I got mentioned (alas my name was misspelled it's Yaron Goland not Yaron Golan) in Cringely's column this week. I read Cringely's column every week so getting mentioned is pretty nifty. What is even nicer is that while Cringely uses my old project, UPnP, as an example of why MS screws things up (not that I can disagree with him, UPnP was a lot of the reason I ended up quiting MS, see my article on the subject) he was kind enough to mention that I don't suck. That was very cool of him.
Interesting read. You don't come across as too bitter about your Microsoft experience. I'm curious about what you think about the new UPnPv2 (web services) Microsoft specification – we just implemented UPnPv1 support for our embedded TCP/IP protocol stack, which seems like maybe it won't do very well in the market because of the timing of UPnPv2? Then again, UPnPv2 will require a full-blown XML parser in the network device (i.e. more RAM, ROM and CPU – not all embedded systems can support that), Microsoft appears to be committed to supporting both UPnPv1 and UPnPv2 into the future, and some UPnPv1 specifications such as A/V and IGD are already mature and shipping product. I'm interested in your ideas about the future of UPnPv1, will it continue to grow in the SOHO environment?
I learned a lot at MS and accomplished some really amazing things there. So no, no bitterness. As for the future of UPnP, the mere support of MS is enough to keep it alive for quite some time. I am aware of UPnP V2 but have chosen to stay away from it because my idea of a good time isn't making the whole world run over web services. Using XML for a discovery protocol strikes me as a bad idea but I'm sure MS will hack it until they more or less make it work. Because of MS's absolute ownership of SOHO I expect the UPnP protocols to have a long but limited (as in only a few specific devices) life.