I don't know an awful lot about hardware testing in this field, but I do know that Intel have released a hell of a lot of semiconductor technologies without anything blowing up in their face to the extent you suggest.
How about the Pentium floating point bug? Or the fact that the first round of support chipsets for Sandy Bridge have SATA ports that fail?
To be frank, in this day and age I would be dumbstruck if Thunderbolt isn't hot-pluggable.
The issue isn't really whether the PHY layer of Thunderbolt can handle hot-plugging. I agree - I would be gob-smacked if hot-plugging (or unplugging) the cable fried your motherboard.
The issue is software. Here's Intel's simplified diagram of Thunderbolt:
When you hot-plug (add) a Thunderbolt device, in some sense you're adding PCIe slots to the motherboard, with cards in those slots. When you unplug a Thunderbolt cable, slots and cards disappear.
The only two announced devices coming for Thunderbolt are RAID arrays. Internally, those will have a PCIe SATA controller with RAID capability.
What would you expect to happen if you took the cover off your Mac Pro and removed a PCIe RAID card while it was running? What would you expect to happen if you added a PCIe RAID controller with attached disks while Apple OSX was running?
I think for most people who understand the situation, the answer would be "kernel panic".
So, I wouldn't expect a puff of smoke (a literal "Thunderbolt") when hot-plugging. I wouldn't be surprised, however, that the word "reboot" comes up a lot - especially initially - when discussing hot-plugging.
For example, I use eSATA devices. With Windows 7/Server 2008 R2 - no problem with adding an eSATA device, it simply appears (and asks if you want to open explorer on it). Removal - you should first unmount the drive before unplugging it. (If you don't, you may get the "need to run chkdsk" when you reconnect. The chance of data loss with NTFS is extremely small.)
My Windows Server 2003 system (actually WHS), does not see the drive appear when it's connected. It's connect then reboot. To remove the drive, I restart the server and unplug it during the POST screen.
I have no idea how XP/Vista/Server 2008 handle the situation - my only instances of those systems are in virtual machines, so they don't see the raw hardware.
Any minor inconvenience there is from managing the chain is massively offset by being able to put the laptop on the desk, plug one cable and use the lot of the peripherals instantly.
Unless, of course, you need to shut down your laptop both to connect it to the chain and to disconnect it.
You can have a clap of thunder, or a bolt of lightning, but there's no such thing as a thunderbolt! Even the abbreviation, TB, is a disaster. Everyone will think you're talking about tuberculosis. :-/
And "
clap of thunder" brings up thoughts of another disease.
Yes, I understand these points. The point is that it's not native to the H67 or P67 chipset. That's all. Herego, apple isn't forced to have a dedicated USB 3.0 controller and then it's not an issue of them simply not offering a port on the machine for it.
But, the
ThunderChicken ThunderBird Thunderbolt controller
isn't native either! It's a humongous separate chip.
It's gob-smacking to hear arguments that "USB 3.0 sucks, it's not native - I want Thunderbolt".