I just spent an hour doing some research in response to Sam Ruby's post Sousveillance where he wonders whether some of the descriptions of Facebook as a social graph roach motel (i.e. information about your relationships goes in, nothing comes out) is accurate. Sam writes

Dare seems to think that the root problem is oppression by the “man”.  In this case, a 23 year old.  Brad seems to view this as a technical problem.

I wonder what I wrote that gave that impression especially in the linked post. In that post, I was simply giving some advise about the kind of social problems you will face when treating unifying social graphs across different contexts and applications as a technical problem. If anyone is whining about oppression by Facebook, it would be Brad’s original manifesto which mentions the site by name over a dozen times.

Data point 1: one day when logging onto Facebook, I saw an offer to scan my AIM contacts and invite the ones that had Facebooks to be friends.  I unselected a few, and then clicked on submit.  Within hours, my network expanded greatly.  IM ids serve as useful foreign keys.

Like lots of popular social networking services, but not Windows Live Spaces, Facebook is fond of violating the terms of use of various email providers by screen scraping user address books and contact lists after collecting their log-in credentials.

However Facebook prevents this from being done to them by only showing email addresses as images which expire after a couple of minutes due to use of session keys. I once considered writing an application to import my Facebook contacts into Outlook but gave up once I realized I couldn’t find any free, off-the-shelf OCR APIs that I could use.

I did find an article on CodeProject about rolling your own OCR via neural networks which seems promising but I don't have the free time to mess with that right now. Maybe later in the year. Sam also writes

Data point 2: Facebook is a platform with an API.  If there is a need, it seems to me that one could develop an application using FQL to pull one’s friend list out of Facebook and share it externally.  The fact that I don’t know of such an application means one of four things is happening: (1) it exists, but I don’t know about it, (2) despite the alleged overwhelming demand for this feature, and obvious commercial opportunities it opens up, it hadn’t occurred to anyone, (3) I’m reading the documentation wrong, and it isn’t possible for applications to obtain access to one’s own Facebook ID for use as a foreign key, or (4) the demand simply isn’t there.

Or (5) the information returned by FQL about a user contains no contact information (no email address, no IM screen names, no telephone numbers, no street address)  so it is pretty useless as a way to utilize one’s friends list with applications besides Facebook since there is no way to cross-reference your friends using any personally identifiable association that would exist in another service.

When it comes to contact lists (i.e. the social graph), Facebook  is a roach motel. Lots of information about user relationships goes in but there’s no way for users or applications to get it out easily. Whenever an application like FacebookSync comes along which helps users do this, it is quickly shut down for violating their Terms of Use. Hypocrisy? Indeed.

Now playing: Lil Boosie & Webbie - Wipe Me Down (remix) (feat. Jim Jones, Fat Joe, Jadakiss & Foxx)


Tuesday, August 21, 2007 2:06:48 AM (GMT Daylight Time, UTC+01:00)
Google open-sourced an OCR lib named Tesseract last year under an Apache license. Version 2.0 was released last month:
Tuesday, August 21, 2007 2:40:50 AM (GMT Daylight Time, UTC+01:00)
Besides the fact that it isn't a .NET or COM API, I tried it and the results were horrible. I assume it needs to be trained but that would take more time than I plan to devote at the moment.
Tuesday, August 21, 2007 5:22:06 AM (GMT Daylight Time, UTC+01:00)
Re: OCR - how about Microsoft Office Document Imaging:
Tuesday, August 21, 2007 6:26:11 AM (GMT Daylight Time, UTC+01:00)
This is completely evil and wrong.

The more I learn about Facebook, the more glad I am that I'm not participating in it.

Tuesday, August 21, 2007 9:36:15 AM (GMT Daylight Time, UTC+01:00)
Luckily I have multiple copies of FacebookSync - its a lifesaver and keeps all my Address Book contacts in sync with Facebook contacts. I agree that it is hypocrisy for Facebook to ban tools such as this. It still works, so it isnt a problem for me, but FB are losing good will with moves to drop the distribution of it.
Master William
Tuesday, August 21, 2007 11:48:26 AM (GMT Daylight Time, UTC+01:00)
My apologies. I initially linked to the wrong post, that has now been corrected.

FYI: the captcha isn't showing on Firefox on either WinXP or Ubuntu for some reason. I'm entering this comment using IE.
Tuesday, August 21, 2007 12:35:24 PM (GMT Daylight Time, UTC+01:00)
Hi Dare, All,

Yep, Facebook is indeed a "Roach Motel". When the opening up of Facebook was annnounced, it was well received. But I think the opening up that was truely intended was for independent developers to be able to build applications to add to the platform (i.e. Facebook), to bring in external data (i.e feeds, etc), and not necessarily in order to take data or content generated inside Facebook, out. At least not in any way that has depth. I can understand why they keep this all walled in: it helps the network to keep growing. At some point, they will have to opening it all up. This will be gradual.

In the early days of IM, when IM was sort of like the "social network", this was also the case. You could not port your "IM environment" to another. You could not "mix" environments. You could not reach your buddies on other IM platforms. Then, slowly, we saw direct IM to IM platform agreements to interconnect (excluding perhaps other IMs), and it has been getting better. So, maybe at some point, Facebook might find that it needs to "get" data from another platform and the condition that platform will put down will be the ability to get data out of Facebook itself. This "other" platform will have to be big on its own, etc.

What I am trying to say is that Facebook can stop being a "Roach Motel", but in doing so, they will have to be able to derive huge benefits from truely opening up (more growth or more cash, or both).

These are just my thoughts.


Tuesday, August 21, 2007 5:19:04 PM (GMT Daylight Time, UTC+01:00)
To Sam Ruby (with apologies to Dare for the hijack):

FYI: None of the pages on your site can be viewed by IE (except for the Atom feed).
Wednesday, August 22, 2007 1:20:14 PM (GMT Daylight Time, UTC+01:00)
Ziv: a more detailed bug report would be helpful. In particular, data from the following page: http://intertwingly.net/test.cgi.

I routinely visit my site using IE, and have people leaving comments from IE. In fact, IE users are the sole source of my AdSense revenue due to a confluence of both IE's and Google AdSense's lack of support for XHTML.

It may also interest you to know that this post is being entered using IE after having clicked on the trackback link from my weblog. Dare can verify that if he likes. The user agent that IE is sending looks like this:

Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 1.1.4322; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30)

(I'm now *seeing* the captcha on Firefox, but my comments aren't being accepted)
Wednesday, August 22, 2007 3:29:35 PM (GMT Daylight Time, UTC+01:00)
I'm not sure what's going on with the CAPTCHA. I'm using Firefox to make this comment and it seems to be working fine for me.
Wednesday, August 22, 2007 3:31:14 PM (GMT Daylight Time, UTC+01:00)
Oh...wait, I remember now. The CAPTCHA has a time out setting so if you spend to long drafting your comment it may have expired and I don't think there's an error message that describes what went wrong.
Wednesday, August 22, 2007 4:29:31 PM (GMT Daylight Time, UTC+01:00)
test from firefox
Wednesday, August 22, 2007 9:56:24 PM (GMT Daylight Time, UTC+01:00)
> None of the pages on your site can be viewed by IE (except for the Atom feed).

I've noticed that too.

There's nothing quite as refreshing as a big f**k you to the entire community of people who don't happen share your browser preferences.
Thursday, August 23, 2007 3:07:35 AM (GMT Daylight Time, UTC+01:00)
> There's nothing quite as refreshing as a big f**k you to the entire community of people who don't happen share your browser preferences.

Ziv has confirmed to me that some pages work sometimes, even with his configuration.

I have yet to reproduce the problem myself with IE.

More data would be helpful. Invective laced comments are somewhat... less so.
Comments are closed.