The article mostly agrees with what Drew said with very few exceptions. The article points at Asha devices (and other devices) that have very small amounts of memory (2mb) and yet perform really well with a GC (and a slow CPU).
The GC study pointed in the Drew article was a desktop study taken out of context.

It says that JavaScript is inherently slow because of DOM. It says that this should not be applied as a sweeping generalization to all managed languages e.g. Java. Then it gives examples including mobile Java performance where small heap devices work just fine.

Most j2me devices didn't use code from Sun.

It's not a rebuttle, in fact he didn't refute any claim other than the GC article. Read the comments where game programming is also discussed.

The separation is useful to understand where the optimization is necessary. JavaScript could be made less painful if it didn't have DOM manipulation to contend with obviously that's not very practical.

Come on... I have the word XYZ and search for XY it won't find it. Won't let me sort search results or apply really elaborate search filters beyond by date.
If I have an attachment and a vague recollection of the email its about as good as gone, I have to remember the exact spelling used within the email.

If it was another company I would let sleeping dogs lie but this is f'ing Google. Their web search is SOOO good and improving all the time, that makes the gmail search so ridiculously outdated in comparison. Their recent changes to search just added more information but didn't actually solve the core problem of "it doesn't work...".

Despite the comments who pick on the Iacocca example (not familiar with that) Schmidt is remarkably technical for a CEO (former Sun CTO with a T) which is really rare. TOR was not something that would be interesting to him, they are doing big data, search mobile and many other things that he needs to understand. TOR would be just general knowledge for a guy who does more than most of us do every day.

They aren't stupid, they understand that terrorism works. E.g. in Israel we had a peace process that was going relatively well between the left wing and the PLO. So the Hamas started blowing up buses and the peace process died. Hamas which was marginalized before used that collapse (and rise of Israeli right wing that resulted) to take over Gaza strip after the fact.
So terrorists can come from a very weak position and manipulate public opinion very effectively. E.g. before 911 no one cared much about Bin Laden, now the "arab spring" is turning into religious governments who shape more after Iran than after the USA... As engineers they are remarkably effective at manipulating us (or really the dumber element of our society)

So the result for this would be a failure due to no index. Ugh. BigTable.

I've been using Codename One and it allows Linux developers to build native applications for mobile devices in Java while working fully in an open source stack. Very cool.

The main issues with FX are the same as those people have with ActionScript. The language is reasonable as a scripting language to tie things together but it feels very odd for most Java developers. Unlike action script you don't get the advantage of the amazing GUI tools Adobe offers.

Web start seems simple when creating hello world, NetBeans gets you started with a self signed JWS version. For production you will need a proper certificate which is harder but doable. Then there is the issue of pack.gz with its odd special case servlet (don't get me started on that one). Then you have the issues of HTTPs deployments not working for some people and the multiple incompatibilities between JWS versions installed which occur when someone has a corporate PC with a specific version of Java installed.
There are lots of issues some summed up here for older versions:
http://www.jroller.com/vprise/entry/should_you_use_jws

Newer versions of JWS (Java 6u10) broke different things but didn't really fix lots of the long standing issues. Most people using JWS just got burned one time too many.