Jest aside the USB-IF's stance (included towards the end) is consumer labels should be labeled on feature sets not standards revisions. E.g. the on the product it should say something like "USB 40 Gbps 60W" not "<cryptic standards name that supports 20 combinations here>". To that aim they are making a strong case for products adopting that style haha.
I am enough of a power user to want to know what potential future use-cases some very expensive device will support. But I guess it's hopeless in a world where most laptop manufacturers can't even implement USB-PD to spec.
I had a little bit of an hostile conversation with a 65W PD powerbank vendor over who violated the spec - the powerbank or my ThinkPad T495. Turns our the answer was simply "yes".
More like we (as consumers and technologists) still haven't learned our lesson.
A device supporting USB4 2.0 gen 2 doesn't mean that hard drive goes over 1 MB/s transfer rate.
A display supporting HDMI 2.1 doesn't mean it will output anything better than 1080i.
Specs are meant to describe interoperability. Conformance provides assurance that they're _is_ interoperability. Certifications (like Thunderbolt 4) provide assurance that they actually support certain features, and that they are supported in a way that matters.
But specs like USB 3.2 are completely forward compatible. That means they describe how a say a conformant USB 3.0 device _is_ a USB 3.2 device. What they add is a way that devices and cables that want to support 20 Gbps can do so in an interoperable manner. They don't mandate 20 Gbps support if you say you are a 3.2 compliant device.
When device makers and cable makers label them with spec internal names and subsections rather than the official marketing terms and logos (like USB 20 Gbps) it is an indicator that they haven't actually proved that they are conformant with the spec, or that they properly interoperate with others who are.
The market can solve this by preferring parts that aren't sold in consumer hostile ways. Problem is, those parts MAY work and usually ARE cheaper ;-)
"USB4 Version 2.0". Jesus Christ, why not USB4.1? Why are there TWO numbers? You'd think they'd want to keep numbers to a minimum since you already have to track wattage and bandwidth too.
Those guys really took the "Universal" in "Universal Serial Bus" to heart. Let's multiplex every conceivable connection device known to man into one giant standard, and throw in the ability to power the devices, too. That should really make things simple.
I seem to recall that there was some proverb about systems becoming so baroque that they eventually become rubble.
It seems that USB is sewing the seeds of its own destruction. No doubt that within 20 years they'll decide the whole thing is so complex that they need to replace it with something simpler.
And I'm still puzzled as to why plugging a keyboard into a keyboard port was considered such an inconvenience in the first place.
I think that USB is an object lesson in why you can't reduce inherent complexity, only shuffle it around.
> why plugging a keyboard into a keyboard port was considered such an inconvenience.
USB was a BIG deal, and a big improvement. It wasn’t the keyboards, it was the printers, scanners, external hard drives, and Zip drives.
Macs had something called “plug and play.” You plugged in a peripheral and it Just Worked. PCs didn’t have that. USB was intended to be the solution, and the most part, it succeeded.
BUT… the PS/2 connector was also a problem. In addition to there only being one keyboard connector on a typical computer—that happened to look just like the mouse connector, and wouldn’t work if you used the wrong one—it had an electrical flaw that could actually damage your motherboard if you plugged/unplugged it while the computer was on.
The original iMac was the first computer to have only USB ports, and no floppy drive. This was controversial at the time but it probably sped up adoption quite a bit.
And the connectors themselves have been great. Aside from PS/2 connectors, most other things use some variation of a DB9 connector. Mechanically, you could plug all kinds of things together that were not electrically compatible.
21 comments
[ 5.1 ms ] story [ 61.4 ms ] threadIs it just incompetence that leads to these naming choices? Who thought this was a good idea?
Jest aside the USB-IF's stance (included towards the end) is consumer labels should be labeled on feature sets not standards revisions. E.g. the on the product it should say something like "USB 40 Gbps 60W" not "<cryptic standards name that supports 20 combinations here>". To that aim they are making a strong case for products adopting that style haha.
I had a little bit of an hostile conversation with a 65W PD powerbank vendor over who violated the spec - the powerbank or my ThinkPad T495. Turns our the answer was simply "yes".
This is like the USB 3.1 Gen 1 vs 3.1 Gen 2 situation.
A device supporting USB4 2.0 gen 2 doesn't mean that hard drive goes over 1 MB/s transfer rate.
A display supporting HDMI 2.1 doesn't mean it will output anything better than 1080i.
Specs are meant to describe interoperability. Conformance provides assurance that they're _is_ interoperability. Certifications (like Thunderbolt 4) provide assurance that they actually support certain features, and that they are supported in a way that matters.
But specs like USB 3.2 are completely forward compatible. That means they describe how a say a conformant USB 3.0 device _is_ a USB 3.2 device. What they add is a way that devices and cables that want to support 20 Gbps can do so in an interoperable manner. They don't mandate 20 Gbps support if you say you are a 3.2 compliant device.
When device makers and cable makers label them with spec internal names and subsections rather than the official marketing terms and logos (like USB 20 Gbps) it is an indicator that they haven't actually proved that they are conformant with the spec, or that they properly interoperate with others who are.
The market can solve this by preferring parts that aren't sold in consumer hostile ways. Problem is, those parts MAY work and usually ARE cheaper ;-)
I'd love something like USC C 3.1.zxcvbn where it goes {form factor} {major}.{minor}.{capabilities}.
These committees tend to have (or at least give off the feeling of) a lot of power with little actual work being required.
I’m sure most here can think of colleagues from the past that would be the kind to fill this role, and that’s your answer.
I seem to recall that there was some proverb about systems becoming so baroque that they eventually become rubble.
It seems that USB is sewing the seeds of its own destruction. No doubt that within 20 years they'll decide the whole thing is so complex that they need to replace it with something simpler.
And I'm still puzzled as to why plugging a keyboard into a keyboard port was considered such an inconvenience in the first place.
I think that USB is an object lesson in why you can't reduce inherent complexity, only shuffle it around.
USB was a BIG deal, and a big improvement. It wasn’t the keyboards, it was the printers, scanners, external hard drives, and Zip drives.
Macs had something called “plug and play.” You plugged in a peripheral and it Just Worked. PCs didn’t have that. USB was intended to be the solution, and the most part, it succeeded.
BUT… the PS/2 connector was also a problem. In addition to there only being one keyboard connector on a typical computer—that happened to look just like the mouse connector, and wouldn’t work if you used the wrong one—it had an electrical flaw that could actually damage your motherboard if you plugged/unplugged it while the computer was on.
So, yeah, maybe don’t hate so easily.
And the connectors themselves have been great. Aside from PS/2 connectors, most other things use some variation of a DB9 connector. Mechanically, you could plug all kinds of things together that were not electrically compatible.