Edit notes: context dependent pre-defined forms?


#1

Sometimes I don’t supply enough information in my edit notes to allow subsequent editors to know where information has come from.

I imagine a future, probably distant, where the edit note section is pre-populated to suit the type of edit being made, with appropriate items, such as check-boxes for points of matching for external links, sentences with blanks for URLs waiting to be filled in, already there to remind me of what a helpful edit note might look like.


#2

Several years back, when I was still new here, certain edits had to be made one at a time. It was tedious, but the advantage was you could give individualized edit notes rather than having to rely on dragnet lists of sources/rationales that only apply to a handful of edits in a cluster.

Sometimes when I know I’m going to be repeating the same task over and over again in edits, I’ll craft a boilerplate edit note that applies to them all, then copy and paste it into an external text document and fetch it as needed.


#3

See, that’s my problem. Coming from WP, they complain about storage space because the database stores a new copy of the entire article every time you click save, they encourage one large change instead of a dozen individual changes. In fact, they’ve told me as an individual (it wasn’t a policy) to not undo/revert certain smaller changes because of that very reason even when a policy or consensus clearly states that an edit is wrong.

So, I come here, and I am making mass changes with one click. It is silly to leave an edit note that is going to appear on 50 edits but each line only applies to 1 of them. I often leave them blank, or use the WP style edit summary (added personnel, corrected errors, etc).

And I really find it silly to include a link in the edit note when the edit itself is the addition of a link. :rofl:


#4

That is just how I saw it too.
Then I found apparent and actual misattributions by other editors and realised that adding a valid relationship requires them/me to have checked that it actually is accurate - as shown by things like matching biographical dates, area, role, genre and/or releases. And that the type of the matches found could be quickly stuffed into an edit note.
Eg “URL matches on surname, bio dates, and label”.

Though yeah, there still is the bloomin’ obvious like Bach being the composer of the Goldberg variations.


#5

Same in most cases. There are a handful of rare scenarios where an edit note is necessary when adding a URL. One case that comes to mind is edit #46511960, where I added an edit note because it’s not readily obvious by following the URL that the account is official since it lacks the “verified” blue check mark. In the edit note I linked to a tweet coming from that artist’s (verified) Twitter account in which he confirms the Instagram account is his.


#6

It’s also important to think of whether the edit note will still make sense 20 edits later.
For the add-link edit case, unless the edit history shows that exact link, the information is lost if the link is later edited/removed.
Another example is notes on merges. Once they go through, one of the entities goes away, so “obvious” comments like “same length/title/artist” stop being so obvious then.


#7

Ooooooh. Hadn’t thought of that. So “verbose mode with URLs” is the way to go for links.
What do you suggest for merges?


#8

For merges my point was more that a terse “same recording” isn’t enough even though right now it’s obvious that title/length/etc are the same.
It’s why scripts like mass merge are so useful because they automatically provide much of that info.


#9

I get that this is not enough.
What would be enough/better?
(Are scripts the only way to get good edit notes?)


#10

Scripts (which I do not use) are only as good as the person that created them.
Last week, I spent a couple hours (we all know that the website isn’t the fastest) correcting the information for one album that was added via a script.

What’s better - using an automated script to add information that takes hours to manually correct, or taking hours to manually add information?


#11

Depends on how wrong it originally was - I’d argue it’s better to have the script import it unless a) the data is not just incomplete but outright wrong or b) it takes significantly longer to fix the import than to add from zero.

At least in my main field (classical music) the main problem isn’t scripts, but crappy hand-entered data. Of course, new editors often (but not always) get better :slight_smile: (but so can scripts, if properly dealt with).


#12

One really good thing we used to have is « This is the nth edit in a set of x edits. »
https://tickets.metabrainz.org/browse/MBS-5128