75%-baked: Only Live Music

I have various half-baked ideas for making money, but I reckon this one is 75% baked. It would definitely attract a significant following if it got going. The problem will be how to make money out of the popularity. Perhaps sell it to google, like YouTube.

---

The day is fast approaching when all radio and video will be delivered over the Internet. You pick a radio station by clicking an icon on your PDA. As we see on YouTube and many other places, this opens things up to the amateur contributor.

Another trend we see is the return to live music. Why do we have shows like "Australian Idol"? Because the whole point of music production is for the performers to show off their natural talent and for the audience to evaluate the performers. Music produced by electronic trickery in studios is increasingly recognized as comparatively empty.

It is possible to stream live sound over the Internet today. Internode is leading the way in Australia: http://www.internode.on.net/radio/. Many of the network providers can do this very efficiently using multicast. So my idea is to start an internet radio station providing only live music. It would later be split into multiple stations providing different sorts of music.

Initially it would be all free with amateur sound engineers and performers not paid. Later some extra channels would be supported by advertising or require a subscription (or both), and those channels would pay production staff and performers, not to mention management. Initially expenses, including some payment to management, will attempt to be covered by subscribers who will be able to vote on which acts they prefer to hear.

The station announcement between performances is "This is OLM: Only Live Music. We don't believe that recorded music is real music, and now you don't need it because OLM streams live music, produced with analog devices, 24 hours a day. Recording OLM performances is not permitted and detracts from their intentionally transient and contemporary nature. This doesn't prohibit technical aspects of streamed radio transmission such as the necessary cache delay. If you miss something then encourage the performers to do OLM again. And become a paid up subscriber so that you can vote on which performers you want us to feature".

The assertion that only live music is real, is a marketing position. It is designed to annoy and make people talk. Participants don't have to agree with it.

The key to OLM is to broadcast performances from different time zones around the world. This can be as simple as a solo performer with a good quality microphone broadcasting from their own home. However it will be better to broadcast actual live performances from venues. That requires some sound engineering, though these days that can be just multiple microphones feeding into a computer running some sound software.

Obviously OLM needs lots of people around the world. The sort of things one needs to do to achieve this are:

• Set up discussion forums for interested people: a wiki and a blog seem likely to be a good start.
• Put information on the wiki pointing to anything that can be found out about how to broadcast music on the Internet.
• Look for support from useful people. Simon Hackett the head of Internode has been involved in Internet multimedia for a long time and might well be interested -- I remember him demoing listening to California radio from Australia, including remote tuning using SNMP.
• Communicate with groups where interested people might hang out (without spamming too much), to invite them to join it.
• Start practicing broadcasting music, preferably as early as possible from different places around the world, to see how well it reaches.

My view is that if some major ISP owner was interested to be part of it then all the technical issues would be addressed. I also think that OLM would attract amateur sound people and musicians willing to work for no payment if the technical issues were sorted out.

2005/03/11: Workflow user interface

2005/03/11: Workflow user interface

My current obsession is workflow, and particularly a workflow-oriented user interface to everything.

Once upon a time computers were expensive and people were (relatively) cheap. So the idea was that people would attend to the computer and keep feeding it. The modern progress bar continues that idea: "here's something to look at while you wait for the computer to finish something". Window systems, and underlying multiprocessing, mean that computers can do more than one thing at one. Instead of looking at the progress bar we go to another window. Now we don't know when the activity is finished except by coming back to it. So we have to remember all the things that we are trying to juggle and keep going back and checking them for progress. So we need a workflow user interface, and what do people use: they use e-mail. That's something they keep looking at and maybe it comes with a notification system. But it is a really bad workflow user interface.

A real workflow user interface lets us know what activities need user input, or want it, or would be willing to receive it. My conception is that activities are organized in a tree and the nodes change colour: green means that this activity can't proceed without input from this user; yellow means the user might push things along, but does not have primary responsibility to do so; orange to red means that the user can only do things that won't affect other players much. For example if the activity is a chess game then the tree node for that activity is green if it is your move and orange if not (since you could resign) and red if the game is over.

Now we don't need a progress bar: the node for that activity will be reddish until it goes back to waiting for us to input. More on this another day.

Wirelessly networked PDAs threaten to become ubiquitous combined with mobile phones. These can hardly afford a multiwindow UI. Also it will be hard for the user to look around for stuff needing attention. Also interaction activities are likely to be the main applications. So for various reasons my workflow user interface is likely to be a good answer.