Launching the Eclipse IDE you have the best of both worlds. First you wait up to 30s for the splashscreen to show up, then another 1-3 minutes with the actual splash screen displayed and program loading your 200 plugins, then 5min for the workspace to (re)compile. Then it's time to kill it because it has deadlocked !

There's many reason why Honeycomb doesn't attract developpers

- first, most apps make money through ads since the average Android user won't buy any app ever, in contrast with iOS. "Apps cost nothing to make, heh ?".
For ads to make money, apps must be downloaded massively (> hundred of thousands). And there's lot more phones out there than tablets as phone are way more ubiquitous than tablets which remain an expensive toy.
The fact that most Android users just wouldn't buy Honeycomb apps is a real problem. Android user wants everything free. Android user won't have Honeycomb specific apps because of that, his fault.

- Making an app optimized for Honeycomb using its APIs is very time consuming. Most of the time you can easily adapt an existing phone app so it scale and handle appropriately on Honeycomb without using Honeycomb custom APIs. That's a great difference with the iPad, on which most iPhone app looked like crap because of no modern layouting taking into account scaling.
On Honeycomb, the new Fragment APIs are kind of complicated and require a lot of refactoring which some developpers may be reluctant to do for various reasons. At this point you'd better write the app UI twice, externalizing the app logic in Library.

So at the moment it makes no sense to write Honeycomb specific apps: it is too much work for too little benefit. Now if the number of tablets grows significantly it might change but nothing is certain.