how many 'translations' some contributors had done (whatever that means).

There were many numbers and it is confusing to choose where to look at, how to calculate, etc.

For example, when we started the contributions, there were 13 000 strings to review. Now there are 1000. But the reviewer's "Source words approved" is 5000 , and there is Target words approved with 11000. Did he reviewed 16000, or only 11000?

We calculate translations by the source words, because languages have differents amounts of words and the source words never change. So, I dont want to know how many target words the translator approved, I want to know how many source words corresponded to those he approved. and possibly, if they approved translations for one string multiple times in a reduced timeframe, i want to only count the source words once.

Also, for example, if I edit a word in a string, and then I decide against and edit again... How many words have I edited, one, or all the words in the string, or all the words in the string twice?

If I accept a suggestion, do all the new words count for me as a translator, or maybe as a reviewer (even if i havent marked them as approved, you could say i reviewed them)?

Expected behavior

It would be great to have a new report, created by someone that has thought about all this, and gives you a number for 'reviews' and one for 'translations' that does not overcount and is not lost between many other numbers that you dont know what they are.

For my point of view, I think the best number is:

For translations: After all the actions a translator has taken, how many words have the source strings they translated? I dont care if they edited the target string 1 or 20 times.
Or maybe I deleted the string, it would be unfair not to count them even when we dropped them.

And for reviews: how many more source words have an approved translation? So, the number words of the unique strings the user has approved a translation of.

I have read the docs at https://docs.weblate.org/en/latest/devel/reporting.html#stats (i think some of the fields are undocumented).

Screenshots

No response

Exception traceback

No response

How do you run Weblate?

weblate.org service

Weblate versions

No response

Weblate deploy checks

No response

Additional context

No response

[nijel]

For example, when we started the contributions, there were 13 000 strings to review. Now there are 1000. But the reviewer's "Source words approved" is 5000 , and there is Target words approved with 11000. Did he reviewed 16000, or only 11000?

Mixing strings and words doesn't work. Source words are words in the source string, Target words in the target strings. This is also described at https://docs.weblate.org/en/latest/devel/reporting.html#stats.

We calculate translations by the source words, because languages have differents amounts of words and the source words never change.

Then look for Source words in the reports and ignore Target words. That's why reports have so many fields because each project wants a different metric for seeing project progress or for paying translators.

For translations: After all the actions a translator has taken, how many words have the source strings they translated?

This should be Source words total minus Source words approved.

And for reviews: how many more source words have an approved translation? So, the number words of the unique strings the user has approved a translation of.

That should be Source words approved.

I don't care if they edited the target string 1 or 20 times.

But right now, it does count all edits, even repeated ones for a single string. This could be changed, but needs a better definition of what to actually count.

I think it's pretty clear that the following could be folded per user:

• User A edits string for the first time
• User A edits string for the second time
• User B edits string for the first time
• User B edits string for the second time
• User C edits string for the first time
• User C edits string for the second time

But there can be outside reasons why the string is edited again:

• User A edits string for the first time
• Source string is changed
• User A edits string for the second time
• User B edits string for the first time
• String flags are changed making it failing check
• User B edits string for the second time
• User C edits string for the first time
• Project policy has been changed to use informal language instead of formal
• User C edits string for the second time

Users can edit each other changes to improve the translation (or to argue about it):

• User A edits string for the first time
• User B edits string for the first time
• User C edits string for the first time
• User A edits string for the second time
• User B edits string for the second time
• User C edits string for the second time

In all above cases, all users will currently have two edits in the reports.

Or maybe I deleted the string, it would be unfair not to count them even when we dropped them.

Deleted strings are not included in the reports right now, and won't be until #1993 is implemented.

[github-actions]

This issue has been put aside. It is currently unclear if it will ever be implemented as it seems to cover too narrow of a use case or doesn't seem to fit into Weblate.

Please try to clarify the use case or consider proposing something more generic to make it useful to more users.