Thursday, June 7, 2007

Microsoft Surface

Before reading any further, check out and make your own judgment.

First thought is it's cool, brings back memories of Minority Report. However, after some thought it really looks too good to be true. And here are some reasons I feel it is doomed to failure:

  1. Think of the iPhone for a moment. Isn't this the sort of thing Apple have promised, just much smaller? It's 1/20th of the price and surely has far more real world uses.
  2. Those demos are crooked. They make the thing appear so easy to use. Far more so than necessary.
  3. Apparently they've yet to write any software other than what's included in the demos.
  4. From the little I've read, it appears that to have the thing identify an item it needs to be barcoded. Now whose barcoding their glass?? Remember that these are at first only being sold to hotels and other public places. So what's the incentive to barcode your belongings if you're only going to be using it briefly?
  5. It's expensive - $10,000!!
  6. They took 5 years to develop it.
  7. The last piece of hardware Microsoft tried selling was the Zune.
Please, I'd like to hear if this is just be being anti-Microsoft or if other agree with me. Of all the computer scientists I've asked, they agree with my opinions.

And thanks for all the birthday wishes everyone. Had a great day yesterday!! :D


  1. Yesterday I saw this fascinating video of another Microsoft product (although by acquisition) called Photosynth.

    I didn't have a thorough look at Surface, but I wonder if the technology is related?

    By the way, before our team (stellenbosch) went to the second round of the IT challenge this year, we read your round-up of it, and it was great to know exactly what to expect!

  2. It doesn't appear that the two are related, although the apparent ease of (how do you put it?) throwing the images around that they show before demonstrating Photosynth could be. This throwing around of images is what concerns me a bit as well - is it really as easy to use as the demos show? Bear in mind that the demonstrator is a techie probably working on the product himself.

    As for Photosynth, that seems like it could be quite ground breaking. Although I'm surprised such a technology hasn't been made available a long time ago. At least this is just software and not MS trying to sell hardware which I feel they should limit. Some concerns I have over it:

    1. How much computing power does it require? If it uses a lot of heuristics it could fair well. However, if it does any brute search it could prove pretty expensive. Notice how there were no demos showing how to create a collection. Remember, they're talking about hundreds and thousands of photos.

    2. The UI is extremely crucial in such a product. The only current 3D model viewing products aimed at the average user I can think of are Virtual Earth and Google Earth style. Int those you can search though and easily be moved to a specific position. In this case I don't see that working. So it requires some pretty unconventional UI to really become widely used.

    Sorry if I'm attacking the products a bit harshly, but I prefer to look at the negative sides of pre-release products.

    Glad to hear my experience in the IT Challenge helped you. I try maintain that site as much as possible, but it takes up so much time that I only end up writing about a small selection of events. I'd love to get down to writing about the IPSC and ICFP as they are two really great contests. Apart from the prizes in the IT Challenge, I really dislike it. Perhaps a topic for another blog post some time.

  3. I'm no expert but I guess it's safe to assume that they use some sort of vector graphics which enables them to do many of the things in the demo.

    Do you think that means they have to convert all the graphics imported to the product? Imagine how long that must take, it could be one of the big concerns regarding computing power

  4. To display those extremely high resolution images at low resolutions they probably make use of mipmaping.

    It's building the 3D models from collections of photos that concerns me wrt computation. They might use vector graphics, but still then they need to have some why of taking each photo and quickly find out where it fits. No matter how fast each comparison is, the possible matches just increase exponentially.

  5. "There's a new way to live the digital world... A new way that feels...(pauses)... familiar. A new way that aaacts like us. A new way, that isn't NEW AT ALL!" -- Microsoft Voice-over guy.

    You got that right Redmond. It ain't New at all. Its a damn iPhone with Legs!

    Microsoft have the annoying habit of making every one of their products sound like they're vital and relevant to our lives right now.

    Sure, Surface Computing may be one of the futures we end up at, but that future isn't now.

    Plus, to be honest, I don't regularly enjoy throwing photos around my desktop. I tend to find it a bit counter-productive :P

  6. Surface computation is not ready for an average user and that is what microsoft is claiming... When Microsoft claims that surface computation would be used by hotels and casinos etc.. the hotels and casinos will be responsible for barcoding their products be it glasses of lemonade or chips in casinos... and yes it is expensive and meant to be espensive because it is being marketed for enterprises... The advertisement just suggests the possibilities. Also with the way microsoft provides its source support and documentation and with the advent of .NET, I cant imagine that programming would be hard. I believe that surface computation would be readily used for various different applications in the near future..