That price is being paid now, inserting t-bolt moves some items from one box to another. Adding t-bolt is the big cost.

Eventually cost factors will drive the creation of specialized t-bolt mashups that handle all the functions needed for a type of peripheral.

T-bolt is still copper for now, I suspect that fibers and optics will benefit greatly from cost reductions if they wind up as mass (100 millions) produced components.

Ugh! I just had a really ugly thought.

If all your gear is built for copper, what happens when the standard does migrate to optics?

You buy all new gear? A modem?

Suspect it depends on your need for speed.

I really wanted to use references to the old Charlton superhero lineup (these days probably best known for being the basis for the characters in Watchmen), but that would have been a little too obscure, methinks.

Just move your copper devices to the end of the chain I expect.

The plan is to build the optics into the cable. So each device will still keep the existing copper based connectors, and when someone needs a longer cable run, they buy a more expensive cable that contains 2 copper to optical converters. One port out of the system could have the long optical cable run out to the first device, then copper cables between the rest.

This style of cabling has been out in the storage and networking world. You end up with a copper cable with two GBIC connectors directly attached.

My eyes are about 28-30 inches away, and thats about the best I can do with the current desks we have. The two screens are already pushed against the cubicle walls, and the desks don't offer keyboard trays. I did have my desk lowered to ensure better ergonomics for typing, and have the monitors on stands.

We are moving into a new buildout with mo space, so I should be able to arrange the desk better to take advantage of the full space of both monitors without having to turn my head so much. At home I usually sit around 40 inches away from the 27 inch display.

New idea / new worry (dinner conversation with other computer security geeks can be dangerous):

PCIe is the protocol used by hardware devices to speak to the rest of the computer, and includes direct access to RAM (DMA). It's previously well known that the Firewire ports on most computers can do the same thing and can form the basis of an interesting physical-access attack to a computer. (It's also used for remote kernel debugging.)

Thing is, when your monitor port now allows the remote peripheral to DMA with your computer's main memory, you've got a new threat... the hostile monitor/projector threat. You're traveling. You connect your computer to the hotel's digital TV or to the conference's projector. What if that projector is evil? Now it gets to interrogate your computer's memory or talk to other peripherals like your disk.

There's just no end of risk here. The only saving grace, I think, is that Xilinx FPGAs aren't fast enough to keep up with the raw data rate on the port, which means you'd have to have an appropriate driver chip. I'm sure Patrick could probably ballpark the cost and components of a suitable daughter board to shove into the back of a TV that would be able to make DMA queries and save the results to a memory card. I'm guessing it's not that hard to do.

What I'm less sure of is what you might do, in your computer, to deal with such threats. Can you imagine your computer popping up some kind of dialog box? "Hey, there's a new device out there. Allow it to connect?" I'm sure those would be popular with users.

Nothing to add here, I just wanted to thank you, Dan. My day has started off well with my favorite sentence I've heard all month:

"What if that projector is evil?"

That sounds like the least scary Stephen King novel.

<Gratuitous>In America, you connect to projector. In Soviet Russia, projector connects to you.</Gratuitous>

Some kind of public key trust relationship that coincidentally implements HDCP?

The crypto would be straightforward, but I'm less confident we could deal with the security policy problems. Could users properly deal with a dialog like this?

"The device `Sony XBR46KBR9200Q' requires access to your system memory. <Allow> <Forbid>"

You could try to "pre-answer" these questions by having a trusted third party that "certifies" hardware to speak the PCIe protocol. That gets you into a world not unlike Apple's app store, where apps must be submitted and centrally approved by Apple. If you did this "properly" and got it to really work, then it would be effectively impossible for tinkerers to build Thunderbolt devices. Needless to say, that's not entirely desirable.

Yeah. I was being at least a bit facetious. Unfortunately, I can see the MPAA and RIAA lapping it up.

Bought one of these Winstar USB graphics adapters:

http://www.amazon.co.uk/gp/product/B0037...ASIN=B0037ZCEFO

It is much smaller than I expected. The hardware, packaging, manual etc are totally generic (but oddly high quality). There isn't a single mention of Winstar or indeed any other company on it. The back of the manual shows images of a 8 other DisplayLink products that are supposedly produced by a series of different companies. I suspect in actuality that they are all made for/by DisplayLink themselves.

It ships with a DVI-VGA adapter and a funky DVI-HDMI adapter with a joint in it that rotates 180 degrees (which would be handy if you were trying to cram it in behind a TV somewhere).

I didn't bother using the driver CD supplied, Windows 7 downloaded some drivers for me.

It works well. I had expected it to use up noticable CPU time, but it doesn't seem to.

Windows are probably a tiny bit laggy when dragging them around, but that could easily just be my imagination.

It copes happily with typical in browser video on Vimeo and Youtube. It copes happily with 360p video in VLC either windowed or full screen.

I copes ok with 720p torrents in VLC at 1:1. It starts to struggle with 720p zoomed to full screen, but then do does my laptop video when it is running on the internal GPU.

I think when it is playing 720p it starts dropping frames, but does so very smoothly (that might well also be my imagination as well).

All this was on my Core 2 Duo 2.2GHz laptop with 6GB of RAM running Windows 7. Will try it on OSX later.

All in all, very pleased with it. If I didn't know it, I wouldn't realise I wasn't still running off the docking station video card.

I'd be very interested to see how you get on under OSX, I'd love to have another usable screen on my mini.

I've done a very quick test under OSX on my Mac Mini.

It generally behaves very similarly to under Windows 7, however it uses a bit more CPU and can't really cope with 720p at 1:1 (fine zoomed to 1:2).

The comparison is slightly skewed, as my Mac Mini is 2.0GHz compared to 2.2GHz on my Win7 laptop. But they normally feel about the same speed.

Apart from heavyweight video and slightly laggy window dragging you'd never notice it wasn't a native display. The adapter I've picked was the highest speced of the DisplayLink hardware, not sure if that makes any difference to the speed when running at 1920x1080 as I am.

It is also very plug and play, i.e. to unplug from Mac, plug into Windows7 only took 10 seconds or so. Makes for a handy poor mans DVI KVM

Sony Vaio Z series, now with a Thunderbolt powered external dock containing a GPU, Blu-Ray drive, and some more ports. Sadly Sony found a way to screw it up though, as they are using some proprietary port, and one USB 3 port to run the dock, instead of the standard Thunderbolt port. Still, it's a start.

Sonnet also announced some Thunderbolt to PCIe devices at NAB.

Some interesting GPU testing from HardOCP on needed PCIe bandwidth for gaming.

This is looking pretty promising for my next gaming capable system to be a laptop again, without sacrificing too much in the gaming performance space.

Hmm. Interesting. I was actually looking at the new Z series as I need a new laptop anyway. I didn't look too closely at how the special dock attached but its good that its Light Peak/Thunderbolt despite a typical Sony custom connector.

Well it was good for a while.

Unfortunately OSX 10.8 broke* the DisplayLink driver. DisplayLink have know this for months, but rather than releasing a fixed version of the driver, they are pushing forward with a new re-architected driver (which they need to support USB 3 for their new chipset on OSX).

They say they "hope" to deliver this new driver in Q1 2013 and they expect it to fix the problem. I'm quite sure it isn't likely to bring any new bugs along with it...

In the meantime my monitor remains blank

DON'T BUY DISPLAYLINK IF YOU WANT TO USE IT WITH OSX.

* two things are broken, the login window is blank and it crashes Finder and Skype repeatedly. The login window I can cope with by tabbing around blindly, the Finder and Skype crashes are a showstopper

I don't know of a single app that doesn't crash semi-regularly in Mac OS X 10.8.x

That includes every App Apple ships, which in my experience, crash far more often than 3rd party.

10.8 also seems to have shit-loads of USB issues, some of which I've been informed of by customers with third-party hardware. Related to my own hardware, I have to do some checking this weekend whether 10.8 supports USB low-power for my devices at 100mA. I think it might be turning them OFF completely when sleeping.

I don't see the situation getting any better with Apple having moved to a 1-year release cycle and fucking with so many low-level parts of the OS willy-nilly.

Apart from this specific issue with the DisplayLink driver 10.8 is no less stable for me than 10.6 which is what I have recently upgraded from.

I can still kill it using VirtualBox if I do the right/wrong thing, but that applied equally to Snow Leopard.

The only Apple app that I really remember crashing since the upgrade is Xcode and that has always been reasonably crashy for me (at about the same level VisualStudio was until a couple of years ago, another area where Xcode lags VS). All in all I am perfectly happy with ML, apart from this DisplayLink driver problem (though the new file handling behaviour still seems arse-about-face to me, "delete" just seems like the wrong text for the button for a start).

Safari crashes multiple times per week even now that Flash is completely uninstalled. Same goes for Mail (though the crashing is the least bothersome of its issues).

I don't use Keynote except once in a blue moon, but when playing around with it earlier this week it was crashing multiple times per hour just sending it AppleScript that according to its help documents it supports.

iTunes craps out every now and then but nowhere near as often as Mail or Safari.

It was a challenge for the groups at Apple, and third parties to keep up to date with slower OS update schedules. What happens next year when 6 months after updating their driver, DisplayLink is looking at another brand new OS?

Right now another PITA is Transmit (FTP) being used with TextWrangler as the default document handler. I can't open multiple documents at a time into TextWrangler from Transmit - neither local nor remote. I can open multiple docs into other programs and I can open multiple docs into TextWrangler without Transmit. If I keep repeating the document open Transmit will eventually just get stuck. I can only think of reporting the issue to both developers.

I don't use Safari, Mail or Keynote

(well ok I occasionally use safari to test websites)

I really have no idea what you do to have so many issues with so many things.

I use safari pretty much all day every day and I really cannot remember the last time I saw it crash.

Same for mail, its constantly there in the background ready for action.

My machine is hardly ever shutdown and is just put to sleep by closing the lid.

The only thing that irks me with mountain lion is the god awful built in versioning on files - its modus operandi is confusing and random.

Not that this helps in any way.

Here are potential problems for crash-related bugs:

1. My machine is a 2009 NVIDIA-based MacBook Pro 17"

I don't think Apple is testing older machines with the kind of effort they used to. We used to test sleep-wake cycles into the thousands and even one failure was not acceptable. This machine very often will not come out of sleep properly. Closer to 1 in 10 than 1 in 5000.

2. I use the machine in clamshell mode a lot.

3. I have replaced the default HD with an SSD. And replaced the optical drive with the original HD.

4. I mount and try to keep a few AFP shares accessible all the time. Mountain Lion randomly drops them while the machine is asleep (or when it wakes - this can cause all kinds of momentary process freezing (which can last for a few minutes).

I see Mountain Lion issues reported by customers all the time, so I also have perspective from others. The many Mail bugs I have documented don't have anything to do with hardware - maybe I'll post them again.

I've got a 2011 MBP, I also use it in clamshell mode with my thunderbolt display all the time. After upgrading to 10.8, the thunderbolt display refuses to wake up from sleep. I've been trying to work with Apple for 3 months now without any progress. They've swapped monitors, had me format, clear every bit of ram, and still... pretty much 95% of the time waking from sleep the thunderbolt display is a $1000 paperweight.

I've got the 27" display right from right before they switched to thunderbolt. Same issue. The machine will wake but the display doesn't come on. Then it will go back to sleep.

I'm using BT Apple keyboard and trackpad with it. What I do in this situation is plug and unplug a USB device to try and force the machine to wake. It works most of the time after 1 or 2 plug/unplugs.

But sometimes it happens just with the built-in display. For instance, I'll sleep the machine and then unplug it from its Henge dock which removes all external devices at one time. For my typical setup that's the display, the display's USB cable and a USB audio device (which isn't being used) plugged into the display. When I then open the system up I'll be greeted with a black screen. Every once in a while I can close it and put it to sleep again, then open it and have it work. Most of the time when this happens I have to force-restart it to get the display back.

When I worked for ATI, always on Apple contracts, if this happened 1 time in 3000 attempts, it was a block ship. Period. We'd be there working 12-16 hour days through the weekend until it was found and fixed.

I have a 27" display from the other end of the price spectrum, and once in a while the firmware in it gets confused about power states and just stays off.

Unplugging/replugging the AC cord brings it back to life.

The monitor I had before this one, totally different brand/model, did the same.
Weird. Like everyone is copying everyone else's bugs.

Hmm.. I'm going to try cutting the power to the display next time this happens. Don't know why I didn't think of it before. Normally I would unplug and replug the display but since the cord is attached to the stand/dock I can't do that.

Still waiting for a working driver for the DisplayLink adapter. I've now given up and bought a 27 inch monitor instead.

Not good DisplayLink, not good (after a lot of badgering we did at least get them to put a note on the driver download page saying there are 10.8 problems).

Don't buy a DisplayLink adapter if you need to use Mountain Lion or value on going driver support in the future.

Now, where are all those Thunderbolt display adapters we were promised