Is this the USB3.0 spec, or is it the USB3 HOST CONTROLLER INTERFACE spec?
In usb1 times (1995), there was UHCI (from Intel, which wanted royalties, only VIA licensed it and they broke it, so two versions of UHCI) and then there was OHCI (from NEC, and copied by everyone else in Taiwan, SiS, Opti, ALI, etc, and a few of these had variations that made driver writing tricky as well). If you have ever seen a USB device which works on one chipset, but not another, this was why. Anyone making a device had to test against all this crap, which broke the promise from the USB-IF: one bus, one physical interface, and no need to test against multiple implementations. The same promise came from Microsoft with the Windows Driver Model (WDM), one driver for Win9x, and what would become Win2000, and every OS going forward - but they had to make two USB HCI drivers just like everybody else, which meant everybody else had to test their drivers against these hardware differences too.
Then with USB2 this got straightened out. One Host Controller Interface: EHCI. Everybody did EHCI, but Intel did most of the work, and got no royalties. Only one implementation that most everybody got right, yay if you were a hardware or driver developer!
Now we have USB3. The USB-IF publishes the bus specification and is totally open. So, I imagine what everyone is in a tizzy about is the HCI, and Intel again wants a single HCI implementation. It is unsaid in all these one paragraph entries that it is the HCI that everyone is complaining about. If that is the case, then I'm rooting for Intel. Writing, testing, and tracking errata from multiple implementations of the same specification is for chumps. If Nvidia, et al, make their own HCI spec then they can DIAF - I won't develop my devices on it, and as a consumer, I won't buy hardware based on their chips.
Yes, reading the actual article, it is the HCI they are bitching about. Nvidia can choke on it. It won't take them two years to clone a design Intel is giving away for free.
It's not like anybody is shut out, Intel has documentation and prototype hardware on what they are doing.
Develop your devices on it? After the driver is written, there's no difference between the controllers. You don't have to develop specifically for either of them.
There might be differences due to different implementations, but that's true even with a unified controller spec. If you think that's not true, you've never tested your devices on an ALI USB 2.0 add-on card, which uses ECHI, but has a lot of bugs anyway.
Sony's just released a 15.5-inch addition to its VAIO S Series that not only adds a crucial bit of extra display acreage, but also bumps things up to a full 1080p.
The most commented posts on Engadget over the past 24 hours.
Now that we've thrown 'em off the trail, use the form below to get in touch with the people at Engadget. Please fill in all of the required fields because they're required.
Is this the USB3.0 spec, or is it the USB3 HOST CONTROLLER INTERFACE spec?
In usb1 times (1995), there was UHCI (from Intel, which wanted royalties, only VIA licensed it and they broke it, so two versions of UHCI) and then there was OHCI (from NEC, and copied by everyone else in Taiwan, SiS, Opti, ALI, etc, and a few of these had variations that made driver writing tricky as well). If you have ever seen a USB device which works on one chipset, but not another, this was why. Anyone making a device had to test against all this crap, which broke the promise from the USB-IF: one bus, one physical interface, and no need to test against multiple implementations. The same promise came from Microsoft with the Windows Driver Model (WDM), one driver for Win9x, and what would become Win2000, and every OS going forward - but they had to make two USB HCI drivers just like everybody else, which meant everybody else had to test their drivers against these hardware differences too.
Then with USB2 this got straightened out. One Host Controller Interface: EHCI. Everybody did EHCI, but Intel did most of the work, and got no royalties. Only one implementation that most everybody got right, yay if you were a hardware or driver developer!
Now we have USB3. The USB-IF publishes the bus specification and is totally open. So, I imagine what everyone is in a tizzy about is the HCI, and Intel again wants a single HCI implementation. It is unsaid in all these one paragraph entries that it is the HCI that everyone is complaining about. If that is the case, then I'm rooting for Intel. Writing, testing, and tracking errata from multiple implementations of the same specification is for chumps. If Nvidia, et al, make their own HCI spec then they can DIAF - I won't develop my devices on it, and as a consumer, I won't buy hardware based on their chips.
Yes, reading the actual article, it is the HCI they are bitching about. Nvidia can choke on it. It won't take them two years to clone a design Intel is giving away for free.
It's not like anybody is shut out, Intel has documentation and prototype hardware on what they are doing.
Develop your devices on it? After the driver is written, there's no difference between the controllers. You don't have to develop specifically for either of them.
There might be differences due to different implementations, but that's true even with a unified controller spec. If you think that's not true, you've never tested your devices on an ALI USB 2.0 add-on card, which uses ECHI, but has a lot of bugs anyway.