With pressing plant entries and zillion live albums discogs is becoming a mess

The problem with discogs was introduced the day the schema was designed. There's no validation whatsoever (or at least wasn't for the first 5+ years); on "country" for example you can sometimes even see text like "No idea". It's not normalized, at all -- in fact the year column of releases you can see things like 92, 1992, 01 and even "a long time ago" or "who cares about time, man..."

Validation is happening more recently at the application level but certainly nowhere the database/constraints. This database is a disaster. To be fair, who could've predicted it would become what it is. Like craigslist and ebay a lot of these old sites just won the timing lottery. They didn't have time to plan things out.

I have a docker container that builds, indexes and makes some attempt to optimize their monthly community data dumps (i.e. everything but marketplace data which is no-copyright). My plan was to build a better release search engine with better matrix search, etc. Some have attempted to incorporate elasticsearch but it's not a real solution. I still have barcode search down to ~4ms but that only covers 60-80% of records released after maybe 1996 or so.

Matrix info seems to be improving. It's a work in progress by all involved.

 

Is it possible to add the music scraper for Albums and Artists from the website Discogs.com?

There is a great database for a very lot of albums/artists


Because the Universal scraper for Album and Artists finds not all my music albums!

I get more info for my music albums on Discogs.com

 

Those are all real albums, each with a different title.

I thought you were talking about small deadwax differences (to sort through) here. The CC albums are official bootlegs, radio show discs, and their real albums. If I cared at all about this band I would want to know about all of these releases.

I have some issues with Discogs, but inclusiveness is not one of them.

 

I correct things as I go along. I have not had anyone dispute my corrections, especially the year or country field. But I did have to call out someone on their mass editing I felt was removing data, and not adding anything to the discography of a label. We got into it, he provided links to support his position (but they were false, did not at all support his position), and some experts chimed in and agreed with me - I prevailed, he was left with his tail between his legs, had to move on. There are people on Discogs that sit down after dinner every night after work, and look for things to edit and remove. They really want to be known as an expert database contributor. Sometimes having nothing to add, so they start removing things, shortening titles, and just basically IMO vandalizing the site. I had to step in and call that one guy out. He was pretty sure that his friends were going to come in and support his position. They all at first ignored the invitation to comment, but I pressed on, One called it a "borderline case" but agreed that the editing was not adding anything, and when in doubt "inclusion" is the rule rather than excluding info. It was not borderline, I was right plain and simple.

Anyway, I just fix small things here and there.

One thing I want to do, and this is going to irk a few people, is the Gram Parker - Live at marble Arch UK promo on Phonogram is listed there several times, most all are counterfeits. I have the real one pressed in the UK. I am going to enter it if it has not yet been entered and add "Unofficial" next to the fakes that are trying to pass off their LP are the real promo. I think if I ad a comment "Not the official releases like "this one" is, and link my own entry with deadwax pics, etc. I might catch less heat. But I know some will insist that the fake is as real as the sunrise each morning.

 

As others have said, the problem with Discogs is not the amount of data, it's the way it's organized. Any trivial difference in packaging warrants a separate "release" (e.g. the position of the copyright notice or, literally, the color of a cassette's plastic shell), but all the master/stamper information is hidden in free-form text fields.

There's no ability to maintain consistency between different editions of the same album, so different entries for the same album will have differing genres, titles and primary artists.

A lot of the key data is based on external sources unrelated to the physical release in hand (e.g. whether it's considered an "Album" or an "EP"), and in many cases, users are specifically instructed not to enter the information on the physical release (e.g. "Country", "Year"), making it difficult to tell if the data on any particular entry is accurate or just copied blindly.

The "live album" problem is due to the fact that they make no effort to organize artists' discographies (irony alert) into any kind of usable format. Radio interviews, unofficial foreign compilations and every archival/promotional/fan-club release ever produced are mixed into each artists' main "Albums" discography, while artists like "Bob Marley" and "Bob Marley & the Wailers" are considered separate discographies and appear on separate pages.

And did I mention that artist discographies are unable to display albums from the same year in chronological order, even when the exact day of release is in the database?!

Add onto that, as others have mentioned, no restriction on what can be entered, no validation on anything that's entered and, in many cases, no way to tell what is supposed to be unique about each entry, and you wind up with a barely-usable database where (true story) it can take you an hour just to find the correct version of a Tommy James & Shondells album.

So yeah, it's got some problems.

 

Someone wants to know about those albums. But almost no one wants to see 150 live albums from one tour crowding out the data for a band that has released a handful of studio albums. They should be hidden by default. Same with pressing plant stuff.

 

Where is the line drawn on when to hide live albums?

Pressing plant info is deep, hideen, and not viewable unless you click on the title and look way down below.

 

Like porn, I know when the line has been crossed when I see it. When there are a zillion live albums from one tour, and those outnumber listings for all studio albums combined, that line has definitely been crossed.

 

You are sounding like a cranky old man. Really, there are 42 album titles under albums there, not that bad considering (they should be so lucky). In post #31 you mention 150 live albums listed. Now you are rounding them off to a "zillion". So which is it, 42, 150, or a zillion?

 

Here's the Counting Crows page.

They have six studio albums.

I'll leave it as an exercise for the reader to determine if the current presentation is more value than crap.

 

I’m not sure why digital versions are logged on Discogs. You can’t buy or sell them, so there’s no price/value info.

I guess if you just want to know something “exists” in the world, even though it’s just bits?

 

Sometimes a title that appeared on LP in the 1960s has appeared on digital download, fully remastered nicely, and yet no CD has ever been issued. So if it is the only release since 1966, I'd want to know about that before buying a slightly scratched-up LP.

 

But if it’s buyable, you could find it by searching the usual suspects as far as sites that sell downloads. If it’s not buyable, it’s kind of like trivia. Doing a general internet search will typically let you know a digital version exists and maybe lead you to a place that you can buy it.

But bottom line, I think it is just the shortcomings of the platform they use and inability to filter out what you’re not interested in or do more powerful searches that will lead you to the info you’re looking for. No doubt people use it for different purposes, and when you’re trying to find a specific version, there’s a ton of noise to sift through.

 

But don't forget that you can submit official physical releases, unofficial physical releases (bootlegs/counterfeits), official digital releases, but not unofficial digital releases. Because that's totally logical.

 

We need to know about counterfeits so that we can avoid paying top dollar for fakes. Bootlegs are very good to know about - especially if one collects them, but they are not legal to sell.

 

lol. You basically just described the purpose of Discogs.

The marketplace works off the database and not the other way around. Not everything that has been released has to be something that can be sold in the marketplace (some examples above).

 

yeah, you are right

 

no, I am having a little bit different question.

 

Odd post. I wasn't replying to anything you asked.