Category: .

The Echo Nest web music API is fairly awe inspiring. For example, they have a web service to detect the sections of a song — bridge, chorus, etc. — given only an MP3. What it amounts to is a bot capable of doing semantic analysis of a sound file.

One thing you could do with it is to distinguish spoken word from music files. A hosting service could then automatically reject music but allow talk, which would control its legal liability.

All in all it’s a radical level of power for the kind of lightweight apps that use third party web services.

One complaint so far: the web service returns XML, but the XML isn’t in a namespace. This will cause pain down the road, and it can easily be alleviated by adding an xmlns=”http://echonest” attribute to the root element.

This is punk rock:

One of the grooveshark folks stopped in at the comments on my post about their not giving attribution to link sources to say that this makes no sense because they don’t link out at all. They’re not a search engine like Seeqpod, they’re a hosting site like YouTube and (originally) iMeem.

I can only cop to retardation, though I swear Hypebot made me that way.

Groove Shark is a search engine, yeah, in that it indexes free range MP3s and helps searchers get to them. But it never exposes the remote URLs, much less the originating web page.

As far as a searcher is concerned the MP3 is hosted by groove shark. When somebody hosts an MP3 they almost always want to be acknowledged, and more often than not they’re doing it to generate page views as much as plays. For grooveshark to sabotage their goals is bad faith.

Why be rude when you can so easily avoid it? How cynical do you have to be to care that little?

Seeqpod also made it hard for searchers to leave their site, but at least they gave attribution to their source hosts by showing the remote URL.

Greg Borenstein in the conversation about the apocalypse of the media industries:

The future composts the past. Someone said that once. Things don’t just go away, they get re-purposed and recycled with new cultural uses. How long ago did TV kill radio? There was a time when families would gather round the crystal set for an evening’s entertainment. What about movies. There was a time in the teens and twenties when more than a million people went to a movie every week. More people bought tickets for Gone with the Wind in American theaters than any other movie since. Radio’s still here; movies are still here. Radio became wallpaper for the car. Movies were for date nights, then for big costumed openings, like 17th century balls.

The confusing thing is that when we think of a medium, we tend to combine the physical format and the social function into one thing in our heads. So, when the social function or meaning changes without an equivalent tangible change to the format, we get confused. The medium has died, but still lives. Zombie Media.

But these changes are amazingly common. While we can argue about whether the internet is a bigger deal than TV or the printing press or fire, we’ve gone through other Zombifications pretty much every twenty years like clockwork for most of the recent past. Let’s just take movies: in the 90s, the hyper-blockbuster went global; in the 70s, the studios went bankrupt and VHS was invented; in the 50s, the studios went bankrupt and TV was invented; in the 30s, the studios went bankrupt and the movie star was invented; etc. Throughout that whole time, absolute movie attendance was falling precipitously while other parallel reinforcing business models were being invented, rising, and falling. And movies’ cultural meaning changed dramatically from disposable mainstream distraction, to national popular entertainment, to rarefied foreign art, to international media spectacle.

If we release our false sense of media stability, things seem a lot less chaotic. These changes will never be resolved. No medium has ever actually been stable for any significant amount of time. We’re constantly reinventing what our physical media mean to us: economically, creatively, culturally. Twas ever thus and twill ever be.

Apache runs CGI scripts that do HTTP GETs, and those GETs don’t have the same power as clients that Opera does. For example, if a PHP script fetches a remote web page, it can’t execute Javascript that it finds. But that doesn’t mean it’s dumb for PHP to be able to do fetches, it just means that fetches have a different purpose in that context.

I bring this up to defend Opera Unite from what I see as unfair charges that it’s bad at file sharing, because I think it’s not designed for file sharing.

What is it? An incredibly potent extension to the toolkit of web apps, allowing them to do hacks which were fundamentally impossible in the past. For example, they could do interframe communications using RESTful interactions, by bouncing them off a local server. And they can receive asynchronous notifications! That’s a big deal.

Of course Opera Unite is sucky at doing the kinds of things that traditional web servers and file sharing tools excel at: it’s running inside a web page. Nothing you or the Opera people could do to fortify it would change the basic deal that it’s a daemon in a highly unstable operating environment. Uploading big files doesn’t make a lot of sense if the page the server is in is likely to disappear at any moment. On the other hand, uploading little things like RSS files and CVS diffs can pack a big punch.

Having web browsers act as personal web servers is a simple and profound technique that will have a big impact on internet music if it becomes common.

One of the major things this makes possible is for web apps to access your own MP3s. So the MP3 is on disk, and is managed by the OS and filesystem, but the software is any web page. This would create very healthy competition among MP3 software, ultimately upping the quality of all player software.

Another big app that this makes possible is HTML pages inside of ZIP files, which is something you’d do if you were using ZIP as an ad-hoc digital packaging scheme. It’s completely possible to use such a file to spread a virus. A clean way to block that attack method is to open the HTML page not via the filesystem but instead via a local web server.

The Opera demos include an app for listening to MP3s on your private PC from a remote location. I don’t think this will be a standard way to listen to music in the long run. You’re not going to keep a web page open all the time, that’s just not how people use the browser. Also, the network connection upstream from your house is too unreliable, and most client OSes are too flaky.

---

Update: some ideas for hacks

Get asynchronous notifications in web apps, like this:

1. AJAX app sends URL of local server to remote server
2. AJAX app polls local server
3. Remote server eventually POSTS event notification to local server

Why can’t the AJAX app poll the remote server? Either it’s too resource-intensive, because the polling loop is on such a short timer, or it’s too slow, because the timer has been loosened to reduce resource usage.

---

Another thing you could do is communicate between pages on the same localhost. So for example you have a mail app open in one page and a music app open in another. They could bounce communications across the local server.

---

Another thing you could do is store remote data on the local machine without inventing a non-RESTful protocol.

Brett said, in a comment in the “media apocalypse” thread

People still watch lots of television. … Whenever I read that “everyone” has an ipod or “nobody” listens to the radio anymore, I take that with a grain of salt. It’s easy to become myopic and forget that there’s a world out there of people who actually lost their television service in the digital conversion because they don’t have cable, or who drive all day and leave the radio on. Some people even still read newspapers!

Logically, I can’t see how to refute the case that newspaper, tv, radio, and records are going down like Jabba the Hut in a hang glider. Except to say that these are huge enterprises and they don’t just go poof. Maybe the most elegant prediction of the future is the one that says all media will follow the business patterns established with blogs, social networks, web aggregators, and search engines. But the world isn’t obligated to be elegant.

I expect to be surprised, I guess. I have a gut feeling that all the value created by the labor it takes to do old media will still be needed, and just as many people as before will be making a living. But I can’t see how that will happen.

It’s a cliffhanger.

Henry Blodget feels that television is so fucking fucked:

As with print-based media, Internet-based distribution generates only a tiny fraction of the revenue and profit that today’s incumbent cable, broadcast, and satellite distribution models do. As Internet-based distribution gains steam, therefore, most TV industry incumbents will no longer be able to support their existing cost structures.

Jerry Del Colliano feels that terrestrial radio is similarly doomed:

the last insult may not be the demise of the Evil Empire [Clear Channel] but the lure of purchasing radio stations at long last for favorable prices at a time in history when an entire generation is not available to be a growth engine.

I would buy a radio station not because it makes money or could make money again, but because it has a brand — a real strong brand – that could lead into a digital media platform.

For myself, I keep noticing how straightforward it would be to make dramatic improvements to the usability of cable television. Why not buckle down and do the work? My wild guess is there’s a clusterfuck going on among the owners, and the only thing they can agree on is to keep milking the cow until they can’t anymore.

And then what? What’s next? Where do all these dollars disappearing land?

Bruce Warila: When the known and the unknown worlds collide, will fans still buy music?

I’m just going to speculate that as the Known and Unknown worlds collide, Middle World artists will see a significant drop in music revenue.

When ten thousand free, just-as-good songs [from Unknowns] (about 600 hours of listening time – created annually) find a mass-market of receptive (key word here) music consumers, the Unknown World is going to sponge up a lot of the ‘enthusiasm’ that fans previously allocated to Known World artists. It makes me wonder: with 600 hours of just-as-good, free music available, will music fans still buy music?

The perfect disruptive business in this industry combines the following: free-sorted-sifted-just-as-good music coupled to repetitive mass-market exposure (for each song), combined with minimal overhead and zero legacy music industry legal friction.

I think it’s possible to create the business I just described, and this is the reason why I don’t get excited about businesses that intend to sell music (now yes, future doubtful). There are just too many artists with lots of just-as-good songs that deserve to be in the Known World club.

One bottleneck with creating the business Bruce just described is finding a way to attract a critical miss of listeners for no-label music. Listeners will steer towards stuff they already know, which is Known and then Middle World musicians. Why would the listeners go to your site for unknowns?

The other bottleneck is finding a way to cover your costs without having a big listener base to start with, since you’ll need to grow slowly, in tandem with the musicians and listeners.

These issues are what I’m thinking about when I gibber about cool netlabel activity like Phlow. Phlow is part of its community, no larger and no smaller, and the space it’s exploring is the disruptive angle that Bruce is thinking about.

Note: I trimmed Bruce’s post way down, and I lost a lot of the flavor. Check out the original.