Coffee & Gems

There is a bug in the Flash player that makes it impossible to implement a RESTful architecture with a Flex or Flash backend. In essence, it is a buggy implementation of an http client - the client cannot get to the response body unless the reponse code is 200.. This is really bad. But what is an embarrassment for Adobe is that: That bug was opened more than a year ago It is marked as priority C only It was closed and had to be reopened after the community complained Adobe engineers suggested as a workaround to purchase LCDS Server So, lessons learned (actually: lessons that proof in practice what was clear beforehand ): It does not matter too much if Flex is open-sourced as long as the Flash player is not It does not matter at all if Flex or Flash are open-sourced if the community cannot apply patches and get them into the deployed player Flex being open source is completely irrelevant.

The Autocompleter looks nice, but why the heck does it send a POST to retrieve suggestions? I had to walk away from it.

I have a fondness for retro computers, history of computing and famous discussions (like the Torvalds/Tanenbaum debate ) . In this context I just came across this W3C mailing list debate regarding REST and SOAP. Nice read.

I just came across a post about Germany's most anticipated Web2.0 sites that dates back to last August. Since this is almost exactly one year ago I decided to spend half an hour of my life and conduct a completely unscientific survey to see how many of these startups are still around. For this reason I simply looked at the web sites to see if they are still here and checked if there are recent posts in the company blog to check if the site is a zombie (if there is no blog I checked the press releases). The result is: 51% are still around, 36% are gone (or not started one year later). The rest seems shaky (last blog post 6 months ago) or unknown (state cannot be determined by looking at the blog). If you consider that these startups were above the rest ("highly anticipated") and that checking if there are even online one year later is a very low barrier I was surprised by the low survival ratio. But one can regard this as positive: "quick failure" is a good stra