Slashdot Mirror
Clash of the Titans Over USB 3.0 Specification Process
Ian Lamont writes "Nvidia and other chip designers are accusing Intel of 'illegally restraining trade' in a dispute over the USB 3.0 specification. The dispute has prompted Nvidia, AMD, Via, and SiS to establish a rival standard for the USB 3.0 host controller. An Intel spokesman denies the company is making the USB specification, or that USB 3.0 'borrows technology heavily' from the PCI Special Interests group. He does, however, say that Intel won't release an unfinished Intel host controller spec until it's ready, as it would lead to incompatible hardware."
Ever the more reason to never give up Firewire until they pry it from my cold, dead fingers.
So will this mean in the end we will have 2 competing USB standards? USB-Intel and USB-AMD? I can only hope that one will get picked over the other before it appears in most products because after the whole HD-DVD and Blu-Ray thing it would be an absolute pain to get a computer with USB-Intel in it when all the products will be USB-AMD.
Taxation is legalized theft, no more, no less.
How is this article, published online by an employee of a company supported by Intel, not biased in its analysis of the situation?
As we have seen with wireless networking gear in the past companies are all too eager to screw the consumer with incompatibilities because of pre-spec products being released. If Intel was doing this I would say good for them, its rare a company would actively try to protect the consumer from these vultures.
This is a replay of the OHCI/UHCI host controller interface standards of original USB.
This does NOT at all effect users, only driver writers.
What is being forked is the USB driver interface, and does not effect device compatibility at all.
As mentioned above, there were two driver interfaces for the original USB standard, and the only people who knew were driver writers and nerds compiling their own custom kernel.
This is blown way out of proportion, and doesn't effect 99.999% of us. Nothing to see here, move along....
Blessed are the pessimists, for they have made backups.
The fine article doesn't have to be bias free. We'll cover every conceivable side of the issues in the slashdot comments, and much irrelevance also.
My personal opinion: USB3.0 is cool, but give me external PCIe v2.0 x16 for the win. And Natalie Portman slathered in hot grits, of course.
Help stamp out iliturcy.
Sounds like the whole OHCI vs UHCI battle of days past is about to repeat itself, this time with a slightly different cast of characters. What a hassle for OS vendors...
Once again, we'll have the VHS version and the Betamax version.
One will win. Avoid whichever one Sony gets behind.
Help stamp out iliturcy.
Who thought of this movie upon reading the headline?
So much for the 'U' in USB.
I couldn't possibly be Twitter - I wasn't replying to myself.
Anonymous Coward
IIRC when I was at the early USB 1.0 developer conferences Microsoft let it be known that they would not support any other Host Controller Interfaces besides OHCI and UHCI.
I agree... the battle just heating up, how can you be biased? Not until there are two definitive sides can you get behind one or the other.
... even if some of them have been OOXML'd lately. (that's not even valid in Roman Numerals)
This does point out one thing, there is a lot to be said for open standards
No matter which version is better technically, if there is one that is not backwards compatible they will have an uphill slog trying to sell it. Yeah, I know, CDs were not backwards compatible with floppy drives, but this is a bit different. If the connector is the same, it MUST be compatible or my aunt nelly will kill someone.
Support NYCountryLawyer RIAA vs People
This isn't about competing USB 3 standards -- the spec is being designed by a group, and there is only one. This is about the design of the hardware used to implement a host controller that can comply with the spec. This is something that any company can develop if they want to, but since Intel is going to license their design of the host controller for free, most companies will just wait for that design and use it to implement USB 3.
..wayne..
USB being cheap crap is a good thing. USB puts the complexity on the host end: great for cheap peripherals.
Personally, I think USB should stop at the theoretical 480 mbps.
The two sides I see here are not Specification A and Specification B but not producing an open standard and producing an open standard.
"there is a lot to be said for open standards"... Yes, Something indeed. Who lead the CD revolution? Sony. Who developed the standard? Sony (and Phillips). They released the standard after they had working products to sell. The "standard" still then cost a lot of money to even look at. (See the wikipedia article on the Red Book standard).
My Point (finally?): Giving the excuse of "we don't want to release it early because then there will be incompatibility issues" (paraphrase) is complete bunk. No company in their right mind would implement a pre-standardized hardware specification (and sink mucho dinero into the manufacturing costs of just the parts to make the parts). And if they do, they aren't AMD/nVidia or Intel. [1]
It would only help in that the other parties would be able to help improve the standard before it is released. Oh, and have equal footing with Intel too, since they would be sharing equal responsibility to creating it.
[1] Counter argument: HD-DVD and BluRay. Nope, that case is an argument FOR what I am saying, not against. If they both would have worked together to produce an open standard, instead of trying to beat each other completely, they both would have had the right product and we, the consumers, would be able to have real competition in the hardware sector.
That is all.
ROMANES EUNT DOMUS
USB is an ugly hack upon hack upon hack. It never really does work right because alot of devices are deviant and/or require special purpose BIOS support (*cough* Toshiba). You can buy USB compliance analyzers only to discover that devices still don't work.
The software support is ugly as heck---remember how many years it took the Linux kernel to get it right? And even then it's a bloody nightmare of random device-specific drivers. (How many protocols do you need to get bits in, out, or through a device anyway?)
USB 3 will only make life worse. Can't we just throw it away already and move on to something cleaner, without any legacy cruft?
(I mostly agree with the rest of it though)
Interestingly (or not) you demonstrate a logical understanding of the technology marketplace. To paraphrase you, if I may, Intel and AMD are fighting about who gets to piss on the idea of competition creates value for the consumer. Any space where AMD and Intel are competing is full of this, and not inconsequentially, lawsuits. Intel has been partnered with MS for a long time, and they worked hard to be the hardware version of what MS was to software.
We can detail the lawsuits ad nausea, but my point is that anyone that was a healthy partner with MS has done to their industry what MS did to software. Like that or not, it is true. In the end, we have Mr Gates to thank for this, no matter how philanthropic he may try to be these days. I wonder sometimes how far exactly he has set the human race back from what will eventually, and necessarily be.
Though that is sort of scifi philosophy, it is true. In the name of riches, the advancement of technology has been slowed, deliberately, and with malicious intent against the betterment of mankind. In this way, I find his generosity a bit pale these days.
Open standards are indeed the ONLY way to create technology and advancement that will last and actually advance mankind in a direction that betters all of us. Despite the socialist sounding tone of that, it is true. We are all better for the sharing of technology from the space race. Technology, and specifically computing/networks are still in the hands of those that would derail it's benefits if there is profit in it. There are those that are trying to change this situation, but it is slow going. Even hardware manufacturers are hobbled by things like the DMCA and it's ilk around the world. Sometimes I'm sad to say I'm American.
Fighting against the 'right thing to do' for the sake of money is not in the best interests of the community, and in the end, it hurts your business. Customer is king, so they say, and when you put hurdles in the way of a complete and exemplary experience by the end user, you harm your business in some way, if not in big ways. It's unfortunate that not enough people will understand that the competitions in the technology markets have hurt them, and they will not understand how to express their frustration that older USB devices won't work with new USB hosts. It will be just one more black magic thing they don't understand about technology type things. They will go to PCs R Us and buy whatever the best they can get happens to be, hoping that it works for a couple of years, not unlike car buyers. So for profits, businesses promote the throw-away society. When there is something new, throw the old away, don't upgrade, don't re-use. How is this helpful to the human race?
Well, just some late night thoughts about this whole thing, and the absolutely ignorant waste it makes of the world.
BTW, there is hardware space competition.... if you are willing to build your own and not buy what the idiot^H^H^H^H^H salesman tells you at worstbuy.
sigh
Support NYCountryLawyer RIAA vs People
Intel has a point: releasing documentation on a non finalized standard creates a fluster-cluck of bad implementations that aren't necessarily compatible with each other. IIRC, isn't that what's happened to 802.11n, pre-n, draft-n, n-ready, looks a bit like n in a dress, MIMO, etc. which just confuse the crap out of a consumer already pissed at USB 2.0 HiSpeed and USB 2.0 FullSpeed crap.
nVidia has a point: Intel not telling anyone else until the last moment would, indeed, give Intel an unfair first mover advantage.
Obvious solution: Release the pre and post release specs with an agreement attached that anyone wanting a copy has to sign. An amount of time that gives everyone a fair chance to get product ready is picked after final specs are chosen. Anyone gaining access to the specs agrees not to release until that time period has passed. Now no one releases incompatible hardware and no one gets an unfair first mover advantage.
If any one of them was really smart and wanted to name it to win, they'd name it either blu-port, usblu, or usb 4.0. I mean seriously, which one are you going to use? One named USB 3.0 or 4.0?
Google's Super Secret Search Algorithm: SELECT @search_results FROM internet WHERE @search_results = 'good'
I was just about to reference to OHCI/UHCI myself. However, since both Intel and AMD are friendlier towards Open Source than before, implementation should not be an issue. That is, as long as we keep Microsoft out of the picture.
This may seem like the odds are stacked in Intel's favour, and I'ms sure they thing so too, by not allowing anyone else near the host controller spec. Despite this, I think that the other board members would fully realise that Intel is a minority against the combined force of AMD, Via, SiS and Nvidia in production of chipsets for desktop PCs.
The USB-IF board knows the danger of losing control of the standard if it is forked, so I'm sure they will in no uncertain terms tell Intel to invite AMD, Via, SiS and Nvidia to the party. Of couse Intel could ignore this and keep going, but it risks finding its spec left out of the USB 3.0 standard
USB2 is quoted as having 480Mbps throughput, however as the grandparent points out USB2 is not a fully-fledged I/O controller just the PHY layer, the host having to do all the heavy lifting.
The upshot is that when you actually use one bus or the other to, say copy files, firewire at a mere 400Mbps trounces USB2 in throughput.
Yes USB3 is in the pipe with vastly improved on paper specs, but then again Firewire has 3200 and 6400 variants in the pipe as well.
Essentially USB should have been left as an interface for keyboards and mice, and firewire aught to have been adopted by intel as the preferred bus for all high throughput applications, it would also have been preferable to SATA.
Firewire might pay for itself in high speed applications where time == money, but it is sever overkill (and too high cost) for many lower speed applications such as mouse, keyboard etc. USB is king of the low speed domain because of low cost: a USB-cappable microcontroller only costs a couple of bucks and a sub dollar micro can do a low speed bit-banged implementation of USB. Adding USB to peripherals is almost free.
Engineering is the art of compromise.
...it's the same HD-DVD vs. Blu-Ray / DVD-R vs. DVD+R shit, except now it's cables.
The only people getting screwed is the customers.
Random Thoughts From A Diseased Mind (Not For Dummies)
Better 2 specs than 50...
Look how many different Ethernet, IDE, Video, Wireless drivers etc, you need to have...
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Please pass me some of what you're smoking!
Quality, performance, value; you get only two, and you don't always get to pick.
Furthermore I think most of the cost advantage of USB comes from economies of scale since so much more USB equipment is manufactured. Once the chips and other equipment is designed and being manufactured, the cost falls as the fixed costs are amortized over an increasing number of units. The difference in variable costs between firewire and USB are trivial past a few million units. Produce more firewire devices and the cost will naturally fall as you spread the fixed costs over more units. Yes firewire is slightly more complicated but not so much as to make a big difference once minimum efficient scale is reached.
As mentioned above, there were two driver interfaces for the original USB standard, and the only people who knew were driver writers and nerds compiling their own custom kernel.
And people who use drivers and custom kernels. I didn't do either, but I still had to deal with users whose USB controller cards didn't work with NT4 because of driver conflicts, and then there were devices that had custom drivers that only worked with one or the other chipset. It wasn't until Windows 2000 that USB was really reliable on Windows.
Open standards are just one way to create technology, and not necessarily the most efficient - Look how long 802.11n has been sitting unfinished. That method of technology advancement works fine for incremenatal evolutionary changes, but disruptive technology just does not lend iteself to industry committees. Real revolutionary advancement doesn't happen by committee, it happens because somebody decides to take a risk and do something different. Such technological shifts do lead to incompatibility, but it's a small price to pay to break out of the grasp of entrenched powers.
It'd be nice if everybody knew what the "right thing to do" was, but it's not always clear. Get 10 industry leaders in a room and they will likely have a different view of the "right thing" than millions of customers.
D6 63 0D 70 89 81 BB 8E 7B 7C 5F 5D 54 EA AB 73
Affect. Affect. That is all.
So basically, USB 3 the real spec is available. This defines the USB interface. Any host controller designed that meets the USB 3 specification would work just fine. If it does not follow Intel's Host controller spec, it will just need to use a different driver. All USB 3 client devices should work with all USB 3 hosts, regardless of what standard the controller chip uses. Is that all correct?
If so, I really don't see what the big deal is. AMD can create their own host controller spec, or they can wait for Intel's free one. Since Intel is publishing this spec for free use by the others, I don't see what is wrong with them having them release to market first. Not releasing at the same time as Intel is simply the cost of using Intel's spec, rather than creating your own.
Or am I missing something significant here?
Stylish sheet to fix many problems in Slashdot's D3: https://gist.github.com/801524
FireWire needs DMA, which is toothed as a security risk by high paid "experts" every now and then. I think it was up on Slashdot for the n-time a month or three back.
FireWire can read all your memory! People with haxor Firewire skills can steal all your data! They just need physical access to your computer - i.e. come into your home and plug something into your Firewire port! What ever will we do?!
No. What Bill Gates has done is to encourage low quality shitty software. Go read the book "The Software Conspiracy". IBM's mainframe used to crash all the time too. Then, they put someone in charge will the balls and authority to fix that problem. That man later founded the Software Engineering Institute at Carnegie Mellon. Microsoft could have learnt from that. Instead, we get Windows. with 85000 APIs and counting. Why is there a need for *5* (at the very least) APIs for doing a directory listing?
the linked articles excluding the intel one are aimed at low level users. This is just Navida and others making intel look bad 1. No one needs to wait for intels Host controller specification it will be compatible with all the other Host controllers using the USB 3.0 spec 2. It sounds like intel is doing the industry a favor by showing over companys a reference on how to implement the spec
null
for many lower speed applications such as mouse, keyboard etc. USB is king of the low speed domain because of low cost
No, we must have one true bus protocol! How dare you suggest the appropriate technology for the job?!
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
While I'd be the first to agree that most software is of incredibly poor quality, this just so happens to be what the market WANTS, as evidenced by the huge success of Windows 95. Bill Gates simply realized this, produced a product that was good enough, and the market rewarded him appropriately, as is the natural way things work.
.NET, which contains the cleanest Windows API available, and is quite a nice platform in general, too. Perhaps you'll like its API for getting a directory listing.
Since '95 Windows quality has improved, of course, but it's still focused around the needs of its actual users rather than what a bunch of smelly software geeks prefer in an OS (and those geeks are probably not running Windows anyway). You think typical Windows users give a rat's ass about how clean the API is?
On that topic, the reason Windows has a large number of APIs is because it keeps evolving, and improved APIs take the place of old ones. The reason old ones are kept around is for backwards compatibility, and to accomodate different people's choices of programming languages. When I need to write Windows apps, I use
Quality, performance, value; you get only two, and you don't always get to pick.
When are we going to end these interconnect wars by going optical or optical with power for both peripherals and internal busses? Last I saw http://www.networkworld.com/community/node/25514 IBM was working on 8 tera bits per second. I'm sure this is more than what is currently needed but I'd hate to pull a Gates and drastically underestimate needs.
Be as you would have the world become.
Only n00bs use anything other than Transparithermite for their viewing window. For additional security, the system is also cooled by liquid propane, and the +5V rail on the power supply is custom engineered to deliver 480V if anything trips the ground fault circuit interrupt.
It's cost me $30k, eight different graphics cards, and two girlfriends, but the peace of mind that a truly secure setup provides is invaluable.
"We have to go forth and crush every world view that doesn't believe in tolerance and free speech." - David Brin
And if win32 is killed, and you can only do .net, then yes. But as of right now, your new sports car is towing that AOL bus with people throwing body parts at you.
And I'm not sure how DRM and a host of other shitty stuff is consider a feature that users wanted.
You seem to be under the impression that just because the API is a bit unorganized that the underlying code must somehow be just as bad. I don't think this is true; much of the old or deprecated APIs are likely just wrappers around the newer stuff. Just because Win32 is written in C, that doesn't automatically make it bad. Just old.
.NET wraps Win32 - the Windows-specific parts do, anyway. That means it can't replace Win32. Microsoft could hide the Win32 bits and force everyone do code in .NET, of course - not that it'd be a very popular move.
And
I'll note that pretty much every modern OS (seeing as most of them are UNIX-based) have somewhat old-fashioned underpinnings, with lots of ugly little bits sticking out, but usually work fairly well in spite of that.
While I have no love for DRM, I understand why Microsoft included it in Vista - people likely wouldn't have been able to play their legally purchased BluRay/HD-DVD discs otherwise. If you don't use DRM-infected media, the DRM in Vista won't bother you.
Quality, performance, value; you get only two, and you don't always get to pick.