Amazon cake update

Yesterday evening I had a call with my contacts at Amazon and a person from the Accounts Payable department. Over the last two days they were able to work out the kinks in their accounting with regard to the MetaBrainz Foundation.

They verified that the outstanding invoices from our perspective were correct, including the now infamous invoice #144. Shortly after the call, a check for $22,500 was cut and will arrive in California by 10:30am today. I’ve also received a complete history of all of the payments made to the MetaBrainz Foundation and I’m happy to say that everything looks great to me.

Also, an issue surrounding this invoice was pointed out to me: If invoice #144 wasn’t outstanding, all of this would’ve been a bit drastic. I agree with that, but what I failed to mention in my last blog post was that if invoice #144 wasn’t paid, then invoice #200 (which is 3 months younger than #144) was outstanding. The matter at hand was that there was a 3 year, or a nearly 3 year, old invoice outstanding. Personally, I really wanted to gain clarity around what happened nearly 3 years ago and finally put the issue to bed.

Finally, I would like to commend Amazon on how they handled matters in the last week. When prodded hard enough, Amazon got their act together, got to work and figured out this mess and then swiftly cut a check. I’ll keep posting updates about this until the check is in the bank and has cleared, but it certainly looks like we’re heading to a complete resolution of this matter early next week.

Thanks Amazon!

Using cakes for social engineering

For the past few years we’ve had some accounting difficulties with one of our customers: Amazon. I have no idea how their accounting and vendor systems work, but apparently we ended up in their system 4 different ways. And payment methods were horribly confused — it was a mess all around.

Invoice #144, for our Live Data Feed service that Amazon subscribes to, has been outstanding for almost three years now (it will be 3 years in January, but I wanted to get this money to come in in 2013). To be honest, there could be some confusion on our or on Amazon’s part — in fact the invoice may not be outstanding anymore. The fact of the matter is that we can’t seem to figure out what exactly is going on, but no money is flowing from Amazon to us and we’re owed somewhere around $20,000. (Which is near 10% of our annual budget incidentally)

For the last 6 months I’ve stepped up my pestering to get this resolved. I’ve been assured progress for the past 6 months, but nothing has happened. Promises of progress, then nothing. Again and again. I finally had an idea how to make change happen: Send Amazon an anniversary cake and post a picture of it publicly!

I even told my Amazon contacts about this idea, but it didn’t really catalyze anything. Then I finally set a deadline of Dec 2nd. The deadline came and went with more unfulfilled promises, so on December 2nd I picked up the phone and ordered a cake. Larsen’s Danish Bakery in Seattle were quite lovely to work with and created this cake for us:

Invoice #144 cake

A friend of mine went to the bakery, snapped this picture and then delivered the cake to Amazon HQ. It was accepted at the reception with promises that it would be delivered to its recipient. Then we started tweeting and Cory posted an entry to BoingBoing “Charity sends Amazon a cake celebrating 3d anniversary of unpaid invoice“.

For almost 24 hours nothing happened, but then I got email from my contact at Amazon telling me that the Accounts Payable team found a problem that was blocking payments from being sent to us, that the problem was now fixed and that they were investigating means to prevent this from happening again.

My contact goes on to say that a check will be cut tomorrow an overnighted to us. And that I should expect one more email telling me who on Amazon will be managing our relationship going forward. And, I have a voicemail pending from a person at Amazon’s accounts payable team to finally resolve the matter of the 3 year old invoice.

Sending this cake was quite effective! For $30 I managed to wake up Amazon, send a clear message that our account was not being managed well, that their AP team has some issues to address and that I wanted to fix our relationship. From where I stand, I see that these issues are on track for being resolved. Thanks for stepping up your game, Amazon!

Finally, I would like to say that all the people I’ve dealt with at Amazon have been polite and were honestly trying to help me. The real reason, from what I can tell, is that Amazon employees are constantly overworked and that MetaBrainz is such a small organization that its hard for them to really find the time to manage this relationship.

I’m glad that Amazon didn’t just cancel their contract with us and I’m looking forward to an improved working relationship going forward.

Annual report for 2012 finally posted

I finally completed the 2012 annual report! This year has been busy, so I apologize for finishing it this late.

Our cost per 1M web hits dropped significantly, we finished our first year in the red and we created 1/4 of all of our edits to date in 2012! Go read the report to find out who was the top editor, the top voter and other interesting tidbits about MusicBrainz in 2012.

Thanks to Navap, Nikki and Reosarevok for helping in putting this report together!

The BBC unveils a service that tweets to artists when their music is played…

… and it is built using MusicBrainz data! Michael Smethurst, a good friend of MusicBrainz, hacked up this service in the space of 2 days and writes:

The original idea came from a friend whose music occasionally gets played on Radio 1, 1xtra and 6Music. Almost always he missed this and either found out later from a friend or never found out at all. But he does use various bits of social media (including Twitter) to make contact with fans and promote his releases and live appearances.
. . .
To power online music services such as BBC iPlayer Radio, Playlister and /music the BBC uses metadata provided by MusicBrainz, a community maintained music encyclopedia. If you use Twitter and you’re a music artist or an agent or a publicist or similar and would like now playing notifications you need to check that your Twitter account handle is in MusicBrainz.

Thanks for creating such an awesome service, Michael. I know MusicBrainz contributors love how the BBC uses their data — I wish more people made such creative use of our data!

Fire damages the Internet Archive

A fire at the Internet Archive (our friends!) has caused $600,000 in damage. Fortunately no one was harmed and no data was lost:

A fire at the Internet Archive’s San Francisco scanning center has destroyed an estimated $600,000 of digitization and scanning equipment. Fortunately no one was injured in the blaze, but the property damage has ruined “some physical materials” that were yet to be digitized, and restricted the nonprofit organization’s ability to record the history of the web.

MusicBrainz just donated $50 to the Internet Archive and asks you to consider making a donation as well.

VMWare image of 2013-10-14 release available

I’ve released the VMWare version of the 2013-10-14 schema change release:

This VM is 8.7GB large and is built on the latest version of VMWare (Fusion in my case). If you’re using a VMWare product, then use this image.

The documentation for this VM has been updated to reflect the latest changes. Please make sure to read this page while you wait for your download!

New VirtualBox image of MusicBrainz server 2013-10-14 available

I’ve released a new Virtual Machine of the MusicBrainz server, based on the recent 2013-10-14 schema change. You can download the updated virtual machine here:

This VM is 10GB large and is built on the latest version of VirtualBox. Due to continuing incompatibilities between VMWare Fusion and VirtualBox, I am going to create a separate VMWare base virtual machine soon. I’m waiting for a vagrant-vmware-fusion license key and once I have that I will build the VMWare version and upload that as well.

The documentation for this VM has been updated to reflect the latest changes. Please make sure to read this page while you wait for your download!

Search server jar/war files

We’ve been asked to provide instructions for how to upgrade a search server installation. We’ve got two answers for you:

  1. Short answer: Use the jar/war files linked below. Deploy these into whatever setup you’re currently using.
  2. Long answer: Check out the source (svn rev 13728) of the search server and follow the install instructions to build your own jar/war files.

Links: indexer jar, servlet war

Search server update for 2013-10-14 release

In conjunction with our main server release we’ve released a new search server. Thanks for your hard work on this release, Paul!

This release contains changes to match the Autumn 2013 database schema changes plus some bug fixes and improvements mostly to do with artist credits and aliases. See the detailed list of issues that are fixed in this version:

Bug

  • [SEARCH-46] – Searching for a release by releasename and artists english name will give no matches
  • [SEARCH-122] – Recording Search only adds release artist credit if different to recording artist credit and VA release
  • [SEARCH-304] – Webservice Json output for aliases when searching is inconsistent for boolean values
  • [SEARCH-311] – Recording, release and releasegroup search results incomplete artist aliases
  • [SEARCH-324] – Release search should return catalog numbers in ascending order

Improvement

  • [SEARCH-132] – Allow to filter release search by quality
  • [SEARCH-242] – The searcher can’t guess that the artist name one’s looking for is alias or artist name ☞ look for both
  • [SEARCH-288] – Include alias sortnames when searching works

Task

2013-10-14 schema change release update instructions

As promised, here are the instructions for updating any instances of MusicBrainz you might have. You will need to perform these steps to upgrade to the new version:

  1. Take down the web server running MusicBrainz, if you’re running a web server.
  2. Turn off cron jobs if you are automatically updating the database via cron jobs.
  3. Make sure your REPLICATION_TYPE setting is RT_SLAVE
  4. Switch to the new code with git fetch origin followed by git checkout v-2013-10-14
  5. Run carton exec -Ilib — ./upgrade.sh (or simply ./upgrade.sh if you aren’t using carton, such as for VM users or fairly new installs).
  6. Set DB_SCHEMA_SEQUENCE to 19 in lib/DBDefs.pm
  7. Turn cron jobs back on, if needed.
  8. Restart the MusicBrainz web server, if needed.

If your server fails to start or cron is having issues, it may be a missing perl module problem. To check for any missing perl modules, follow the instructions in INSTALL.md.