News for hackers. Stuff that matters.
Re:Porting USB applications to the web. Part 1: libusb by tinus_hn on Thursday January 20, 2022 @12:31PM (#30011930)
The next step is that you buy the device, but you don’t get a driver. You can conveniently use it from the manufacturers website.

I’ll leave it to others to speculate on what happens to the website in 2 years.

Re:Porting USB applications to the web. Part 1: libusb by Palomides on Thursday January 20, 2022 @12:38PM (#30012018)
plenty of USB devices need proprietary configuration software (like any fancy mouse/keyboard), razer has a whole pile of online crap already

if you want something else to complain about, this depends on chrome-only USB support

Re:Porting USB applications to the web. Part 1: libusb by matheusmoreira on Thursday January 20, 2022 @01:54PM (#30012996)
Proprietary manufacturer software is so bad. It's hard to describe how aggravating it is. At least Razer is popular enough so people have already reverse engineered the drivers and implemented open source versions. I had to do it myself on my laptop and it was certainly an interesting experience. I learned that the manufacturer's driver was intercepting every single keystroke just to send a "light up this key's LEDs" command to the keyboard every single time. Why couldn't they do it on the hardware? It boggles my mind to this day.
Re:Porting USB applications to the web. Part 1: libusb by lp0_on_fire on Thursday January 20, 2022 @12:47PM (#30012130)
> plenty of USB devices need proprietary configuration software (like any fancy mouse/keyboard), razer has a whole pile of online crap already

Which is the primary reason I will never purchase another Razer product.

Re:Porting USB applications to the web. Part 1: libusb by tinus_hn on Thursday January 20, 2022 @12:56PM (#30012248)
My example would be some 3D printer which you can only use from the website, and then after a few years you suddenly need a subscription to use your device. And a few years further support runs out and you can’t use it at all anymore.

Because you can’t force a company to keep a website up and running. But you can keep a driver on an offline computer.

Re:Porting USB applications to the web. Part 1: libusb by jandrese on Thursday January 20, 2022 @01:06PM (#30012395)
It won't be cached on the Internet Archive because the site is hidden behind a signup link and DRMed to your printer. The company will go after third party developers with legal threats like all closed hardware companies.
Re:Porting USB applications to the web. Part 1: libusb by RReverser on Thursday January 20, 2022 @01:10PM (#30012428)
Companies that want to do so, already embed "ring home" mechanisms in their apps that periodically connect to the Internet and check that you still hold a valid license or refuse to continue working.

It seems this depends a lot more on the company policy rather than specific technology used for the implementation.

Re:Porting USB applications to the web. Part 1: libusb by RReverser on Thursday January 20, 2022 @12:58PM (#30012293)
You can likely visit the old version of the website via Internet Archive or save an offline snapshot yourself if you want, so it would be no different to driver on an offline computer.
Re:Porting USB applications to the web. Part 1: libusb by folkhack on Thursday January 20, 2022 @01:15PM (#30012495)
Have to disagree here. Anything remotely complex with JS usually doesn't work as-expected after archived on Internet Archive. Keep in mind we're not talking about a LAMP-driven WordPress blog with some jQuery sprinkled in - we're talking about an actual piece of software running in the browser ie: device driver.

In good faith I would not expect something which uses experimental browser APIs to maintain functionality when archived by Internet Archive's automated crawling solutions. Same applies to offline saving.

Re:Porting USB applications to the web. Part 1: libusb by RReverser on Thursday January 20, 2022 @01:18PM (#30012529)
Good point regarding the offline capabilities, I'm not sure how well the Internet Archive will handle Wasm, but for the rest - WebUSB was already stabilized and isn't considered experimental (although I admit that didn't stop some other non-cross-browser APIs from being removed in the past).
Re:Porting USB applications to the web. Part 1: libusb by petee on Thursday January 20, 2022 @01:39PM (#30012795)
Archive is a little flaky on this front, but the other issue i see is when the company decides Archive is illegally hosting and files a DMCA or other lawsuit
Re:Porting USB applications to the web. Part 1: libusb by monocasa on Thursday January 20, 2022 @01:47PM (#30012900)
The difference is that webusb requires the specific website allowed to interact with the USB device to be listed in a USB descriptor. What that means is that you can't replace the functionality with the same UX. So for instance Arduino can make a generally available web arduino IDE, but I can't replace that if it goes down.
Re:Porting USB applications to the web. Part 1: libusb by devanl on Thursday January 20, 2022 @02:06PM (#30013147)
The initial WebUSB proposal included the notion of "allowed origins", but it was eventually removed in favor of user freedom.

The device can provide a landing page URL for the browser to show when the device is plugged in, but otherwise any site meeting the secure origin requirements and that has user permission can access a USB device, same as any other website.

Re:Porting USB applications to the web. Part 1: libusb by monocasa on Thursday January 20, 2022 @02:15PM (#30013304)
Oh, that's great to hear. I might take another look at it then.
Re:Porting USB applications to the web. Part 1: libusb by cluoma on Thursday January 20, 2022 @01:41PM (#30012833)
Given that the 'pile of online crap' has been around for a while now and seemingly not going anywhere, having access to this in the browser, and on any OS, sounds pretty convenient.
Re:Porting USB applications to the web. Part 1: libusb by dstaley on Thursday January 20, 2022 @12:45PM (#30012107)
This is also available in Microsoft Edge. Still, would be awesome to see support in Firefox and Safari!
Re:Porting USB applications to the web. Part 1: libusb by dmitriid on Thursday January 20, 2022 @05:36PM (#30016266)
> Still, would be awesome to see support in Firefox and Safari!

Both Safari and Firefox consider WebUSB harmful, and will not implement. Their reasons have been voiced loud and clear to the Chrome team (as they were voiced regarding many other standards).

Chrome, however, is the dominant browser and engine, so they dont' care. At all. It's now a "standard", and you will never other browser implementers' positions on web.dev (which is now a full-on unashamed Chrome propaganda vehicle).

Re:Porting USB applications to the web. Part 1: libusb by RReverser on Thursday January 20, 2022 @12:51PM (#30012190)
To be fair, Edge is also Chromium-based (I guess that's what the previous author meant, rather than specifically Chrome-only).

But yeah, I'm also hoping for WebUSB to make its way to other browsers, but for now I'll take any improvement to building & distributing apps across many operating systems at once :)

Re:Porting USB applications to the web. Part 1: libusb by tehbeard on Thursday January 20, 2022 @12:55PM (#30012224)
The OSS community has an easier time supporting it than when there's only a crappy windows binary blob?
Re:Porting USB applications to the web. Part 1: libusb by mschuster91 on Thursday January 20, 2022 @12:59PM (#30012302)
A crappy WebASM blob or heavily obfuscated JS blob is similar to a crappy windows binary blob, though.
Re:Porting USB applications to the web. Part 1: libusb by kitsunesoba on Thursday January 20, 2022 @01:47PM (#30012910)
Probably worse, because at least there’s well established tooling for reverse engineering Windows binary drivers, allowing for someone to eventually write FOSS drivers.
Re:Porting USB applications to the web. Part 1: libusb by RReverser on Thursday January 20, 2022 @01:56PM (#30013030)
I don't think it's worse because: 1) Wasm was designed to be easily decompilable / convertible to a text assembly representation by design, whereas for other platforms it requires fairly sophisticated disassemblers 2) Wasm can interact with the outside world (any I/O) only via explicitly defined import points, which are easy to intercept and/or replicate even if you don't know what's inside the binary itself.
Re:Porting USB applications to the web. Part 1: libusb by ThePadawan on Friday January 21, 2022 @03:03AM (#30020548)
I get what your opinion on the subject is but your example is hilarious to me since literally every time I have to download a device driver, it's on a manufacturer's website that looks (and performs) like it's been in continuous operation since 1986.