Blacklist sites

Tags: #<Tag:0x00007f37c1d59ee0> #<Tag:0x00007f37c1d59d00> #<Tag:0x00007f37c1d59af8>

There is an editor who adds streaming links for artists to their site. Said site acts like an aggregator and has no content of it’s own. Streams are actually provided by Youtube:
https://musicbrainz.org/edit/61830738

Do you think a blacklist would be a proper measure or would it eventually lead to new problems? Like when the domain changes owner and is used for other stuff.

6 Likes

A better idea would be some guidelines for what should and shouldn’t be used for that relationship. For example, an official licensed streaming site such as Spotify should be used, while an unlicensed site or an aggregator such as in the example you cited should not.

Edit: Here’s a STYLE ticket I put together.

8 Likes

I agree that we should not link to this site, especially not as “streaming”. The pages aggregate a lot of data from other sources, but one thing it does not provide is streaming of the artist’s music. There might be an occasional YouTube video embedded, but if the site says it allows streaming an artist’s music I would at least expect a significant amount of the artists entire catalogue to be streamable, not just one or two random songs.

6 Likes

Your ticket looks like a good start to circumvent what could be a big problem.

1 Like

I would also be inclined to say that adding URL relationships should not be auto-edits.
URLs aren’t primary data for a music databse IMO, and they have higher chance of abuse, so having them votable would be reasonable (and might cause cases like this to be flagged a lot earlier).

I disagree:

2 Likes

I’m in favor of this suggestion. What would the disadvantages be by making those always votable (especially beginners)?

@Freso: You know about my many spam reports. In case you didn’t look through them yet, I’m seeing a bit less spammers but they often do auto-edits.
It’s not easy making useful edit search queries for auto-edits. Not being able to properly use “applied edit count of editor” is a real bummer.

I tried using a whitelist of trustable editors who are adding many links but that didn’t work as well as I hoped. If the requested URL is too long the connection get’s reset. :frowning:

This is what I’m currently using to find spam and other bad URL rels.:
open, auto-edits and all URL edits [open]

If you want to see how much spam was removed by me use this search and look out for “spam” in the notes (forgot that note on a few occasions).

I’m also not reporting spammers older than one month and keep those in my bookmarks instead. Currently 206. I’ve also made a list with common spam keywords but I doubt it’s complete. Have seen Vietnamese and Thai spam entries.

2 Likes

URKs that are carriers for a known external ID (discogs, wikidata & co) are one thing (although I might prefer those being stored as the id (as is done for things like isrc, jasrac, …) but visualised as a hyperlink; for one thing, that means if a url scheme or domain name changes, you have only one place to make the change).
Free-form urls are something else. It’s nice to know where and artist’s website is. But that’s not nearly as relevant as their recordings, releases (and grouos thereof), and relations to other main entities.