I would like to let you know about the current status of http://freedb2.org and provide a few hints about where it is going from here. First of all, there is now a news page at http://freedb2.org/news.html and as time permits I am posting news, documentation, and answers to frequently asked questions there.

Secondly, freedb2.org has just been subjected to a day long slashdotting which levied a total of around 70,000 queries against the database and peaked at around 4,500 per hour. At no time did the server load get out of the green zone, and feedback received from visitors indicated no perceived slow down in performance.

It is extremely unfortunate that owing to the events at freedb.org this week, it has been necessary to throw freedb2.org into service prematurely. However I am now confident that freedb2.org is capable of taking over the load while a long term solution is found.

Please help complete the beta testing by trying out your freedb client software with freedb2.org. To do this, substitute the freedb2.org domain for the domain which is currently used in your preferences (usually freedb.freedb.org) and select the HTTP version of the protocol.

I would appreciate any information that you can provide about problems that you encounter and I will also do my best to respond to any concerns you may have about events this week. As there is not yet a forum available on freedb2.org, please email me directly at dba@freedb2.org or post here if it is of general interest. Thank you.

Hello,
First, thanks a lot for providing this priceless service

So I went to try it in EAC but can't seem to get it working.. the original URL (which still worked as of now) was

http://freedb.freedb.org:80/~cddb/cddb.cgi

Replacing it with http://freedb.freedb2.org:80/~cddb/cddb.cgi didn't work though, maybe someone can help with that

Greetings and thanks again

It should be http://freedb2.org:80/~cddb/cddb.cgi

Submissions from EAC to that URL don't seem to work though.

Ah yes of course, that works wonderfully

An another note, I am not familiar how freedb works other than querying album information.
Is there any way to correct entries? The first disc I tried just now came up with 6 or 7 entries to choose from (why is that?), and the one I chose randomly promply had 2 false track titles.

Just if anyone cares it was this http://freedb2.org/?fDisc/1807315
The one who entered that confused sign with sing, and E5150 with something that makes no sense (if you know the meaning).

Unfortunately freedb2.org only supports submissions via email at this time. This was implemented first because I intended that freedb2.org would be used as a mirror of the old freedb.org server while it went through an extended testing phase. The extra code to allow freedb2.org to run as a master server was to be written during that time. Since a lot of priorities need to be changed now, I can bring forward implementation of submissions via HTTP to cater for clients like EAC if they don't support submissions via email.

There is also an ethical issue. It is my wish to release all submissions to freedb2.org into the public domain, just as musicbrainz.org does. However there should be some community consultation about that first, and I will also have to write a bit more code to follow through on my promise to make updates available as frequently as possible (in real time via RSS if I can get the bandwidth).

If the submissions are not immediately available then it would be dishonest and unethical to promise that they would be available 'real soon now' and still expect you to trust me. This is the main reason why I have not yet actually made a formal announcement that freedb2.org is fully capable of receiving submissions via email.

Would it be possible or desirable to have that community discussion here? Does EAC support email submissions? I need your advice.



I've been developing a web application that will allow everyone to edit and correct submissions via forms on the web site. In the meantime you can correct bad records that you find by editing them in your client software and resubmitting them via email. This is how the master freedb.org server handled it, but again I'll have to write a little bit more code before the corrected entries replace the erroneous ones in freedb2.org's server. Note that any submissions that you do make will be stored safely in the database until they can be properly processed by that currently non-existent code.

Now the $1,000,000 question is, does freedb2 overcome the poor design of collisions that freedb sufferes from?

In the long-term what about extra data (per-track genre, label, composer, etc)?

Finally freedb has a flaw, its most valuble asset is its users, but is also its flaw (wikipedia suffers the same with deliberate bad edits), would there be a review system to vet submitted data?

I also thought of a wiki- like system.
About the info submitted to freedb2 being public domain: I don't see any point in copyrighting the submissions, as the effort of creating them is rather small, and there is no intellectuel property to be protected. I never thought about my entries submitted not being public domain because I always thought of it as a community service.

There is a lot of info about the structure of freedb and ways to overcome the design problems on my site at http://asmith.id.au/freedb.html. Since all the data in freedb2 is in the form of records in a database there is no problem with collisions the way there has been in freedb.org. The old filenames and categories are stored so the records can be accessed using the obsolete CDDB protocol, but the records will also be accessible via unique keys which have no collision problems at all.

Freedb2 is a relational database built on PostgreSQL. Extensive design changes can in many cases be made very quickly and easily. After a bit more community discussion we should be able to arrive at a very useful set of enhancements to the freedb database design.

I've devoted a lot of time to writing software that analyses and optimises the entries in the database to remove spelling errors and inconsistencies in capitalisation and punctuation. Once things are up and running smoothly I'll be resuming that work so you should see a steady improvement in the data quality. As I mentioned above, it will be possible to edit the records to correct mistakes, directly on the website, but some form of moderation would certainly be a good idea.

>remove spelling errors and inconsistencies in capitalisation and punctuation

We went down that road 4 years ago, changing the data coming from freedb as you mentioned, it is a difficult one - in the end there are tracks which will be affected by your routine which should not be, it could be 2 in every 100 cds is affected in an adverse way, atleast with the implementation in the client that option could be switched off, on the server the changes would stick.

Is Joerg on your 'team'?

Seconded. Lots of my CDs have track names in languages such as German, Italian and Latin. I also have CDs where the name of the composer of a track might be different depending if you anglisize it or not. I also have tracks where the initial letter should not be capitalised or where I really do want capitals in the middle of words.

All of these sorts of things will get corrupted by an automatic spelling and sanitisiation routine. I strongly urge you to commit to the database "as is" but use your routine to flag up records with potential issues so they can be reviewed by a human being.

About copyrights

By no means does writing the track names from CD's back cover to computer constitute a work in such a manner that it would be eligible for copyright. On the other hand, inveting the track names would fall under copyright at least if it's combined with actually doing the songs too. So go figure. I can see no problem with people who send track names. That isn't anything original. A common definition for a copyrighted work would be that no person in the same situation would end up with something similar.

There is actually a decision made here that states that a line drawing depicting a sheep isn't copyrighted as anybody would draw something with great similarity. Though the booklet which included the drawing was considered copyrighted as a whole.

How are the character sets done? Can any submitter choose his own or are you forcing UTF? Any strict policies on that?

I have been using a different approach to improving the data where the database is being considered in its entirety rather than just one record at a time. By taking this approach it is possible to put the large number of duplicate records in the database to good use and treat them as "votes" in an automatically moderated system. Obviously it isn't perfect either but it is possible to safely make repairs to several hundred thousand records this way. I should also mention that freedb2.org keeps a complete change history of every record, so even when changes turn out to be undesirable they can be reverted to an earlier version.

When I last spoke with Joerg on Monday he had decided to take a very long holiday from freedb. He has devoted most of his weekends for the last six years to running freedb.org and understandably, he has had enough.

Will the freedb2 code become open-source?

Edit: Sorry, just read the news page and saw that you are already offering the source. Good!

I have converted as many records as possible to UTF-8, certainly all the records which were encoded as LATIN1.

However there are still about 30,000 records which were in some other 8 bit encoding. To properly identify the encodings of those records will require a number of techniques including the incorporation of language dictionaries into the database. Some preliminary work was done with very promising results.

If at all possible, please submit all new records encoded as UTF-8. Where records are not encoded as UTF-8 I hope to be able to retrieve the encoding information from the received email submission and make the correct conversion to UTF-8 before importing the record into the database.

Yes it is and will be, and just as soon as I can arrange it. I've detailed the specifics of what will and won't be open sourced on the news page at http://freedb2.org/news.html and I hope you will find some interesting things there to get started with.

I am the sole author of freedb2 and have been writing it since December 2004. It is my intention to release almost all of it under the BSD licence as soon as possible. Releasing the data is currently a bit problematic because I simply do not have the bandwidth available to upload the data and operate freedb2.org at the same time. Once freedb2.org has been moved to the new server provided by http://commandprompt.com I will be able to upload the data to a file sharing site, or start seeding a bittorrent.

Folks I'm going to have to sign off for a while and get some sleep. It's been days... Back in about 8 hours.

Thank you, knight in shining armorA, for saving the DAE-community out of its misery.
You may don't notice it constantly, but we ARE fucking thankful for that ...
I'd like to support you, but I don't got the time, yet, I'm moving to another city...
guess that's a double bad timing according to the end of freedb then ^^...

Freedb2 is not meant to be a long-term solution? Long-term being a relative term these days as evidence by the end of freedb.