Thanks for the detailed write-up Hans. Full disclosure: I work for Laszlo Systems, which is the corporate sponsor of the OpenLaszlo platform. There are a couple of points you made that I wanted to address:
You mentioned that the simplistic Flex application you made performed substantially faster than the OpenLaszlo one. My experience is that comparable real-world OpenLaszlo applications perform about the same as their Flex counterparts, in application size, startup time, and runtime performance. A simplistic app may highlight issues that do not become apparent as the application increases in complexity. I'd like to see the OpenLaszlo application you built; it's possible that there's a simple fix that would make it perform equally - or even better ;). You'll find a lot of highly complex, publicly-deployed OpenLaszlo applications at the URL below, and I'm sure that you'll find that they perform quite well:
Regarding the double-click handling, I think I know exactly what you're referring to, because I filed this bug on it some time ago:
If you look at the history of the bug, you'll notice that we received a contributed fix that addressed the issue, and that we tried to integrate it, but ran into issues doing so. Those issues were never addressed by the contributor, which is why you had difficulty listening for the ondblclick event of grid.
Note that I'm not trying to blame anything on the contributor in this case; we are always grateful to developers who wish to help with the OpenLaszlo project. Instead, I'd like to point out that open source software - which you professed to love in your opening paragraph - is often dependant on the voluntary contributions of willing developers. This is a fact-of-life of open source software.
It may even be a testament to the power of OpenLaszlo that you were able to work around the double click issue. Do you see yourself being able to apply a similar workaround to a missing feature of the Flex grid?