#metabrainz

alastairp: i see, makes sense. will work on adding it.

alastairp: lol. "Se gasta 50€ en un portabebidas de carbono para aligerar peso, sobrándole a él 20kg" - sounds about right

(from the comments)

:D

like the people who remove the plastic caps from their wheel valves for the same reason

mayhem: State of Community?

alastairp: ^ can you bother mayhem

Going back to recommendations, can we please have ‘reverse’ recommendations for everything? e.g. if we can work out your fav plays can we easily also do least fav? I just think it would be funny and v. shareable!

aerozol: especially for April 1

hah!

antirecommendations

Freso: we see it on the stream

Freso: and we hear you

"Ok, recommendation system, do your worst!"

"State of Community"

Huh

I see all black

aerozol: yes, I've thought about that. "We thought you'd hate these tracks, so we made you a playlist of them"

aerozol: Brian Whitman of the echonest did a hack at a musichackday many years ago, where he did recommendation/clustering of the music on your computer, and if you had anything that didn't fit into your "recommendation bubble" it would physically delete it from your computer

alastairp: D:

we should do that but with friends/followers

aerozol: isn't that what social media is for?

yeah but we should “remove” those people :P

From our life? The world? :D

‘delete’

Freso: LB username should work all of the time. please report to us specific cases if that's not working

it's actually finished

but we haven't merged it

alastair's famous poke-based development model

alastairp: merged too, not deployed yet :). well deployed on beta already but not prod

really good point that all sites have separate ways of reporting things. definitely something to centralise somehow

!m lucifer

You're doing good work, lucifer!

alastairp: speaking of, how are those python musicbrainzngs PRs going? :p

lol

reosarevok: requires more poking

🏏

mayhem: please pick something up and poke alastairp

this, maybe?

Maybe

Part of that (such as "edit_note_thanks") should remain in MB

but message / notification could be meb-wide

alastairp: I’m not sure I’d count it as finished for my purposes until it’s live yet. :)

mayhem, a session on LB roadmap planning?

mayhem, mbid mapping improvements including releases.

*release name

also, using mb_metadata_cache in many places but that came as part of mapping session.

and another session for YIM planning

If anyone comes up with topics for next years summit between now and next year, they can be added to the wiki already: https://wiki.musicbrainz.org/MusicBrainz_Summit/23

mayhem: actually I'd like to talk about that notification / messaging SQL design to see if people feel it's good enough or what should change, especially if we want to use (parts of) it at a meb level

Any business development ideas? (I have none.)

bitmap and I checked it ages ago but it never moved further because I got busy and we never got much more feedback anyway

Pleeaasseee!

+1 to YIM Preview

Yeah, we need to get on that Twitter trending trend.

👍

> Blue Monday is the name given to a day in January (typically the third Monday of the month) said by a UK travel company, Sky Travel, to be the most depressing day of the year. Scientists have dismissed the idea as baseless pseudoscience.

pff, scientists

^

reosarevok: Freso: is there a minutes doc?

What did I just get volunteered for?

writing minutes?

alastairp: None that I’m aware of, but also not going to deal with that while also managing the stream. :)

can you hear enought for this?

Mostly?

OK, I'm making one

comment by anyone, editor reosarevok, other @meb ask me if you want an invite to edit

Thanks

mayhem: I don't actually rate music in MB, mostly rate books in Goodreads :)

Should we consider displaying the number (or percentage) of responses at each rating level in addition to the average rating? This would provide some information as to the distribution of rating.

But I probably would do it in LB

Mainstream anglo-centric music is statistically insignificant.

Or show the distribution as a tooltip / popup?

Freso: lol

(semi-fair, but lol)

recording with most distinct ratings by different users in the db: https://musicbrainz.org/recording/cf916fa5-6d17...

alastairp: I’m disappointed that that’s not Rick Astley. :(

Definitely a good number of sockpuppets there though.

AC/DC Hells Bells top "real" recording?

Freso: PM incoming

uh oh 😰

But if an average rating is 3, does that mean it's mediocre or whether half the people love it and half hate it but nobody thinks it's average.

rdswift_: that can be interesting to show as well :)

freso points out that many of the users who have lots of recording ratings are very well known editors in the MB ecosystem -> picard power users?

(not posting their names here for privacy reasons, MeB devs ping me if you want the details)

I also know a number of people through the years have used (or wanted to use) MB as a central place to keep ratings across different media players, tagging formats, etc., etc.

Possibly the same people in alastairp’s list. :D

For what it's worth, RYM does *not* (AFAICT) indicate what ratings mean to their users (they even have forums threads where users talk about "what is your rating scale")

but there's a huge dropoff - 70k recordings for the top user, going down to <10k after only 7 users

How the heck would any person provide a rating that ISN'T subjective?

By other people voting on them, like with tags... oh, that's how ratings work :D

;-)

Goodreads’ rating scale _used_ to say that 1 star was "didn't like it" through "it was ok", "liked it", "really liked it", and 5 stars being "it was amazing", however it seems like their recent UI update has removed this hints. Also, similar to RYM, lots of people had their own rating scales, disregarding GR’s suggested hints.

I generally follow that schema though, both for GR and MB and other 5-star/rank type systems.

The main thing is that people understand that 1 is lowest and 5 is highest. How they differentiate levels between the two is generally their personal understanding.

Yeah.

I started rating on Rotten Tomatoes (who I don’t think let’s you just set something to ‘read’/watched like Goodreads?), and I decided to rate 5 as average instead of 7. Like, 5 - average, but I don’t mind average movies. When my friends see my ratings they fly into a rage

(because they are ‘low’) but I don’t think it breaks Rotten Tomatoes

I just mentioned to aerozol, I typically look at an artist's wikipedia page to find their "critically aclaimed" album to listen to first

I suspect ratings would reflect that as well

We effectively still have a scale of 10 in MB

Since we do .5s

^

So it would be trivial to convert anyway, if desired

Aren’t they stored or served at 0–100 too?

*as

But for the user it's still only 1-5 that they canenter.

Yes, something like that

*can enter

I think the rating + number of ratings are two things that should always go together. On RYM I will sometimes listen to a lower rated album that has thousands of reviews. Because I’m more keen to listen to their hits than something nuanced

alastairp: added introspection endpoint, need to figure out if we want to restrict it though/and how? we have a userinfo endpoint too so maybe we could merge both. anyways can explore in oauth i guess or later.

rdswift_: oh, huh, I forgot that :)

Sorry

lucifer: bitmap mentioned that MB actally has something similar to that endpoint, though it's not documented

if there's an RFC, and a good spec for this type of endpoint, probably OK to keep it public

alastairp: hmm, i do know of userinfo endpoint in MB but that is documented.

is there a reason why 3rd party app developers might want to have this?

right, i just read the oauth.net page on it and it recommends restricting it actually if needed.

"what scopes does this token have" might be useful

ok, great

yeah, scopes thing makes sense.

ideas to restrict it: separate flask app in CB, running in separate container, and use consul service to connect

rdswift_: would you like me to buy you a plane ticket to the next summit?

We feel that you should be present. :)

would love to meet you rdswift_

MB digest auth could probably continue to live in MB?

mayhem: I tried to get him to go as mentor for a GCI summit and he refused :(

mayhem, alastairp: should I be looking through the State of X presentations and summarizing those for the minutes?

(not tonight, obviously)

(thankfully it's all recorded anyway)

Also, tomorrow we should try and figure out a good way to order people, not sure if anyone is following the zoom screen often? Combining IRL and zoom raised hands seemed complicated today :)

rdswift_: that isn't a problem coming into spain. they care about consumer electronics -- they want to tax those.

personal medications, no problems at all

If it is preferable, we could have a ping system through IRC, so ping-to-talk, for remote people 🤷‍♂️

(I would like a good way to not interrupt mayhem :D )

Okay, I'll give it some real consideration then. Thanks.

“The future of ratings” topic, feedback wanted: https://community.metabrainz.org/t/the-future-o...

mayhem, alastairp: oh another thing we could discuss in a breakout session tomorrow is how to make troi easier to use as a library. it currently feels a bit clunky...

lucifer: sure

aerozol: that was quick

aerozol: "Your feedback will dictate the future of ratings." *help* dictate? Just in case we decide something that is not specifically proposed?

Okay, will update. I was trying to make it weighty!

Sure, I agree with the sentiment! :)

2 other people already answering, so seems like we'll get some discussion

aerozol, Freso: I pinned the topic globally for two weeks so people will see them, feel free to change that if you feel it is too long / not needed / whatever