by Mike Masnick
Wed, Jul 2nd 2008 9:16pm
by Timothy Lee
Fri, May 16th 2008 12:09pm
from the privacy-or-competition? dept
Inforworld reports that Facebook has cut off Google's Friend Connect service from accessing Facebook's APIs. Facebook claims that Google Friend Connect "redistributes user information from Facebook to other developers without users’ knowledge, which doesn’t respect the privacy standards our users have come to expect." Techcrunch has more details about what Google was doing and what Facebook objected to. Facebook is getting a fair amount of flack for this decision, and it's not hard to see why. Given that Facebook has just rolled out its own competing service for linking third-party websites to Facebook, we can't help but wonder if the privacy issues aren't just a cover to avoid having to interoperate with a major competitor.
Still, Facebook's privacy concerns aren't totally bogus, and this dispute does illustrate the point we made on Monday about the challenge of building an open API while preserving user privacy. It's true that users ultimately have control over which applications and sites they approve to access their Facebook data. But users aren't necessarily going to know which applications have good privacy policies, nor are they necessarily going to want to invest the time and effort to figure it out. So it's not necessarily a bad thing that Facebook is imposing at least some minimum standards on sites that use their API. And while Google obviously isn't a fly-by-night organization, Facebook may be worrying about the precedent it would set if it started allowing sites to funnel information gleaned from the Facebook APIs to third party sites that Facebook had no control over at all.
But on the other hand, the fact that Facebook can and does arbitrarily disable Facebook apps isn't going to be good for the health of the Facebook ecosystem. If I were a software developer, I would certainly be reluctant to develop for an "open" platform like that. And in the long run, that's a big threat to Facebook's dominance of the social networking universe. Facebook is big, but it's not as big as the rest of the web put together. If a company like Google can figure out how build a usable, open social network atop hundreds of websites, it will give Facebook and MySpace a real run for their money.
by Timothy Lee
Mon, May 12th 2008 4:06am
from the privacy-challenges dept
The New York Times reports on MySpace's new Data Availability project, which will allow third-party websites to automatically import information from a user's profile, saving the user the hassle of re-entering it on a bunch of different sites. As the Bits blog says, this is a smart move and could be the start of an important trend toward making MySpace a more outward-focused platform. Facebook wasted no time in announcing a program of its own called Facebook Connect that will have similar functionality. And now it looks like Google will rolling out an extension for Open Social that will provide some of the same features. We've said before that the achilles heel of social networking sites is that they're so inwardly-focused. In the long run, it's going to be difficult for any site—even one as large and technically savvy as Facebook or MySpace, to get users to stay inside of a walled garden. The site that figures out how to be a platform that other sites use for identity management will have a huge advantage in the long run. MySpace and Facebook appear to regard this kind of outwardly-focused platform as the next frontier in social networking.
They're going to face some serious challenges in the privacy department, though. We discussed the privacy dilemma with Facebook apps back in January, and both Facebook and MySpace are going to have to grapple with the same set of issues with their new strategies. On the one hand, they need to lock things down sufficiently that one bad (or compromised) application can't suck down a ton of private user data and do bad things with it. On the other hand, if they are too restrictive, it will limit the usefulness of their platforms and discourage third-party websites from using them at all. Striking this balance, and coming up with security mechanisms that give sites the information they need without giving away the store, will be crucial to these initiatives' success.
I think it's possible that these problems will prove intractable. Ordinarily, when we talk about an "open API," we mean an API that anybody can use without any kind of pre-approval, and that doesn't constrain how data received through the API is used. Obviously, MySpace and Facebook aren't going to want to offer an API that's open in this sense; there's too much potential for mischief if an application can trick a user into authorizing a malicious application. So unless they can come up with a really elegant mechanism for limiting the spread of information, there's going to have to be a vetting process, which will mean extra overhead that limits how quickly the platform can grow. It's going to take a lot of ingenuity to make these platforms open enough that a lot of potential partners can participate while keeping them closed enough that they don't become vehicles for the bad guys to cause problems.
by Mike Masnick
Tue, Apr 8th 2008 12:27am
from the took-'em-long-enough dept
So, now, finally, nearly four years later, Google has come to its senses and announced its entrance into the web platform space with its aptly named AppEngine offering. In many ways, it's similar to Amazon's offering (which is a good thing!), though much more integrated, which could prove to be either a problem or a benefit depending on what you want to do. Amazon allows for a much more a la carte setup, which could appeal to many, while you have to really embrace Google to enjoy the benefits of its setup. A big open question is pricing. A huge part of the appeal to Amazon's Web Services platform is that it's crazy cheap. You really have to be working it quite hard to build up any sort of significant charges. Google hasn't released info on pricing yet, offering AppEngine up for free to the first 10,000 developers (who appear to have snapped up all the open slots in less than two hours). That free service has some limitations: initially 500 MBs of storage and enough bandwidth to serve approximately 5 million pages per month. There's some suggestion that the final service will always be free up to that level, with charges starting if you go beyond that. If so, that could certainly appeal to people who just want to try some stuff out for free.
While this may seem like something that will only appeal to serious techheads, this could be a really big deal. A lot is going to depend on how well AppEngine really works, and how open it really turns out to be. However, if it really does provide another super cheap (or even free at low levels) full service, highly scalable platform for all different kinds of applications, things could start to get very interesting pretty quickly. Between this and Amazon's Web Services, the very concept of developing online applications may finally start to change in significant ways for the better. The easier it is to develop and deploy highly scalable web applications, the more innovative and creative solutions we're going to start to see.
by Mike Masnick
Thu, Jan 3rd 2008 11:58am
from the not-allowed-to-make-things-more-valuable dept