#metabrainz

I mean you made the ticket

Yes, does the ticket fit what you wanted

hm..

Just want to make sure I got the point

i mean, yes I guess?

If not, edit it :)

bleh. how do yo udo indent on our wiki (putting a " " before a sentence jsut maes it like in a grey box with monospacefont)

anyone?

use :

cheers

:-)

yvanzo, bitmap: back to Wed meetings, right?

alright!

mhhh wet meetings :D

is there no better way than to make a ticket for each new area? :P

Nope, that I know

automating areas management was a suggestion for gsoc btw

yvanzo: Do we have a reliable source for this? Neither Wikidata nor Geonames (which are two mostly used now) are good enough.

How so? Aren’t these our current references to create areas?

yea. doing it automatically is how we ended p with half a million 200pop US dupe-name villages ¬___¬

(I still say we should remove any of those unless in use (artist, recording, place event or user is attached) and thne re-add as requested)

I mean I got no issue with some dupenamed us village beign re-added if a concert was held there or an artist came fro mthere or whatever

but lets be realistic? many of those will *never* hav a use in mb

it's also this dilemma: either we remove them, or we add so many villages fro mevery other country. that is impossible and unmanageable

unused villages shoudl be removed and thne readded as needed

that way our database doesn't look like some US village directory or whatever

I mean. to a russian this will look odd probably :D

the idea was not to dup the whole wikidata/geonames, just to automate using it to create missing areas on demand

and.. that worked badly :D

it has not been implemented, or am I missing something?

i thought we were talking aobut old areabot or what

i may be mistaken then

oh, alright, I don't know much about areabot, it has probably been dismissed a while ago

yea it has

also that is why drsaunders is doing it manually

but there is still a lot of unused villages in the database :/

That many?

honestly

i don't know the numbers. but I think there was quite a few

That would be nice to have some feedback about the old areabot, for the record, to prevent having the same issues again.

because and this was the thing with areabot, is that for some reason, us-villages in wikidata was labelled with.. I don't remember. some spesific property which it was to interpret as "this should be added", a thing we hadn't known in beforehand

I think there shouldn't be a problem with that tbh

Right now we show the most used areas on top

I cannot find anything about AreaBot on the wiki. It is just mentioned in comment to MBS-8361

MBS-8361: Use geonames data for mb-areas https://tickets.metabrainz.org/browse/MBS-8361

Villages were marked as villages, we were supposed to add those - the problem wasn't that, but that the rest of the world just didn't have proper data

reosarevok: true, and that helps *a LOT* but there is the issue with al lthe wrongly added relationships before that was fixed :/

That's probably better nowadays, but still not great

Wikidata still has crappy coverage of many non-Western areas

indeed

(it used to be AWFUL which is how we had every tiny village in the US but were lacking most big cities in, say, Romania, not to mention something like Indonesia)

but iirc it was also that areas which outside of the us wouldn't be lablled as "village" because the pop was too low, *is* marked as village becasue it's soem "official" designation in the us or something

in addition to what yo ujsut said

I don't remember tbh

I mean. as long as there is something musically related i'm all for adding it. but there are going ot be palces that have a population of 200 and wil lnever

(or when, eventually possibly. we can *then* add it)

I feel having everything is better than not having stuff unless we can find a neat way of auto-adding

But

iNaturalist has an interesting way of dealing with missing species

...

When the user requests them, it automatically searches a few databases, and adds them automatically based on that

cool

Us doing the same based on Wikidata would probably work (I guess that's kinda what yvanzo suggested?)

hmm

atleast having an automatic ticket creation!

Where rather than needing a ticket, you'd have a way to say "search for this area in Wikidata" at the bottom of a MB search

Like we have "Add a work" or whatevs

oohhh

but what if it's not found?

(or, user error)

Well, then we'd still need to add it in Wikidata I guess

We could still keep AREQ for that if needed

But it'd automate a lot of cases I suspect

hm! yea

and if things are well linkied in WD the relationships can be propagated in mb as well

(iNaturalist does it so that if it's not in the automatic DBs, you can request it manually)

sounds liek a plan tbh

(but with the addition that we run a cronjob to remove every village with ### pop that is not in use)

(maybe i'll be wrong and there wil lbe liek 10)

Yeah. iNat adds the parent taxons automatically based on that data, and that should also be doable in MB: "if we have the parent in MB, link it, if not, add the parent, until we get there"

(we'd need to decide what intermediate things to skip, etc, but the general idea should work)

sure

tbf, I don't actually agree with skipping intermediate thnigs. I mean if we are "going to have all the things" and "feel having everything is better than not having stuff" then it shoudl aply to larger areas, and soem historical areas too?

releases, people and recordings can easily have "intermeidate linked area" credit without having a smaller area

or it can be old and information fuzzy

or area used to be non-intermediate but a higher level but was subsumed later. etc

can areas be linked with other areas with a date range?

so you could like logically say "blup area was a part of thingycountry from 1200-1500" but then was a part of othercountry 1520-1907" and "become soverenge state in 1908"

eh. ignore the last one

that'd be the start date of the area i gues

but like

Currently AREQ already requires a reference such as Wikidata or Geonames.

what do we do for a medieval work written in some place in some date in a region that no longer exists, but does have a n equivalent area today

but borders are not identical so you sohudl rather use an historical name

Historic areas are currently not supported.

and that is a huge problem imho

not even near historical such. but old ones

anyway we do actually have some.. we have east germany afaik

and .. chezoslovakia?

at ome point we *have* to enable historical areas.. what happens when todays borders are altered?

say catalonia becomes a country.. how do we deal with that?

make a new entity and igonre that is used ot be aprt of spain until such time

?

all old links of artists to this area would be erranous if moved to new country "catalonia" whne the person f.ex was born and lived and died in "spain"

We should probably replace area fields with area relationships (which have dates).

that works if the old area is equivalent to the new one.

not always so

Norway of pre kalmar-union had a prt of what is now sweden

(as an example)

w.ex a writer fro mthere. how do yo udo? do yo uwrite "was from norway (daterange)" or "was from sweden daterange"

what if they lived during the change?

what if they chnged theri own oppinion of it ?

what if data is unconclusive

taht's live, capturing the complexity of border changes is beyond MB goals IMHO

for the "herjedalen" scenarion it's fairly ok to just use norway with he data range (as it was a prt of that then) but for situations where the old area stopped exisitng. or a new area created

I just use Estonia for Soviet Estonia people

if there is a floss project that handles border changes and historic area data, we should surely use it though.

yvanzo: yes and no. it's unfortunately something we're going to have to deal with eventually as the data we colelct gets better and better. music s not too unoften closely connected to "area" bot historical and current(ly changing)

Because it's more relevant like that and it's not like they generally chose to be Soviet

yep

But I can see how it's tricky

that makes sence..

yes

That said, I kind of agree with yvanzo that we should figure out how to deal with music stuff first, and area stuff later :p

i mena yea

(we still suck at a lot of directly musical bits)

bitmap: around btw?

(and this is why i think "nuke the unused villages"! :P)

But that's effort! :p

tsk

(and the MBID is probably already in Wikidata, etc)

both those would be easy to do with some report regex thingy

atleast I'd be very interested to see a list of "areas ith no link other than to one other (parent)area"

and order by listed pop

yvanzo: do you have some time to check some PRs?

yup!

BilleeKhoj: the pop thing is harder, but the first bit is easy. One sec

and population, not pop music , before you go all dads on me, reosarevok

Pop music would be easier!

:P :D

Since we store that as tags, but not the population :p

yvanzo: I really want the genre stuff to get merged. https://github.com/metabrainz/musicbrainz-serve... - do you feel that's problematic? (IMO we can look at how alias stuff will work once we actually implement it)

but like I'll shut up if its like 10