One advice is that if you are focusing on a next generation platform, focus on next generation applications (applications that were not possible in the past, not migration of existing client server application). And be ready to be patient for that new wave of application to emerge.

It is the client server focus that threw me off.

Best,
Edwin

One really, really secure ring to rule them all. Like everything else in the enterprise, assuming that one security authentication layer will solve all security issues within the enterprise is simplistic at best. To the extent that external platform providers produce useful widgets and services, security for an application which interacts with multiple services and domains is “a hard problem with both technical and usability impacts,” to quote your last blog entry
Innovator’s dilemma redux. Every new wave of development architecture has spawned a new set of commercially successful development tools (e.g., stand-alone companies or acquired at a high valuation). Client/Server spawned PowerBuilder, Notes, Oracle Forms. 3-tier/J2EE spawned BEA, NetDynamics, JBoss etc. We believe that the new wave of rich internet apps and the democratizing approach of assembling apps primarily from pre-existing widgets and services will create enough a disruption in the market. Existing players will try to service this new market by adding lumpy kluges to their existing products. Most new players will get Silicon Valley disease and spend all their time fascinated with the possibilities of this shiny new technology rather than understanding how the technology addresses fundamental business challenges. We think Web 2.0 development tools represents a discontinuity and as such will be difficult for BEA et al to address (and even more difficult for techno-happy solutions like RoR) – we can agree to disagree on this, as only time will provide the proof on the existence or non-existence of this market
On making money supporting open source. I fully accept that JBoss did not make a great deal of money providing commercial support for Hibernate (although getting bought for $400M by Red Hat on a $15M revenue stream is a trick I would not mind emulating). I think you misunderstand the reason JBoss supported Hibernate, however. JBoss needed Hibernate to succeed in order for their overall J2EE stack to succeed – the revenue they made off of Hibernate support was completely beside the point. In the same way, we at ActiveGrid need an open-source AJAX toolkit to succeed in order for our overall strategy to work. We just happen to have picked Dojo (again, reasonable people can disagree on whether this is the optimal choice, but we are confident in our decision). It is immaterial to the success of our strategy whether we have a profitable Dojo support business. It is crucial to the success of our strategy that we be able to count on enterprise adoption of an open-source AJAX toolkit.

You raise very good points. There is a great deal of uncertainty in the market right now. To us, that uncertainty spells opportunity!