Topic on Talk:VisualEditor/Diffs

Topic heavily vandalized. Hopefully closing it should help.

• 
  Wikitext diff showing a paragraph being moved and "subtly" vandalised
• 
  Visual diff showing a paragraph being moved and "subtly" vandalised

The community has been asking for improvements to the diff-engine. The above example, copied from VisualEditor/Diffs, cannot be implemented until the diff-engine itself is enhanced to detect that a paragraph has moved and changed. And if that is done, the first priority is to upgrade wikitext diffs to show that a paragraph has moved and changed. 95% of editing is wikitext editing. Visual work is an almost insignificant 5%. The persistent neglect for core editing needs, in preference for largely-unused visual projects, is undermining the project in general. In this case it's literally impossible to build the described visual diffs until our core diff engine is upgraded. Doing the work to implement this diff in visual form, without also doing so in the wikitext diff, would be actively perverse.

P.S. I'm sorry if my tone was overly negative. Repeatedly seeing this pattern in other places has been getting very frustrating.

Wait, is that a genuine screenshot of a moved&changed paragraph? Or a mockup? If that's a genuine screenshot, why didn't the improve-diff project improve the wikitext diff?

> cannot be implemented until the diff-engine itself is enhanced to detect that a paragraph has moved and changed

Incorrect, Visual diffs have already been implemented. They are already working on all wikimedia wikis.

> In this case it's literally impossible to build the described visual diffs until our core diff engine is upgraded

Incorrect.

> . The persistent neglect for core editing needs, in preference for largely-unused visual projects, is undermining the project in general.

Incorrect again, https://meta.wikimedia.org/wiki/2016_Community_Wishlist_Survey/Categories/Reading#CW2016-R094, https://phabricator.wikimedia.org/T121469

>the first priority is to upgrade wikitext diffs to show that a paragraph has moved and changed

Perhaps edifying yourself would useful before making assertions.

These comments seem to have been mostly made out of ignorance.

Did you even read my last comment? Anyway, back to the point...

Simple Diff got a negligible amount of support compared to other things. So thank you for making my point for me, and thank you for digging up exactly the links I wanted to cite. Development priorities are backwards. Why is this diff working in visual, but not for displaying this diff for wikitext? task T121469

P.S. The new change to Flow's mode-switch is a real hindrance. Who the heck makes a submenu with exactly one active item? Normally that's just a bunch of annoying and pointless extra clicks, but for browser-related reasons at the moment it's being an active obstacle.

> Simple Diff got a negligible amount of support compared to other things.

First, there have been requests for this for years before you became an editor. The parent page of this thread shows prior work related to that.

Second, your assertion that the priorities are backwards is still wrong. The phabricator you linked to clearly proves that they did work on improving text-based diffs, and there is more evidence elsewhere that disproves your theory even more. Easy to dig up for anyone who's interested.

> Why is this diff working in visual, but not for displaying this diff for wikitext?

It seems to do an analysis on the html by comparing it to html changes made recently, it can even work without any wikitext at all.

>Why is this diff working in visual, but not for displaying this diff for wikitext? task T121469

If you mean why doesn't it work outside visualEditor, that seems to be because this is just the first step before that (https://phabricator.wikimedia.org/T105173).

First of all, can you please post by username instead if IP?

Secondly, thank you for proving my point! I asked why the example diff works in visual diff and DOES NOT work in wikitext diff, and your answer was to link to Yet Another visual-task obsessively soaking up work in precedence to our primary systems. The reason so many in the community actively resent VE is because development priorities are backwards.

Our primary (95%) editor and core systems are repeatedly neglected or sabotaged in the name of the insignificant secondary editor(5%) VE.