Expert to Expert: Meijer and Chrysanthakopoulos - Concurrency, Coordination and the CCR

·White paper on Designing Efficient Background Processes.
·PDC talk on Designing Efficient Background Processes.
You can find more contetn in the PDC sessio: Designing Efficient Background Services --> https://channel9.msdn.com/pdc2008/PC19/
Firstly, Charles & Chittur - thank you for a really interesting video.
It's great to see this sort of thing is being worked on - from my experience (I've got the Win7 beta on my laptop) it appears to be helping and I'm pretty impressed with the speed improvements (things are not 100% but that's par for the course with something
like this).
Also, do you think now this more mature design is available, you (Microsoft) might be able to try some "tough love" with those hardware/software manufacturers who aren't doing a good job. I'm thinking of the sort of scenario where you have some gadget that's
incredibly useful, poor quality code probably won't make me stop using it (unless it's seriously flawed) but it drags down my whole machine to some extent... and all because they've got some lame service running permanently that needn't be.
Typically speaking, even if you are able to pinpoint the performance on the item (which isn't always the case unless you're more technical than I am, but you can get a fair sense that it's negatively affecting things) you will get ZERO response from the software
writer's / manufacturer's support - they already have your money, the guy dealing with support might not understand the issue or he might be jaded since there are countless false problems so he's going to ignore your legitimate problem. Maybe sometimes they'll
be dealing with it internally, but you won't ever get any acknowledgement.
It would be great if you got onto their case either discretely "having a polite word" or (preferably to my mind, since it's more open) by publishing lists of widely used services where design could be improved by simple adjustments (i.e. run as a task under
task host).
Two particular cases that spring to mind, which must affect huge swathes of PCs:
Sun's Java update check service
Adobe's Acrobat Reader update check service
These two things run always - they have a legitimate purpose (yes, it's good to get updates on these bits of software) but it doesn't merit their own schedulers starting every single time I boot Windows! Also, if they used Task Scheduler, you could tweak the
frequency of checking yourself (and there's no need to learn a new UI - Task Scheduler exists!)
Anyway, just putting my two cents in...
Cheers,
Neil