Project Stargate: "Always On" Skype video connection between remote offices

UPDATES: Stargate is Live! | Also, see SimpleGeo's Implementation

We recently expanded PointAbout from its office in Washington DC, opening a new office for AppMakr in San Francisco (as an aside - interesting personal-interest story about the move here).

One problem with geographic separation is that the people in the other office are very abstract. You might see work coming from them in the form of code, releases, email, etc., but it's very hard for the actual people between offices to make a connection. And that inter-personal connection is very important to us as a company, so I set out to figure out a way to solve it.

My co-founder Sean aptly named this project "Project Stargate," after the science fiction Stargate movie in which shortcut paths to the Universe are opened up through a round, glowing portal (really wanna geek out? More details on Stargate here), so the name has stuck.

I'll be adding to this blog as we figure things out, and I'd really like to have your thoughts in the comments section below, especially if you've done any work to solve this problem (any Y Combinator startups out there working on this vexing issue?). Just to be clear, we all have AIM, or Gtalk, or Skype call & video chat abilities - that's not the problem.

The problem is that to initiate a call or IM chat requires effort. And while it may be a very small, minimal amount of effort, nonetheless it's effort, which means that the magical moments where someone turns to another person and says "hey, I have an idea" or "can you look at this for me" are lost. And that's really, really significant. Since effort is required to establish a connection, the serendipitous connections are lost. That's what I'm trying to bring back through "Project Stargate".

My idea is to have a persistent, "always on" type of connection between the two offices. No screen savers, nothing to power up or turn on, no connections to be established. It's just an always-on "portal" between the two offices.

Here are the issues I've established so far in this project, which I'm working to resolve:

Big brother: This is a concern of Sean's, and well founded. It's amazing how a person in a room is no big deal, but a camera can feel creepy. And creepy is definitely not the feeling we're going for here. We're experimenting with various ways to mitigate this issue, among them:

We've chosen Logitech microphones with an on/off switch. That way, there's no audio unless someone wants to talk. This re-introduces the problem of effort, but I'm hoping it's more of a "Nextel Push To Talk" type of thing that might not be as big a barrier to spontaneity.

We've toyed around with the idea of putting wax paper or similar in front of the camera, to blur the image. That way you'd be able to see movement, but not the exact person or what they're doing. Since you see movement, you could initiate an interaction with the person, without them feeling like they're being watched constantly. Sean and I even toyed with the idea of having the wax paper (or translucent plastic, etc.) be on a spring in front of the camera so it would have to be physically flipped up by the user and would spring back down when not in use. Maybe a more advanced way to do this would be to have a big button that serves both as a microphone on/off switch and a "privacy screen" flipper, so a user could just hit a big button on the desktop to establish a full connection, while otherwise the unit would have no audio and only fuzzy images. I imagine this could be accomplished with software as well.

Having the window box of your image be bigger: Speaking of software solutions, one problem in all the software I've tried using is that the box which shows you is always very small. I wish it were re-sizeable. Sean's feeling (and I agree) is that if the person who was in the room with this device could more clearly see themselves, then they'd at least be more aware of the fact that they're on camera, and therefore wouldn't be surprised by it. None of the software I've found allows for this, or seems to have considered this. I guess all the software out there assumes that when a connection is established, it's an active connection. I'm looking for more of an "always on" passive connection.

A lazy susan: We're considering putting the monitor (and camera attached to it) on a lazy susan, so it can easily be swiveled around, out of view, or turned back towards the user for a conversation. Another option would be to put just the camera on a lazy susan, so it could be pointed at a wall but easily turned back around towards the user.
Persistent Always On Connection: This part's pretty easy to do. I was able to get all the hardware for just about $1,000.

I bought two eMachines (note - the ER1402 model was available on Amazon Prime but doesn't seem to be anymore) and downloaded the Skype 5.0 Beta (Windows only, unfortunately, or I would've used Mac Minis), and a Logitech HD Pro C910 webcam (does 1080p video). Don't forget the HDMI cable, the monitors, and speakers. Total cost: $544 per workstation. Not bad. I tested various video conferencing software, including G-Talk (worst quality), Skype 5.0 HD Beta (better but not great) and Logitech Vid HD, which was far and away the best quality, and free with the C910 webcam. The Logitech software is what I have running in the video below. It runs fullscreen with good quality. Not "your favorite program on HD TV" great, but very good, especially compared to others running on the same connection speed.

We may also try offering our video connection up to AppMakr users - this is just an idea and could become overwhelming, but it's a neat concept that an AppMakr user could see us at work and connect with us via video quickly & easily. This would be phase 2, though.

PS - I love what Trevor at Anybots is doing to work on this issue as well. I'm not sure how a roving robot that feels vaguely human compares to a standard workstation, but if anyone has tried it, I'd love to hear details.

Here's a video showing the setup: