Appeals judges toss out FTC win: What Qualcomm did to its rivals was 'hypercompetitive, not anticompetitive' An appeals court has reversed an earlier ruling that Qualcomm broke America's antitrust laws, dealing a blow to the FTC, which brought the case. Rather than crippling competitors with its war chest of patents and licensing demands, the US chip designer seemingly did nothing wrong, we're told. “Anticompetitive behavior is …
Too true, there have been more than enough examples by now of companies getting their patented methods included into standards and then racking up profits through 'interesting' licensing. I know we have FRAND and similar but they tend to be reliant on companies playing nicely which never works 100%.
How about the standards committee does some proper work to value the contribution of each piece of work to the whole standard and then set AND PUBLISH a price per use of that standard with a breakdown of who gets how much from that. Contributors get a guaranteed income and anyone wanting to use it has a clear up-front cost
Unfortunately, that's not generally how standardization works (the clue is in the name) - standards are usually based on the implementation of existing technology / IP that is then enshrined within the relevant standard (i.e., there is no innovation within the standardization activity itself).
For example, the 'C' language standard was initially based on the compiler implementations of "K & R" 'C' that were available at the time the standardization activity was started. Innovation takes place else where (often within open source compilers such as clang and gcc) and new features are then submitted for inclusion within updates to the standard. Most of this happens through the submission of papers to the standardization working group(s) responsible for the standard.
However, something like a Special Interest Group could (should) be established that can drive (and control) the innovation process - membership of such a group can require (within it's membership T&Cs) that any ideas may only be presented to the SIG when the SIG has unfettered, unencumbered rights to the IP (including the right to publish within a standard).
Standardization bodies are now much less likely to incorporate patent encumbered IP - mainly because of the recent problems arising within the mobile comms sector.
A really good example is the v.90 standard for 56K modems back in the mid 90's. There were essentially two companies out there that had a viable process: 3Com/US Robotic's X2 and Rockwell/Lucent/Motorola's 56Kflex. v.90 (and later, v.92) took bits from both (to hear USR's response, v.90 was mostly X2 but with some 56Kflex sauce stirred in; I imagine the same was said on the other side as well), and was incompatible with both.
I know that Cisco is one of the 40 tonne kaiju in the internet standards realm, which is certainly a pain at times, because they have their own special way of doing things.... (Microsoft was also one in the past, and probably still are.)
It would be more difficult to create a standard in this sphere without standing on someone else's patents than it would for a competitor to create a product that doesn't infringe.
All new tech in the#is area is based on previous products and research.Bundled up with that is a lot of patents. Even a SiG is loaded with lobbyist pushing their area and installing themselves at the approval areas as well as doing a lot of the grunt work.
Say there was a big overhaul and certain technologies or companies were not allow to be part of a standard - well a sufficiently resourced company like Qualcomm could just carry on producing non-standard but better tech and may well become a defacto standard.
Only patent reform could have any real effect - for a start 20 years for tech is ridiculous.
The length of patent lives is one factor that influences how much risk a company is willing to take when investing in R&D for new technologies. Spending billions when others could just copy your blueprints in a short time makes little sense without the adjustment of exclusivity rights to enable some return on that investment. So long as there is vibrant competition (as there most certainly was in the cellular implementation space last decade) regulators don't need to do adjust these parameters.
Having a capped, sliding effective discount of a fee adjusted for the intensity of standards usage under an open license is a much more fair practice than what the true semiconductor monopoly: Intel did in the 90's and 00's. It didn't competitors through product competition via engineering efficiency, but through fiercely guarding its IP instead of making it a license-able standard and suing competitors who legally reverse engineered solutions into submission by swamping them in legal fees. They also disabled their industry standard compiler's optimization compatibility with competitors' explicitly binary compatible chips, and used entirely exclusionary rebates to deny contracts and exposure to AMD at top tier OEMs (a similar arrangement with Apple was the one lone point the judge writing the summary against Qualcomm conceded in favor of the original opinion.)
If cellular solutions ever get "good enough" then after Qualcomm's patents expire, people will simply make better implementations of older standards and won't be clamoring for new contracts for whatever Qualcomm is offering; they'll have to find a new market or ride into the sunset. (This may very well happen to x86-64 as related implementation patents expire next decade.) Indeed when the standard gets mature, a lot of customers did bring implementations in house as a cost savings measure.
However, the industry all maintained relations with Qualcomm for both implementations and IP with all the major cellular implementers signing licensing agreements because Qualcomm continued to reinvest its royalties back into R&D leading the charge into what the industry considered a desirable improvement of cellular technologies in 5G. It has largely been a virtuous cycle resulting in better products and lower costs for consumers much earlier than otherwise have been possible, and licensing is not an unfair take of product value as some VERY large customers of Qualcomm looking to reduce their input costs would have you believe.
As I recall Qualcomm got their leg up by having some IP related to CDMA. They leveraged that into the standards process for 3G and later LTE. That's how they "suddenly" became a player. They leveraged that toe-hold with some, apparently "hypercompetative" packaged licensing deals of their Arm SoC and broadband modem combos. Soon they were setting quite pretty in the smartphone market and could throw around their weight even more. FRAND might be fine for the modem parts, but it might just be a case of "Yeah, so... err... sorry. We don't quite licence the modems separately, we can give you a good deal on the combo."
Having Qualcomm involved isn't the problem. That future standards group just needs to tighten the contractual terms for FRAND to insure Qualcomm isn't allowed to play those games with the 6G standard.
That wouldn't change anything until all 5G and older is obsolete. Considering that Europe doesn't seem to be even trying to retire 3G, we'll all be dead by the time 5G stops mattering.
Yes, they are the most reversed of any federal appeals court, but that is still only an average of 2.5 decisions per thousand according to the fount of all knowledge (Wikipedia). I believe this includes cases that are not appealed to the supreme court, as well as those which are appealed. It probably also includes cases the Ninth Circuit refuses to review, since that decision can also be appealed to the Supremes.
They're most reversed due to being the biggest, mostly. Whether they have the highest reversal *rate* depends a lot on what terms you look at -- you could also make a case for the 3rd, 6th, or the 11th.
The 9th's reversal rate tends to get highlighted by conservatives in the US because it's perceived as the most liberal circuit -- it's their way of arguing that decisions they don't agree with are invalid.
In general, actually. Most judges get confused by things like rounded corners as denoting a copied design - in that example, it was to the extent that the Galaxy S3 (or was it the 4?) was designed in part by the legal team to ensure small things like the corners were not aligned.
Judges understand the law as it relates to the topic. This doesn't give them the ability to understand technology or it's impact in society.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
