Make Discogs a Wiki!

As I mentioned the other day, I finally joined Discogs, following my surprise discovery of some of my earlier work on a listed Polish CDR. I’ve enjoyed browsing the site from time to time, but had generally stayed away– Discogs always looked like a rabbit hole I might spend an inordinate amount of time adding to.

I took some time to create a full release entry for one of my earlier albums, the same one that had been the source of tracks for the Polish compilation. If you haven’t used Discogs, let me describe it in a single word for you: ANAL. discogwiki.jpg

Not that this is necessarily a bad thing. When you’re serious about looking up an exact album, especially older stuff, the difference between a track being 2:33 or 2:45 can make a real difference! And I have to hand it to them, Discogs has done an admirable job of creating a useable, somewhat-friendly interface for entering all the possible information about albums in. Still, given the amount of time and effort it took for me to enter my own album in– one I thought I knew pretty well, ha ha– really opened my eyes. When you start imagining a site of this size tracking song lengths, cover art, limited editions, formats, stereo information, etc… well, there’s a lot of opportunity for mistakes to be made!

Even I made some initially. I put “40” for the number of CDs in the release itself, thinking this to be a space for the number of individual editions manufactured. I tangled up a performer’s alias, and used a slightly incorrect format term where another would have been more appropriate.

The good thing was that later that day, I was discussing my first Discogs experience with some fellow SLSK Noise chatters, who noted these errors and helped me fix them. The bad thing?

I found out that my entry probably won’t be seen by a Discogs moderator for around four months!

Yes, you read that right. At this moment, on the same internet that you and I use, a person can enter information into a popular site receiving hundreds of thousands of visitors each day and not see that information made accessible to the general public for months! It’s almost hard to believe, and makes you wonder what the problem could be. According to the Discogs News forum, problems with having too many newly-added releases in queue (in other words, waiting to be checked-over by a moderator before entering Discogs publicly) had become so great that Discogs management “put the brakes on” the ability of users to make new submissions– whereas previously, users were granted three “points” good for submissions and updates for every full new entry; now members found their points suddenly and drastically reduced. Why? To put it simply, the 300-ish moderators couldn’t keep up with the sheer amount of information being presented to them!

Discogs Community Manager “Nik” said:

The length of the queue is more or less directly related to how long it takes for a submission to be processed. There is absolutely no point in throwing more and more submissions onto the queue if we cannot process them (and the queue gets longer and longer). Last September the queue was just over 20,000. By the start of 2007 it was 30,000. It rose over January and February to 40,00 at the end of February, and was continuing to climb. That’s double the queue size in six months. And the waiting times have risen proportionally. To do nothing was not an option. This is a pill that we have to swallow in order to manage the queue and waiting times for submissions.

Is it?

I wonder what would happen if Discogs were re-invented as a Wiki. Rather than rely on the “expert” opinion of a small group of volunteers, why not let group consensus prevail? I can envision the moderators still having a place– even Wikipedia has it’s stewards and admins– but otherwise, let the public add as they will. It’s hard not to feel for users like “Stoffeler10,” who seem to enjoy adding to the wealth of information, and now find themselves cut off:

“Sorry nik, but this is sick. People (like me), who are submitting for years now, spending hours for this data are now able to submit 7 releases. Thank you very much. What an award for all the work.”

or “Todeskult”:

“Sorry nik, but this is bullshit: I’m experienced submitter (482 new entries done, not counting the edits), now I got only ONE new release in queue and I’m unable to move 2nd release from my drafts to my pendings! It’s not fair!!”

I agree, it really isn’t fair. It doesn’t even seem like a particularly viable fix. Common sense says that as the site grows in popularity, the amount of submissions will increase. Although Discogs is scrambling to add more moderators (ten more per month, according to one of Nik’s forum posts) the problem will most likely continue– it’s hard to believe that a supply of volunteer “experts” is an inexhaustible resource!

That’s where making Discogs a Wiki starts to make sense. If you’re just going to continue adding users with privileges to keep up with the tide, why not just let everyone have these powers? With marketplace sales of wanted albums as a driving force, the Discogs community has a vested interest in keeping information accurate. However, under the current system, users are forced to choose between spending points correcting small, known errors, or adding new content!

User “Jester1973” writes:

“My sole complaint about the current restrictions are that I have several Edits pending with single obvious errors noticed after the fact or by moderators. I have no issue cancelling and resubmitting these, but being over my limit, I actually can’t. This means leaving Edits with errors easy to correct sitting in my personal queue waiting for restrictions to be lifted so I can cancel and re-submit correctly. “

Besides, moderators are only human. Getting the “expert” label doesn’t necessarily make one so. Think back to my own submitted release. If anyone should be an expert on it, I should, right? But I initially made mistakes with my own material. I can only wonder what an “expert” moderator will think of a limited-edition 40-count dual CDR (one audio, one data) release packaged in 40 unique LP jackets taken from popular artists, and containing that artist’s original LP as a gift! In short, it’s an “expert’s” nightmare– but the perfect opportunity for the Wiki structure to shine.

Users can question entries, update at will, and sort potential corrections out. Users are free to contribute, without worrying about having enough point capital to “spend” in order to continue adding to the growing wealth of information. It sounds great to me– please, make Discogs a wiki!

Update: Discogs user “Doctor Trance” writes in the forum:

“I’m for this, as you would know if you’ve read some of my other posts/threads.

“Nik’s thoughts in that other thread was that he pondered it, but thought their might be too many duplicate and incorrect releases, but if you took all the mods right now, and made it their job to look for discrepancies like that, it could work. In addition to that, only allow users with a certain rank should do it, not just everybody. It would be a slight step above Wikipedia, in that not everyone could submit just anything. With nik’s new submission limits going according to how well users submit, the new ranking system could only allow not only the most experienced submitters, but also the most accurate.

Wikipedia is also a database about everything under the sun: theories, biographies, definitions, etc, including many things that aren’t necessarily factual based. On this site, we are only adding factual information found printed on musical formats, so it’s not rocket science.

It’s actually the RSG that prevents things from going in smoothly. If the site were simply artist name, release name, and song listing, the place would have been twice the size it is now, with twice as many releases here. After all, these are the 3 top priorities on anyone’s list in searching for music. We’ve gradually allowed every credit, format, period, and semicolon into the system, that appears on a release, so it makes it more difficult to just jump to wiki style. If Discogs had remained in a more simple form, it may have already become a wiki site, not to mention a more gigantic one.

I’ve also always been in favor of wanting to see all releases by one artist in here, as opposed to only half, but a perfect half in that it is as accurate as one can get with those that are here. Before submitting here, I use to come here thinking that every electronic artist in this DB had only ever released stuff listed on their Discogs page. I was always disappointed when I would learn from other sites that these people actually had much more than what Discogs would show. Then coming in as a user, I found what the reason is why artist pages are so incomplete: as Kergillian puts it, the “tightly reined” database.

The site will never truly become a complete database unless it goes to a somewhat wiki style. Even with the huge queue, there simply aren’t enough users with enough releases in their hands to get every artist’s release in here.”

In another, older thread, Doctor Trance shows some evidence that Discogs may indeed be moving towards something like an open wiki model:

“The amount of Y-votes use to be 4, then went to 3, then to 2, so 1 is the next logical choice before letting someone have full control of submitting their own stuff. I’m sure there were those who said no to reducing it every time, but the site seems to have pressed on just fine with each reduction.”

I’m finding this to be a truly fascinating discussion– a collision of volunteerism, public use, private enterprise, information security, and the internet! It’s very much a mirror of many similar questions facing nations today. Here’s an excerpt from my last comment at the Discogs forum:

“I can’t help but notice the similarities between this discussion and the current security-related debates going on in the US… I can’t help but wonder if folks like Killaswitch and Perham were among those happily calling for the bans on liquids at airports, evil light-brites, and deadly toenail clippers….”

10 Responses to “Make Discogs a Wiki!”

  1. Kergillian Says:

    That would be disastrous. Sites like Wikipedia are among the most unreliable in existence, as all information is questionable, most added without sources, half of it plagiarized, and often the entries are hijacked to suit random users’ personal agendas.

    What makes matters worse is that far too many people now use Wikipedia as a form of information bible, using it to provide information for other places – even as news and database resources, which corrupts the entire information system.

    Whenever you allow people to add information to any form of database without any leash or ‘security check’ [i]before[/i] that information becomes active, all of the information you’re entering loses its validity.

    Granted, even in a somewhat closed system like Discogs, errors sneak in, or database-wide system changed make updates necessary as time moves on – but the number of those errors and change-related updates would increase a thousandfold if random people could simply add and update information at a whim.

    We also already have a few problems here and there with people movng information back and forth, or updating information back and forth. In a wiki, these issues become hourly battles – or less.

    What makes Discogs so special and reliable is the fact that it [i]is[/i] so tightly reined – the information is widely used because – on the whole – it’s so reliable.

    The new system should speed things up. The queue is already down to almost 25,000, and is falling farther every day. The average wait time for submissions is also dropping quite rapidly. Soon, the submission levels will rise again, and the queue will be more controllable, and the response times therefore faster.

    Remember that there re hundreds of millions of releases out there to be added, even more when Classical, World, Children’s and other genres are added. Better for them to take a little time than for them to have to be constantly monitored and corrected.

  2. Kergillian Says:

    (and don’t forger that pending submissions are visible to the general public – so you don’t have to wait until the releases are ‘accepted’ for them to be seen and accessed by DIscogs users..)

  3. phish63 Says:

    just read this and as kergillian stated above…..horrible idea. discogs DOES NOT need to become like wikipedia. i will relate this to a t-shirt i designed for a group in college… said “we are not cocky, we’re just better than you”. in this case, WE is discogs and YOU is wikipedia.

  4. startlingmoniker Says:

    Uh-oh. Phish fan. *klaxons sound*

  5. startlingmoniker Says:

    Kergillian– You might want to read some of the studies done about Wikipedia’s credibility. Check out this one, where the author states:

    “Staff were either given an article in their own expert domain or a random article. No difference was found between the two group in terms of their perceived credibility of Wikipedia or of the articles’ authors, but a difference was found in the credibility of the articles — the experts found Wikipedia’s articles to be more credible than the non–experts. This suggests that the accuracy of Wikipedia is high. However, the results should not be seen as support for Wikipedia as a totally reliable resource as, according to the experts, 13 percent of the articles contain mistakes.”

    Actually, Wikipedia itself has a large section addressing its own reliability, with links to many interesting studies, many of which show Wikipedia to be suprisingly accurate.

    Let’s contrast this with Discogs, though: if a moderator makes a mistake, he or she may no-vote the entire entry. Instead of allowing many users to pound it into shape, it’s simply gone. Also, any open wiki can run circles around a moderated system– if a new album comes out at noon, an open wiki can have the same info posted directly thereafter, whereas Discog’s moderated setup makes them wait for quite some time. If the entry does go up immediately, it would certainly only be by “cutting” in line, and having a moderator pick and choose which topics to allow first. This isn’t very democratic, nor is it good for rapidity.

  6. Obscurica Records Says:

    Hm, it’s been changed, but for awhile there, the Stimbox wikipage site read thusly:

    Sinec then, the revision has been revised. Nothing is keeping someone else from reverting it back. And forth, and back. I think the above entry is the single best example of why Wikipedia is inherently unreliable, and no moves should be made to democratize Discogs in that particular direction.

  7. inexpressible Says:

    Not much to add to Kergillian’s statement.

    IMO Discogs wouldn’t work the way wikipedia does. I understand some people have a problem with their subs staying in the queues for months (which is NOT the standard btw.) – but the result speaks for itself: Discogs is the #1 music database. And patience used to be a virtue.

  8. Kergillian Says:

    1) Random internet articles about the validity of an internet encyclopedia hardly convinces me. Especially since all ‘scientific’ evidence I have seen – even from print journals – will be looking at conrcete and/or niche topic articles, and will have a sampling of so few articles that it doesn’t gve a proper look at the whole picture.

    Example: Nature Journal’s ‘scientific proof’ comparing Wikipedia to Encyclopedia Britannica looked at 42 articles on science in each and compared them. FORTY-TWO articles out of 1.5 MILLION. *And* they were all on similar, scientific topics.

    That’s hardly an effective sampling. That’s like doing a topical survey on politics by going to the Democratic and Republican conventions and asking ten people at each who they’re going to vote for, and then writing an article about how the nation’s voting habits show a dead heat in the presidential race!

    That aside, Discogs is not an encyclopedia. We don’t write articles. We track information into a database using a precise formula (ie: various fields to track the information).

    Here are just a FEW of the errors that we already need to worry about, but which would be exacerbated a thousandfold in a wiki:

    1) People simply writing the artist name down without checking to see if a variant name exists in the database already, making a half-dozen or more artist pages for a single artist.

    2) People marking down the name (ie: John Smith) without checking to see if it’s the CORRECT John Smith, thereby completely mixing up the artist pages for the common artist names.

    3) People mixing up label names so that there will be dozens of label pages for the same label, with no connection to one another

    4) People screwing up cat#s so that there is no sorting, or so that there are multiple, identical releases due to cat# errors. Also, people writing the wrong thing down as a cat# (like the matrix #s or UPCs)

    5) People adding duplicate releases because they mistakenly perceive that they’re version is unique, or don’t realize it’s already in the database.

    6) People skewing, mixing up or leaving out Role information, or doubling up roles due to a msiunderstanding of ANVs and Aliases.

    7) People mixing up ANVs and Aliases creating a mess of the Artist pages, putting Aliases in the Artist’s page, or creating new pages where there should just be an ANV.

    Those are just a FEW of the common problems that a Wiki would create, thereby corrupting the entire database. And that’s not even including typos (which are ESPECIALLY bad in a database due to the necessity of artist, role and label name precision) and grammatical errors.

    And then there is the problem of websubmission. The rule of ‘don’t submit unless you OWN this release and hold it in your hand’ is not infallible, but it is a very good principle to hae to ensure the submitted information is correct. It would be IMPOSSIBLE to enforce in a Wiki.

    The Discogs system is constantly being improved, with a few speedbumps along the way for sure – but the system itself ain’t broke and don’t need fixing :)

  9. dmaxx Says:

    I completely agree with Kergillian here. Making Discogs a wiki would be an enormous mistake. Most newbies are just not capable of making correct submissions.

    I rather have a database with less entries but trustable, than a database with more entries but totaly unreliable.

    I believe that Discogs will be (close to) complete someday, but in meantime it’s best to just remain patient. Take it one release at a time.

  10. 0x00 Says:

    Rather than a wiki, I suggest dmoz as a model. Nik is (or was) an editor there and I think he likes their autonomy and trust metrics. You apply to edit, get assigned a small category (on Discogs that might be a label), then with experience and ambition you can show your editing log to a higher power and be assigned a category further up the tree. We don’t have trees but a 3 or 4 tier hierarchy is still possible: a group of labels, then a genre, the whole database… you get the idea. It’s nicely scalable while retaining good control, in contrast to Discogs and Wikipedia respectively.

    It was mooted (a good while ago now) that mods/editors could make small fixes to pendings and add them to the database without further user interaction. Creating ‘power mods’ was a concession in this direction, but still the bottleneck persists at exactly the same point, quite needlessly. They’ve ploughed on with the original submission queue while adding genres, adding linkable roles, walking blindly into a combinatorial explosion. Restricting submission limits is an embarrassing but inevitable band-aid at this stage.

    Nik is amenable to good ideas, and I’m sure his diplomatic skills have been well honed these years of explaining/covering Kev’s incremental, creeping featurist, blind-alley, bottom-up design shambles. Good luck. Moderators don’t last long before burning out at Discogs, not by chance.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: