> “As we started to update our branding we did a lot of focus group studies with many different types of consumers,” he tells The Verge, “and none of those people understood the messaging and the branding, and they don’t understand revision control or spec names.”
You didn't need a focus group to tell you this (though I'm glad they did one). Just look at every single comment thread on HN about USB branding, at this point it's a meme.
USB-IF is a forum of so many different companies the focus group probably was needed for CYA finger pointing if no other reason. If this branding doesn't work out they can blame a bad focus group and convene a new one rather than war among themselves.
Given the terrible product naming[1] most hardware companies[2] have, at this point I’ve just come to the conclusion that most people in command don’t have any idea of how other people behave and reason.
[1]: Sony WF-1000XM3 are AirPods competitors, but you’d never tell your friends about your Sony WF-1000XM3
Sometimes I wonder if that's intentional. You won't suggest the WF-1000XM3, so you recommend "the Sony ones", and your friend goes to their preferred store where the XM3 is unavailable/out of stock and just orders some lower quality/older generation Sony ones rather than a competitor
But then that friend hates his earbuds and tells everyone how much Sony products suck. So in the end it hurts the brand. That's why having a small set of quality products with good names is best.
Sony has done a better job with other lines -- with Walkman, Discman, Vaio [Letter], PlayStation, Alpha, Aibo, and such. When they have a huge array of products, as with cameras or headphones, they soon become unnameable. They do have a set of headphone models called "LinkBuds" now, though. So they're not as clueless as you think.
The newer USB transfer speeds are basically using the same sort of line code as the Ethernet standards anyway. So doing something similar with USB would result in...
1.5BASE-UT (USB 1.1)
12BASE-UT (USB 2.0 FS)
480BASE-UT (USB 2.0 HS)
5000BASE-UT (USB 3.0 / 3.1 Gen 1 / 3.2 Gen 1)
10GBASE-UT (USB 3.1 Gen 2 / 3.2 Gen 2)
10GBASE-UT2 (USB 3.2 Gen 1 x2)
20GBASE-UT2 (USB 3.2 Gen 2 x2)
This is still ridiculous. The USB certified logo should be one image, then a similar sized wattage and Gbps number displayed next to it. The pertinent information is still too tiny. The cables themselves are still way too easy to mix up. Connector's inner plastic should have different colors
The color coded rings on resistors is a mightily bad idea, at least as implemented on resistors.
I'm not that colorblind, I see traffic lights, fruits between leaves and other things others have problems with, basically everything except resistor rings and distant lighthouse colors (yes, the colors on a map make sense, but trying to decode them IRL is as hard as resistor codes.)
The worst thing is resistors with weird background colors for the color bands. I can usually figure out beige resistors, but the blue ones? no chance of getting those right.
Yep. Just do it on the cable, heat shrink some wrapper around it and make it extremely obvious on the box. Try walking Grandma through what cable and charging brick to use over the phone when she calls you while she's at walmart
I recall coming across the color guide for the various features, like blue for USB3, and red for… power delivery, maybe?
But then one day I was on AliExpress or Alibaba or something and seeing that you could just buy cables from some manufacturer with whatever color you wanted. And I lost faith in the reliability of color id for cables.
Right, but it’s still easier to explain “either green or blue”, and the colors are in rainbow order for a reason.
It’s not just for grandma, most everyone would have an easier time remembering the “working” or “fast” colors than having to think about Gbps or watts, myself included.
But why are you defending it with novel solutions and colour grading schemes for something which could simply not be named the same thing because it isn't the same thing?
You can interpret the colors as naming. I mean, it’s similar with HDMI, DisplayPort, Bluetooth, Wifi, etc., they all regularly get new variants with new capabilities and compatibility restrictions. Naming each version something entirely different doesn’t solve the compatibility problem. It’s like saying HTTP/3 shouldn’t be called HTTP.
I have a bunch of USB-C cables. Some will fry things some won't. How will a naming convention help me keep track of them once they get chucked into in the basket under my coffee table
No in my view. USB should be USB. That rectangle thing you poke into a port somewhere. Not just on chargers, motherboards or literally aircraft and cars.
If it's not that. Call it something else. LEET, PHAT, BIGPORT.
I really don't understand. Universal Serial Bus... in 10 different formats make's it no longer universal.
Honestly I'd prefer that! Have fried two devices using the wrong damn cable/ charger. Sure the device should have had better power IC's etc to prevent anything from happening when I plugged in the wrong one, but still, they all look the damn same!
that's pretty much not true anymore. the wild west after USB-C first came out is over, testers are available enough and sketchy vendors get shamed loudly enough that unless you're buying cables at unreasonably dodgy stores, you're going to get something spec-compliant.
Snap-On still sells (along with many others) flashlights that charge via USB-C... only when using an A-C cable. Just as long as we're shaming non-compliant vendors.
> fried two devices using the wrong damn cable/ charger
Doubt the USB-IF can do anything about that. Those chargers either don't use the USB logo or use it illegally without passing compliance testing. Don't buy from those sellers.
no there should be only one cable that does work for all use cases.
do you really think most consumers now what the 40gbps/20gbps or 5w/240w or whatever even means for them?
No, consumers mostly don’t know this, but the market also needs lower tier cables. USB-C high throughput, high voltage cables with thunderbolt and display port (I think that is the max of what a cable can carry) are quite expensive.
Lots of people need reasonable cables to charge their phones.
They still have the problem of "Hi-Speed" being the slowest speed. They should've killed that name off too and required that manufacturers use the designation ½Gbps to be in line with the other speeds.
Actually "full speed" (USB 1.1, 12 Mbps) is the slowest speed that still somewhat common (in keyboards, etc). High speed (USB 2.0) is 480 Mbps.
I do agree though, as you and many other people have in the comments, just getting a numeric Mbps rating would be way better than the 3.1 x2 gen 4 nonsense.
The whole concept of "SuperSpeed" made it more difficult to understand what was going on, so I'm very glad to see it go. A rare hiccup in the USB consortium's otherwise unbroken streak of making each generation more difficult to understand.
The all-time prize has to go to the German wine producers, though, who regard a wine named "2001 Selbach-Oster Wehlener Sonnenheur Riesling Spaetlese Feinherb" as very precise and helpful.
It is a dry (feinherb) white (Riesling is a white wine berry variety) wine that has been harvested late in the season (Spätlese), making it have a higher alcohol level. Wehlener Sonnenheur is a geographical vineyard location (should probably read Wehlener Sonnenuhr, which is semi-famous and located at the Mosel, near Bernkastel), Selbach-Oster is the name of the winemaker, and 2001 is the year of production.
I fail to see what is complicated about that. But then I am German.
Most of those characteristics actually seem par for the course even in the US. For instance, Trader Joe's sells an in-house "Trader Joe's Reserve Merlot Sonoma Valley 2020" [0].
At a bare minimum, you'll get vintage (year), winemaker, and grape varietal, possibly with some additional qualifiers, e.g. Reserve as above, or late-harvest. Riesling in particular is such a widely used grape that can be either dry or sweet, so breaking out "dry Riesling" is not atypical.
Oops, thanks for the correction. I let my poor French bleed into my poor German. I have even stayed in Wehlen at the Gasthaus Prüm so it's especially embarrassing.
I think your point is relevant to USB, though! The USB folks know what all the words mean. They know what a USB 3.4 Gen 2 cable is, and what a PD cable is, and what a SuperSpeed cable is. Because they are the experts. It's not complicated to them, just as the wine description is, truly, not complicated to you. But your average person who just wants to connect their monitor to their laptop is left adrift.
P.S. that wine (well, the fruchtig version, not the feinherb) is one of my very favorites.
I think that's fine, every word carries actual information, though not speaking German will render the last two words useless and not being well versed in German wines will do the same to a few words at the start, so perhaps the only universally understood tokens are 2001 and Riesling (and perhaps not even Riesling?)
However, there are lots of laws called things like Rindfleischetikettierungsüberwachungsaufgabenübertragungsgesetz and I don't know which level of German language proficiency you need to be able to read that fluently, but probably one of the higher ones. Even native speakers sometimes struggle with lengthy concatenations and I feel that German legalese is a whole different level of crazy. So perhaps the crown for the most opaque naming scheme should go to the German Bundestag?
That compound word is obviously never used in conversation so I don't think proficiency factors into it. I doubt it's frequently spelled out, even in legalese. This is no different from how Americans use catchy acronyms for new laws so they don't have to spell them out in full (PATRIOT Act being a more blatant backronym example).
>That compound word is obviously never used in conversation so I don't think proficiency factors into it.
It is in discussions about how the duties on surveilling the labelling beef on packaging are distributed (which is what this law was about). The official abbreviation was RflEttÜAÜG ...
So if I've got two computers, each with a port labelled 40 Gbps and a USB cable labelled 40 Gbps, what's the protocol I can use to copy files at close to that? FWIW I used to run PLIP back in the nineties (IP over parallel port) to share my dialup Internet connection between my desktop and "laptop", so I'm not afraid of anything.
> the Ethernet over USB4 interdomain protocol, also known as USB4NET enables two USB4 PCs to establish a network connection between each other when connected using a USB4 cable, akin to connecting an ethernet cable between network cards on two PCs.
I wrote about this a month ago at https://news.ycombinator.com/item?id=32714807 it gets confusing because the USB4 protocol uses the word "routers" for hosts, devices, hubs which of course is used by Ethernet for something totally else.
It's rather unfortunate that you still have dig so deep to look up the auto-assigned IPs and then manually type them in Explorer.
It really ought to be as simple as file transfers from a PC to a phone - plug the cable, each side gets a prompt on whether it wants to provide access, and once you confirm, the other end shows it in the drive/device tree in Explorer.
Ah I missed it, TYVM! It's great to see that such a feature made it into USB4 now that it's going to reach these speeds.
> I do not quite know what happens if you were to plug three hosts together via a USB4 hub. As my post above details, USB4NET properly travels over the hub but which hosts pair, I can't even guess.
Can't wait to see people trying this and reporting!
This wouldn't be necessary if Big USB had better versioning numbers to begin with. They're right that customers shouldn't see "USB4 Gen whatever x LOL", but it shouldn't be a name underneath, either.
The problem is that there are too many "axes" to encode and simple version numbers would have never worked to encode that all: protocol version, port version, top-rated speed, top-rated power draw, optional features, etc. Not every device needs 80 Gbps and 200 Watts, and if every USB "4" cable had to support that minimum it would greatly increase costs per length to support a tiny fraction of devices. It would drastically simplify things when you go looking for a USB cable for a device, but the cost market of USB cables would look a lot more like, say, HDMI cables: just about only short cables and quite a bit of expense to them.
This new branding initiative may be on the right track, encoding the two axes most obvious to end users of cables: speed and charging strength. (If cable makers move to the new branding. They don't have to. That's the real confusion that USB should fix but can't. Branding is a suggestion, not a requirement.)
The problem is that techs push for branding to be things like a spec semantic version, but specs often just define options such that vendors would implement them in an interoperable way - while profiles are what define and test interoperability against mandatory feature sets.
Spec lines like USB 3.x and HDMI 2.x are meant to be interoperable sets of ever-increasing options, not an upward climb of mandatory minimum capabilities.
Vendors who didn't use SuperSpeed nomenclature before might have been doing so because it was clunky, but also might have been doing so because they didn't want to go through the effort of being certified against a profile (and in some cases, had nonconforming products)
This is simpler naming, but it remains to be seen whether implementors will suddenly care about certification. Those motherboards with the "USB 3.2 2x2 USB-A" red ports on the back are AFAIK un-certifiable and even non-conformant. No amount of marketing push for simpler names is going to help if vendors feel they get more value from just making stuff up.
Names are fine, the problem is if you call something superspeed what are you going to call next years speed? Hyper speed? Ultraspeed? Turbospeed? You rather quickly run out, and that’s if it doesn’t start sounding ridiculously hyperbolic before that. And of course superspeed is super slow by todays standards.
So you have to either use marketing names that don’t mean anything by themselves, or numbers that naturally increase.
Well, USB-IF made a mistake with USB High Speed, Super Speed (+), Full Speed.
Since 3.2 it was pretty simple:
- SuperSpeed USB 5Gbps
- SuperSpeed USB 10Gbps
- and so on.
However, to use, those vendors must send devices for testing and compliance. Many just don't do certification because they don't want, and some don't do that because they're making straight up non-compliant: SuperSpeed USB 20Gbps a.k.a. USB 3.2 Gen 2x2 can only come in USB-C form.
USB 3.2 Gen 1x2 which has the same speed as USB 3.2 Gen 2x1, but it only comes in USB-C unlike USB 3.2 Gen 2x1. When you see a motherboard with USB-A ports and those specs - those are not certified.
Then Super Speed branding tells you nothing about Power Delivery.
Or even USB 1.1, which called the faster of its two modes (12 Mbit, or perhaps more honestly Mbaud, per second) “Full Speed”, at which point anything you name an even-faster one is going to be confusing.
> it’ll now also have to list a charging speed like 60W or 240W.
Ah, no more trying to do determine what dielectric the cable is using: they’ll label the cable with how fast the charge can propagate through the cable. A significant blow for consumer choice!
It's even worse. A 20W USB-PD charger might indeed put out 20W, but not at the voltage your device needs. It might put out 20W at one voltage, but only 10W at another.
The whole thing is a goddamn mess, and it's a mess because the alliance is a cartel, not an actually useful technical standards group. They're a bunch of fat swine contributing little to nothing but extorting licensing fees from everyone.
This should never happen, provided everyone is actually following the specs.
If a charger outputs X W and a device needs Y W, for any X > Y it is guaranteed to output the voltage and current combination the device requires.
Every <=15W charger supplies 5V @ 3A, every 15-27W charger additionally supplies 9V up to 3A, every 27-45W charger additionally supplies 15V up to 3A, every 45W+ charger additionally supplies 20V up to 5A. A higher-power charger is therefore always a superset of a lower-power one. A charger may offer a variable voltage, but this variable voltage may not exceed the highest fixed voltage it offers. With the right cable, it may offer currents above 3A for lower voltages.
Devices are required to be able to charge from any charger providing at least the device's power rating, and it should even provide a similar user experience. It is not allowed to depend on the optional variable voltage or higher current a charger may offer.
Really, the USB PD specifications are fine. There is nothing wrong with them - except perhaps the fact that they are over 650 pages long. If you run into issues with USB-C charging, it is almost always caused by either the device or the charger manufacturer simply not following the damn specs.
I only remember a couple of gross violations, both in the early days: apple macbook charger and the Nintendo switch.
I think things have quieted down since. Of course that is due to the limited choice in chipsets (cartel? Or just bc nobody wants to design a complex, low margin part?). But nowadays dodgy mfrs can compete on who can electrocute or set alight the customer to save a couple of pennies, but the protocol conformance will be outsourced.
For the last 5 or 10 years, USB has seemed like a standard designed to make you purchase more than what you need. Whether it's the capability of the cable or the number of cables (through trying to rectify disappointing purchases). I'm sick of everything they are.
Those are irrelevant? Consumers only care about: speed and wattage. You're technically never supposed to care or even know about which USB standard it is, all of those Gen 3.2 2x2 aren't supposed to be consumer facing.
USB 3.2 Gen 2×2 “consumer” name is^W was "SuperSpeed USB 20 Gbps" that is what USB-IF been suggesting since the beginning. However, to use that name, they have to submit to testing and compliance. Otherwise, they can use whatever name they want.
As long as there are no two different standards that are capable of the same wattage and speed—everything is good.
I am still not convinced that Gen 1x2 actually exists. Sure, it's in the specs, but can you actually buy chips which support 1x2 but not 2x1?
Sure, in theory it is easier to design because you are dealing with two 5Gbps signals instead of one 10Gbps signal so you have a bigger error margin, but 2x1 was introduced in 2013 whereas 1x2 came along in 2017, had a worse real-world data transfer rate, and required way more complicated chips due to the mandatory USB-C handling.
Gen 2x1 uses a slightly more efficient error correction method than 1x2, so even though the number of bits going over the line is the same, the effective data transfer rate is higher.
Meanwhile, Gen 1x2 uses half the bandwidth that 2x1 requires, which makes the design of the chips, PCBs, and cables a lot easier. If a Gen 2x1 link fails to operate reliably, it will actually downgrade to 1x2 if possible.
The world was confusing when we had distinct ports that looked kind of the same but were not the same: serial ports, parallel ports and PS/2 connectors.
So they invented a universal serial bus. And now we've got 218312893 variants of it with different connectors, speeds, etc.
It just hit me that (of course) this was coordinated. I'm against many forms of advertising, but somehow I don't mind this. They needed a name for the letter and looked around. I can't imagine the brands having paid all that much to be the name of an Android version.
USB Yushityu 2007 Mimetic-Resolution-Cartridge-View-Motherboard-Easy-To-Install-Upgrade For Infernatron/InterLace TP Systems For Home, Office Or Mobile (sic)
This is a really stupid comment and you should probably delete it. What they've done here is pretty close to what I've been wanting. They're now advertising the port speed, which is all anyone cares about. They're not advertising meaningless version numbers anymore. The only improvement I could offer now is to advertise the speed specifically. What they are labeling as
40gb
20gb
10gb
5gb
is in actuality
More like
2x 20gb
2x 10gb
1x 10gb
2x 5gb
1x 5gb
Just specify exactly what the port/cable does. That's all you need to do. Version numbers mean nothing to me.
> They're not advertising meaningless version numbers anymore
USB IF is the only party responsible for making version numbers meaningless. They are _extremely_ useful for other hardware products - HDMI, DisplayPort, WiFi, etc.
There are scenarios where they would be useful. But USB IF decided to make them meaningless on purpose.
Everything about their history related to version numbers seems anti-consumer to me. Intentionally obfuscating capabilities and confusing consumers was a terrible idea.
Yes, it pretty much started when they decided not to "confuse" consumers the first time with USB 2.0. Claiming said fear they made all the USB 1.1 Devices USB 2 just with a different speed name. That way you could have USB 2.0 Mice and Keyboards.
So you had USB-Low Speed for keyboards and mice and USB Full Speed for Full Speed and Hi Speed for High Speed (later Super speed for speed that was just super).
Not confusing at all and it resulted in a time when you had to actually be darn careful not to get a USB 2.0 "full speed USB Hub" or other device. Later the same thing happened with regards to USB 3 a few years later.
Manufacturers making these (for a time until the relevant usb3 parts become too cheap to bother with these near fakes) always pop up.
At this point it will be a long painful process to replace all the hardware that doesn't follow the new scheme and USB version numbers could have been the much simpler solution.
One thing missing from their specs is the glut of "Power only" cables on the market. By failing to giving recommendation there Chinese manufacturers will all just do whatever - Probably abusing the High-Speed based labeling and that will pollute those and again cause confusion.
How hard is it to give them a symbol to print on their shitty power cables so people don't have to buy a cable tester?
You didn't need a focus group to tell you this (though I'm glad they did one). Just look at every single comment thread on HN about USB branding, at this point it's a meme.