Back in the day, chances are you pointed over to your physical collection of CD's, vinyl, cassettes, 8-tracks and other assorted music formats. A few years ago, you might also have started to point to your computer's hard drive and/or iPod, but now, well, your music collection is a bit more scattered among even more devices and services -- multiple computers, MP3 players, songs you downloaded into your cell phone, and maybe even your collection of music on a subscription service like Rhapsody.
Truth is, I've become a bit dissatisfied with how limited I am in how to organize and collect my personal music library/collection compared to the numerous ways in which I'm consuming music -- not to mention all the ways I'm discovering/finding/sampling new music. The notion that one must first "acquire" a song, then place it in a repository before it becomes part of a personal music library is quickly becoming antiquated to me. The mere act of downloading a song just to get it into your music repository and make it part of your library for easy access and listening starts to seem a bit tedious. I mean, I spend a fair amount of time finding new artists and sampling new albums from my favorite artists on Last.fm, Pandora, and Lala, but it's so much more difficult to include these finds in my music "collection" since they are segregated from my library of song files being managed by whatever is my music manager of choice (iTunes, Windows Media, SongBird, Winamp, etc.). Despite the industry and government's attempt to distinguish royalty rates between streaming (Internet radio and on-demand), downloading and all gradations in between, the general public continues on, consuming music in its various forms with little worry about which is which. Indeed, as I've mentioned before, as the world becomes more networked, near ubiquitous access to music anytime, anywhere and soon, anyhow, the delivery/storage method of consumption increasingly becomes irrelevant to the consumer.
Then there was the Yahoo! Music fiasco. For a year or two, I signed up for the Yahoo Music Unlimited service at a favorable introductory rate which was similar to Rhapsody's subscription and tethered download service. During that time, I amassed a hefty collection songs I "added" to my collection which basically meant I tagged/bookmarked them into my collection. Eventually Yahoo discontinued their service and opted to transfer subscribers over to Rhapsody. By then, the free alternatives like Last.fm's and Deezer's on-demand streaming offerings were maturing at a rate where I didn't really feel like getting on the Rhapsody bandwagon. Perhaps one of the most frustrating aspects of the whole experience was that this collection of music I had assembled simply disappeared and I was left with no record of what I had bookmarked and had to leave it up to my memory to recall all that music and search for it elsewhere. In my mind, even though I didn't buy or download any of those tracks, they still rightfully belonged in my library of cherished music. For all the anticipation of Nokia's "Comes With Music" program and the ability to keep a few of the songs I download after the life cycle of the service, if I chose to end it, I'm still sufficiently stung by the Yahoo experience to ever allow another company to exert such control over the size and shape of my collection -- and how easily it can be lost.
In an ideal world, your personal music collection and the way in which you find/collect (note collect can mean many things: download, find a stream, etc.) music should be:
- consumption-independent: including a song in your collection shouldn't be restricted by whether you consume it as a downloaded music file (nor its file format) or a stream
- platform-independent: including a song in your collection shouldn't be restricted by the platform you found it on, whether it's Rhapsody, Last.fm, imeem or Pandora
- repository-independent: including a song in your collection shouldn't be restricted by which repository the song is located (be it in the cloud, on your hard drive, or any other device)
By decoupling the notion of a personal music library from these three requirements, the personal music library becomes a completely metadata-driven representation. Of course, this concept is hardly new or earth-shattering. In fact, the XSPF playlist format demonstrates the concept quite well. On the scale of an entire personal music collection, the iTunes Music Library XML file comes even closer, but is restrictive about what song formats can be included. Again, varying file formats and music media managers' favoritism for differing file formats (and the exclusion of pure metadata sources of collecting music) derail the ability to unify songs into a single personal library.
What's also missing is a layer that can represent the various ways each song in a library can be consumed. For every song represented in a metadata-driven music collection, it can be consumed in multiple ways. Allowing a personal music collection to include as many consumption nodes for any given song allows the user to see and adapt their consumption based on multiple scenarios. As the user adds more consumption nodes (buying a download file, adding a subscription service, visiting new on-demand streaming sites), the way in which they can consume the song expands as well.
Think of a data portability concept like OpenID, Facebook Connect, OpenSocial that can be applied to your music collection where your songs are the friends/contacts in your networks and the additional social data available on each of your social network profiles provides additional data to access a song from that node.
Once it encompasses "everything", the different sources you go to consume the music becomes just another "node" which can be added or replaced with other sources. Having a music library format that embraces the various nodes or sources to consume the music (re: the where you "get" the song from and how it's being made available to you whether it's on your hard drive, located on remote server, as a stream) frees the song's requirement to be accessible from one source. Thus, the various sources expands the breadth of your collection's accessibility without limiting what music can be appended into the personal collection.
To at certain extent, you might envision your personal music collection as in the cloud, subservient to no device, service, website or file format -- but ingesting and updating music from these nodes nonetheless. In an interview with HP's Russ Daniels on Ars Technica, Daniels offers his vision of the cloud in sync'ing and saving data and although his analogy involves a user's contact list, the model would be quite similar to managing your personal collection of music:
"One of my favorite examples is, I have a car that has an entertainment system that supports Bluetooth so I can use it as a hands-free headset. But to be able to do voice dialing, I have to load the contact list from my phone into the car. Which means that the car and the phone both have to implement a complex Bluetooth protocol correctly, and they don't. It's very difficult to get those kinds of point-to-point complex protocols debugged in the context of the large number of devices that might need to interact.This is a powerful model and concept for music -- and one I'd love to see implemented in an open and portable architecture to counter the services any companies may develop and turn into a walled garden (while I like Gracenote and Omnifone joint-venture to save/sync music into your home and car, I'm curious to see how proprietary it will be).
But the whole idea is flawed because my phone's not authoritative for my contact information. The phone has a local cache of information that it gets from Outlook. But Outlook's not authoritative for my contact information, neither is Gmail, neither is my Vonage, neither is Grand Central, neither is the six or eight other things that I have in my life that think they are.
What none of them do is the simple thing of, "tell me the URL for your contact service." Additionally, it has to be a service, not a repository, because in fact the contact information that's relevant for me includes the global address list for HP, and I have to be able to have that invoked... I can't replicate that data and keep it synchronized, so I need to be able to use a federation model behind this single endpoint to answer those kinds of queries.
If you get that piece right, then if an application or device wants to cache that information locally for offline access, that's great. If one of them wants to proxy for another one, that's great, because those are all based on standard protocols and implementations—we do know how to do well.
So when I think about the authoritative data source, it should be in the cloud. And it's a service, not a repository, so I can deal with the complexities of the real world where there are many potential data sources. However, I have to get every piece of software that's involved to do the right thing—to delegate responsibility—rather than acting like they're independent owners and whatever they have in their local state is good enough."
Casting a Wider Net: Capturing Intent and Affinity into the Collection
To extend this even further to define what music belongs in one's collection, why wouldn't any song I voted up, favorited, or rated up during my experience on Pandora, last.fm or anywhere else also rightfully be included in my collection?
If you liked a recommended song from Pandora, you would do something similar -- you might even bookmark the song or give it a thumbs up. On Last.fm you might "favorite" an artist or "heart/love" a song, or add it to your personal playlist on MySpace. The subtle piece here is that, before you even buy a song to add to your collection, you've already performed actions that demonstrate your affinity for a song and your intention to add it your collection -- once you buy it, download it for free, etc. This affinity/intent vastly expands what one normally thinks of as a music collection -- previously, your music collection was a repository of sorts, now, your music collection is more synonymous with a collection of "all your favorite songs", regardless of whether you "own" it, or "subscribe" to it, or "added" it to a favorites list -- if you like the song, it should be part of your collection; the act of downloading or buying is really today's intermediate step to get it into your collection. Once you've demonstrated your affinity for a song, the only other crucial question is how do you want to consume it. As an MP3 file you can store it on a device that's available anytime? As a stream whenever you're connected to the web? How you want to consume a song is really secondary to the expression of your affinity to the song, or your explicit (or implicit) intent to have it included in your personal music collection.
Innovating On an Open, Portable Personal Music Collection Format
With an open and portable format, applications and services can innovate new and useful services for users on a level of personalization that would otherwise be less efficient. A great example are services currently leveraging your iTunes XML Music Library file as an efficient method to know your musical preferences. Prior to that, the other options were:
- asking you to input/search for your favorite artists one-by-one
- monitoring your listening habits
- upload your entire music collection into the cloud
- Lala can mirror your iTunes Library and put it in the cloud and give you access to your entire collection as streams over the web
- SonicLiving can provide you alerts, notifications and personalized calendars of when your favorite artists come to town
0 comments:
Post a Comment