07 August 2009

Twitter outage creates panic

According to CNN, the twitter outage left users feeling as if they had lost a limb or left home without their cell phone. It is suggested Twitter needs competition to provide alternatives when an outage occurs, as they inevitably will. There were (or are?) a couple of alternative social microblog services available (is jaiku still running?). Of course, this won't help if multiple sites get attacked at once.

What would help is Google Wave. This outage is an incredible opportunity to demonstrate the potential for resiliency a "federated" or distributed social media system has. Content in the Google Wave universe is independent. Every user can have a copy of a bundle of posts, comments, content on their own device. Multiple copies can exist on different servers. It could be possible for a group to continue working, or at least work offline with their content, during an outage and then when connection is reestablished, the changes can be merged back into the conversation. This is what we can do with email now. We can read messages in our inbox (as long as it is not webmail) even when the network is unavailable. We can keep messages around as long as we want. We can write draft replies, take notes, resuse content through quoting or editing the text we recieve and at any time later, forward to others or send the revisions back to the sender.

What did I do during the panic? I just waited for Twitter to come back up. I only post once a day (if I am feeling up to it).

06 August 2009

You can't just put content behind a blank wall

I caught a discussion of Newscorp's new plan to get users to pay for online news content. It will be difficult to sell news online because there are so many fragmentary ways to get the news for free. If any scheme for getting online users to pay for news does work, it just has to be easy. No matter what news sources online do, they must make paying for the news easy and transparent like iTunes. As easy as putting a coin in a paper box at the corner bus stop. The pricing is not as important as convenience.

Also, the customer must have a feel for the worth of the content before they buy or they must get a cheap bulk subscription so the content is cheap enough to take the irrelevant, incomplete, incompetent or useless with the relevant, complete, competent and useful content. I hate sites that put up a poorly written summary and a login or subscription screen. It breaks the rhythm of navigation on the web when a link leads to nothing. It stops you cold and punishes the user for following a link. It would be a sad web of balkanized content with links as obstacles. If content is to be shuttered behind closed doors, it must be quick and easy to open those doors with some kind of universal pass like OpenID connected to a micropayment system.

It started me thinking again about how to get online users to pay for content again. You can't just put content behind a blank wall and expect it to work. No one will ever find it, be able to search for it, search engines index it. Its not enough to provide a meta data summary like a bibliographic catalog does. Meta data will never be the answer to our search problems, at least not as long as humans are responsible for providing it. Nearly everyone ignores meta data, fails to include it, or includes incomplete or incorrect meta data. Who is going to keep all this meta data up to date? No, this is unworkable. Meta data must be generated automatically from content and that is subject to a high error rate using current technology.

The solution google books provides gets much closer to a real solution to the problem of hidden content. Instead of trying to describe the content using faulty and hard to keep up meta data, why not grant access to a sample of content? This gets much closer to a successful model for selling content online. When I read a book in Google books I get a random sample of pages around my keywords. Each user received a custom sample of content tailored to their interests and needs. In my experience reading a few pages of a book without restriction, as I would in a bookstore, gives a feel for the content. I am more likely to buy the book if it proved useful repeatedly over several searches. Yes, sometimes I find what I want in the sample pages, but I generally bookmark the source, take down the title in my notes and will cite the source in any work derived from the information gleaned for "free" which is actually a fair exchange I think for citation and a link.

I do not understand the hostility and opposition to Google Boooks. I am willing to pay less but buy more books in electronic form for reference purposes. If I find an interesting book in google books but it is not one I would pay $30 for a hardback I would pay $10 to download to my book reader. If I have to pay $30 for one book, it is going to be the one I value most and need the information most, which I want to keep around for a lifetime, not a casual read or reference work.

There are books I would buy on the reader as convenient portable references. I would buy more ebooks at lower cost to fill out my "search space" of texts on my ebook reader. If a book adds to the information I have available on a subject but only partially or tangentially, I can't afford a $30 hardback, but I can afford three $9 works related to my subject to add to the search space on the reader

An idea I had a long time ago, when I was wondering how to pay for hosting my first website, was the "vanishing page" model. This would work a bit like PBS where content slowly disappears unless readers pay a small fee to keep it available. The individual reader does not pay for each content page, butsimilar to donations to PBS, a small number of readers or viewers pays for free access by others (this actually gives the donor a feeling of superiority, if it were not for me...). Mechanically, the web page would be publicly available to all readers and search engines but a count of page views would be kept. Each time the page is viewed the number of views or days left would be decremented by some amount. A button to make instant micro-payents would be displayed ok the page along with a thermometer displaying how close the page is to being removed from the site. If enough people donate, days (or credits, it could be a ratio of views to donations similar to bitorrent) are added to the life of the page, if not, it is replaced by a summary and a button to start donating again.

What we need are ingenious "social engineering" methods to get people to buy content online, similar to the ones used to manage "soft security" on wikis. We need soft methods, like Google Books, which gives readers a peek into books that might interest them.