Topic on 2017 wikitext editor/Feedback

When I fill in a citation from an external website using the automatic form (for websites which do not have a good translator--and there are many), I do not get an opportunity to fill in the rest of the citation or make changes to the automated piece-parts. All I get is the "Insert" ref button.

Topic:Thdkejww5xpfqltq would help with this, but so would some part of the flow which enables me to make further changes to the citation, presumably using the same modal popup ("do you want to make changes to the automatic citation?"). This is also "a problem" in VE, but at least in VE I can proceed to edit the template visually.

Agree with this, needs an "edit" option as well as "insert". Also needs an opportunity to name the reference; currently you can only do it by editing ref tag after insertion.

The general idea is that first you insert it, and then you edit it (in this case, you edit the wikitext directly). I think that the fairest thing we can say about this model is that it has strengths and weaknesses.

That's a bad idea in VE too, but there it only introduces an unnecessary extra click to edit the inserted citation. The behavior of the new wikitext editor is likely to introduce errors, because citation templates are fiddly and many people avoid editing them manually, and is a regression compared to the behavior of the old wikitext editor + reftoolbar. Sourcing is already such an uphill battle that making it harder to efficiently produce complete, accurate, and correctly formatted citations seems like more than a minor weakness.

There's some information on user research at phab:T107078. Naturally, since I personally want this change, I'm skeptical of any data that disagrees with me. ;-)

Me too ;) But I'm especially skeptical of data about "user behavior" that seems implausible and inconsistent with my experience as a user, and even more skeptical of statements that aren't data but allude to the existence of data somewhere-or-other that isn't available for broader review. (For example, implicit in the claim you link to is the assumption that the abandoned edits would have been desirable ones had they been made. It seems to me that it is not a bad thing if people back out of making claims if they discover they can't correctly source them.)

• Median edits per registered account: Zero.
• Number of edits for you and me: A lot more than zero

It's conceivably possible that you and I are not exactly typical. ;-)