Gracenote Defends Its Evolution

In the beginning was a music recognition database called CDDB, and it was good. Now, people accuse Gracenote of stealing its success. CDDB and Gracenote architect Steve Scherf sets the record straight.

Very bizarre outcry from the techies...

I've never really understood why people are angry at GraceNote. If you put information out into the world, expect others to copy it. Expect some to take it and make it profitable. Expect someone to get some gain out of it that you might not be able to or even want.

Yes, there are various State-run ways to try to protect content or ideas (copyright, trademarks, patent, etc). These are useless for everyone but the ultra-powerful who can afford to litigate copyright infringement. Don't believe me? Try to battle someone copying your music, art or words.

My own sites ALL repudiate copyright -- I release it into the public domain, and even tell people to stick their own name on it. I make my profit two ways: I gain incredible information from the replies on slashdot or on my blogs or forums (that's free information from you to me), and I leverage that information into my "real life" of consulting and speaking engagements.

If you reply on slashdot, theoretically you own the content of your post. But how many people take your post and use it to form their own opinion? Who owns the newly formed opinions? In my mind, no one, ever. Sure, you may have submitted some CD information to CDDB, but who is to say that the information is unique to you -- and even if it was, who cares what CDDB did with it if you gave it away freely. Even if you put a restriction on it, how are you going to stop CDDB from changing its business model? If Linux all-of-a-sudden was ripped off completely by a big company and sold commercially, how would you fight it? With what funds?

What Grace Note did might seem mean or wrong, but I don't see a problem with it. People volunteer information for free all the time (see slashdot or any blog's comments). Other people use this and work hard to find value out of that information for others. It is the continued labor of working that is valuable to the market, not the one-time work that someone hopes to make repeated profits on.

Why did people submit data to cddb?

I submitted data to cddb. Why did I do it? Why take the time to type in the tracks?

Because I thought that I am submitting my data to the public. I thought that if I submit my data, so will others, and we'll have a public resource that everybody can use. But suddenly, that public resource turned private - I could not use it freely as before. They tricked me into giving them a resource, and then treated it as if it is their own property.
It is as if I gave a dollar to a public project - say a server to run slashdot on, thinking that if everybody contributed a dollar to that resource, then the public will have a resource - slashdot will have a fast server. And then slashdot suddenly turned around, took the $100k that people contributed, added another $100k from their own money, and said that now you can only access slashdot under certain conditions.
It is true that what they did was legal, but I think it was highly unethical. They for sure tricked me out of 5 minutes of my time.

Re:Why did people submit data to cddb?

Now, I don't remember what the original submission disclaimer for CDDB was, but I don't see anything legally -- or morally -- wrong with what Gracenote has done.

I see something ethically wrong with one thing Gracenote has done.

Gracenote has sued other companies (such as Roxio) that have used FreeDB, saying Gracenote owns software patents to CD-identifying technology. That so many people worked to contribute to a freely-available resource, only to have that resource closed and then have the closer use lawsuits to attempt to stifle competition came as a slap in the face. Now, this was five years ago, and maybe Gracenote has behaved themselves since then, but after that I chose to use FreeDB instead.

And no, Gracenote did not "release the database to the FreeDB," FreeDB copied a two-year-old mirror that had been made before Gracenote was formed, before it closed the database. Gracenote's position has been that the data was owned by them. In fact, they used the arguement that XMCD added copyright tags to each submission setting the copyright to the CDDB maintainer, copyrights which then passed to Gracenote when they were formed and said maintainer was an employee.

Re:Why did people submit data to cddb?

I have no problem with them using the info that I contributed for their own cause. I have no problem with people selling linux, or the wikipedia on a DVD. My problem is that they do not make the public part of their database - the part that was contributed by 1000s of users - freely available to the public. The fact that they only stole 5 minutes of my time doesn't make it any less of a theft.

If they had announced ahead of time "please contribute to our database, and eventually we will change the access rights so that only qualified clients can access the database.", I am not sure that I and other people would have contributed our time (i.e. money) to them. (and I mean client in the sense of computer program, not customer).

Re:Why did people submit data to cddb?

Exactly. It's comparable to Wikipedia and other public knowledge infodumps. In the case of Wikipedia, the whole thing is run by a nonprofit foundation with all its policies, including all the legal terms you're releasing your contributions under, out in the open for all to see. It's a pretty safe bet that the Wikimedia foundation won't all of a sudden charge mandatory access fees and get rich off of what its users have spent years building up in good faith.

Re:Why did people submit data to cddb?

I don't know how it works now, but if I recall correctly the big change that happened when CDDB became Gracenote was that they started insisting on a fee to be paid by application developers wishing to write code that would connect to their database. You would need some kind of unique developer key to connect, which cost money. This immediately meant that all the open source software that was previously able to use CDDB was suddenly without an information source for some of their functionality. It took some time for freedb to be set up with servers that could handle the required load, and for all the end-user application software to be able to redirect to those servers. This is the fundamental reason why a lot of people got very upset with Gracenote, because there was a time when all the information freely submitted to CDDB was unusable, as none of the pieces of OSS that people were using were able to access the data.

So, you may not pay money yourself, but that doesn't mean that money is not changing hands in order for you to be able to use that feature.

Q.

From the article...

This may sound hypocritical when you consider that Gracenote's own client software is closed source. To be frank, I have had little say in the matter of open-sourcing Gracenote software, so my opinions on the subject don't necessarily reflect that of the company.

How can the company be adequately defending itself if these pleasant comments are coming from a guy who's not really in charge at all? Having read the article, I have some respect for this employee, but it hardly means that Gracenote the firm no longer merits blame.

What's there to set straight?

I like the tone of "we did nothing wrong, it was the investors' fault, and besides, all our functionality and data are already available through freedb".

Yeah right, so the community had to duplicate a lot of the work that was "donated" to CDDB, while Gracenote profited from it without giving back. His point that the data before CDDB went commercial can still be downloaded is flawed; we're interested in what happened *after* you took all that hard work that you got for free and started charging for it. Besides, that's not "giving back"; that's "whee, we're making a boatload of money here, but hey, have some leftovers of the WORK YOU DID FOR US which we happened to leave behind!".

That's ok, I think the community did a good-faith effort and look how things turned out. I'd say no hard feelings, but I also don't think CDDB can expect a lot of community support or understanding in the future, pretexts and explanations nonwithstanding.

Well

If all else fails use C4
(Think about it for a minute and you'll get it)

The gift is a blessing to the giver

Why should someone be upset that Gracenote is using community donated data commercially? It's all still out there free (freedb? don't have linky). If you give something away (CD information, $5 to the bum on the street, winning lottery numbers), what the recipient does with it isn't your problem or responsibility. Either you are giving it away, or you are trying to elicit payment of some kind (without specifying what you want - should you be surprised that you don't get it?), in which case, you aren't giving it away. Anyone concerned that their data is being used has a problem with the entire concept of 'donate'

Was it good?

In the beginning was a music recognition database called CDDB, and it was good.

Anyone who has worked with CDDB would disagree. Jamie Zewinski provides a detailed summary [jwz.org] of its shortcomings. That someone steps forward as its "architect" makes me chuckle.

irrelevant

> We lost a lot of sleep over the situation (I did, at least), because it was clear we had to change
> or become irrelevant.

So what? An open source project is as popular as its users want it to be. I'm never going to pay to use that sort of service, because there's a free one out there, and it just isn't worth any money to me. Becoming closed-source and non-free is surely more likely to make it irrelevant, not less.

The REAL issue...

...is not so much that Gracenote took CDDB and closed it. The REAL issue is that Gracenote's contractual agreements with vendors like Apple (f.e.) preclude Apple (f.e.) from implementing a choice in track databases for iTunes.

You use Gracenote in your software, you're prevented by your license from allowing users to choose freedb.

That's suck turned up to 11.

What good is the original db being available, open, free if no one can realistically use it in the real world?

Ha ha

What's Gracenote, is it money from Graceland???? :^P

Don't bother... The questions are never answered.

Wired: You built your business upon data donated in good faith by your users.

Steve: blah opnion blah done before Ti Kan blah.

Wired: To charge them for the data that they sent in? Doesn't that seem wrong?

Steve: blah blah investors market blah FreeDB still exists..

Wired: But you forced the community to produce FreeDB as a last-ditch resort. It was a needless duplication of a huge amount of work.

Steve: blah not greed blah GPL blah blah.

I read that whole smarmy article hoping that we'd finally get a decent answer. No dice. It's just a bunch of wandering by a guy who has gone to the McNamara school of interviewing ("don't answer the question you were asked, answer the question you wish you were asked"). But it's easy enough to counter this trick: just keep asking the question that you want answered.

Wired, you let him off the hook easy.

Profit is ok, but screwing your supporters is not

I wish Gracenote all the best in making a profit off the data they collect. What I find disappointing and a betrayal of its own users/data providers is Gracenote's latest for-profit initiative to use its database to pursue those that the RIAA and labels very loosely (or often inaccurately) define as copyright violators. So, it could be possible that data I provide to Gracenote could be used against me by Gracenote to assist the RIAA or their cohorts in one of their heavy handed copyright suits. Nice...

I always wondered if it was based off my idea..

In 1995-1996 I was running a popular web site I set up called The CDPLAYER.INI Project.

It worked with the Windows CD Player / Media Player application which identified CDs as long as the tracks and titles were in an INI file in your WINDOWS folder.

People would e-mail in their albums as text snippets and I would add them to the INI which users could download. There would be a new version practically every day.

It hit the buffers when the file got to 64K, which was the maximum size of an INI file in Windows 95 - then it had to start being partitioned and the need for a custom application became apparent.....

Useless === Worthless

CDDB/FreeDB, and their ilk are essentially useless becuase they have allowed any jack-ass to submit their version of song titles, artists names, genre, etc..

One need only query one of these databases to see the huge variety (hardly any free of major errors) of entries for a single album - I recommend U2's "The Joshua Tree" as a case in point.

iTunes plays a big part as well

I wrote a little freeware app for the Mac (NetCD) which used the then-free CDDB, and its development ended as I watched the MacWorld Keynote where Steve first announced iTunes. I wasn't bitter and still am not... iTunes did it right, and I was happy to see it (and know it was free).

The fact that iTunes used CDDB (and they actually managed to engineer a different agreement that was better than what the rest of us developers had... probably because Apple paid Escient to do so) was what really ensured that FreeDB would stay on the sidelines. When the CDDB was free, there was no need for FreeDB; during the short time after Escient bought the CDDB and before iTunes came out, FreeDB was growing steadily but hadn't achieved enough fame to move ahead of the CDDB. When iTunes came out, Joe User, when asked where the track names were coming from, would answer "iTunes puts it there." The CDDB (and FreeDB) was nurtured by geeks and hobbyists; Escient's (and Gracenote's) version was/is used and abused by consumers.

It's a bit late, isn't it?

There was an appropriate time to apologize and explain Gracenote's side of the story. It was when Gracenote took all the information we had put into the database, locked our client programs out of using the servers, and made deals to get rich from our data. It's really far too late to do anything about Gracenote's reputation now.

FreeDB sucks (everything is 'Folk'), but I'll take it over Gracenote any day.

he doesn't deny anything

The chap completely accepts that he took user contributed information, then locked it up in a private database. How is this setting the record straight? He has explained his motivation for doing so, but the fact is he still took the user contributions to build his product. He had no moral right to do this. Once you are accepting contributions from users, you have to accept that users feel a sense of ownership and entitlement. Many years ago (while I was still at school in fact) I put the effort to type in some early drum'n'bass cds into this public database, in return for the effort it saved me. I never thought for one second that some chap would start charging for access to this information and I certainly wouldn't have bothered typing any information in had I known it would become a commercial outfit within a couple of years. If only I could remember the cds that I entered into the database, then I could demand that Gracenote remove my info from the database.

This is why GPL license > BSD license.

Full interview text

At the risk of being slashdotted, the full text of the interview can be found here [moonsoft.com]. I appreciate Wired putting out this story, but they (quite understandably) edited the interview very heavily. There's a lot missing, so some of the points don't really make it through.

What about legal threats and patent trolling?

I think the biggest problem with gracenote is that they will threaten any company using freedb in their product with legal action based on their evil patent portfolio where they basically patented databse lookups over a network. He tries to point out that they've let freedb go on, but really they've worked hard to take it down.

CDDB is dead

MusicBrainz [musicbrainz.org] is the future.

I'm curious....

...as to when this Frox thing referred to in the article came out. I helped my then pal Ian Giblin implement his idea for CD recognition and track storage in 1993/4 when he wrote his RiscCD application for the ARM-based Acorn Achimedes running RISC OS. I don't think he, and certainly not I, was aware of anyone else doing this at the time. Does Ian deserve some recognition (that I can vaingloriously bask in ;)? A quick google for frox just seems to refer to some transparent caching ftp proxy software...

Oh he can go sc3w himself...

"More importantly, the focus and dedication required for CDDB to grow could not be found in a community effort. If you look at how stagnant efforts like freedb have been, you'll see what I mean."

FreeDB has had problems from day one because Gracenote sued companies who tried to use alternate lookup systems. They sued FreeDB at one point over the database's content and raised questions over patent ownership and copyright ownership of the database. They've been complete bastards and he can go F himself over a 100% disingenuous statement like the one above.

Steve who?

I started using CDDB in the mid-90s, but I don't know Steve Scherf from a hole in the wall. The name I associate with CDDB is Ti Kan, and even wikipedia lists Ti Kan as the inventor of CDDB [wikipedia.org]. It doesn't say anything about Scherf being the "co-creator". IIRC, Ti Kan also had a really nice Audi Coupe Quattro that was featured in european car.

Re:It was good?

I think that the blurb was more of a failed attempt at invoking a bad cliche than it was factual. Face it - the writing and editing here have always sucked and will always suck in the future. You just have to tune out the bad cliches, bad grammar, bad spelling, and bad humor. That usually leaves you with a link to an article, and if you use Google you can usually find the actual information that the article is re-reporting. That said, if you can't beat 'em, join 'em: never RTFA and post lots of comments. Doing anything else will grind the system to a halt. :P

Re:Patent abuse

AC, patents aren't for data, these are for 'methods of information transfer' if i can generalize...