Its honestly a REALLY good idea. Still pisses me off that windows has had a QR code for years but it just goes to a generic support page.
That said: There are plenty of environments where a QR code is not viable. Secure environments where you cannot have a camera is one. But also most server rooms where the KVM has been abused for years and is covered in filth. What you can squint and scratch down on a piece of paper and what your phone can process are two very different things.
Linux so easy enough to have both code and text but I do have concerns on the broader impact of this being normalized.
IMO there are exceptionally few cases where it is acceptable for a QR code to not be immediately adjacent to a textual representation of the same content.
But take a look at computing and UX in general. There is a reason that a common refrain at the college and entry level job levels is “These kids don’t know what a folder is because they are used to everything at the top level”. And… there are very good reasons to not deal with folders in google drive or whatever. Hell, everyone lamented the loss of the start menu but how many of us still just do “winkey, ‘makemkv’” or whatever to launch something? Which is how you get thought processes that hide that until they are outright gone.
And the same with error messages. Hell, I was in a meeting not too long ago where we had a very serious discussion about whether we should even still emit error data to the console for an application when NOBODY ever thinks to copy and paste that. So what are we gaining when the first day of any support ticket is “Okay… can you get me this file from this folder? Okay, open up explorer and click this box and type c colon slash…”
I agree, it’s a great idea. The limitation is that the maximum amount of data a QR code can hold is about 3kB. Assuming the encoded data is compressed - either mostly ASCII and compressed, or a custom serialization format referencing external lookup tables - you could reasonably fit around 10kB of debugging information into a single QR code. That’s not too shabby.
Its honestly a REALLY good idea. Still pisses me off that windows has had a QR code for years but it just goes to a generic support page.
That said: There are plenty of environments where a QR code is not viable. Secure environments where you cannot have a camera is one. But also most server rooms where the KVM has been abused for years and is covered in filth. What you can squint and scratch down on a piece of paper and what your phone can process are two very different things.
Linux so easy enough to have both code and text but I do have concerns on the broader impact of this being normalized.
IMO there are exceptionally few cases where it is acceptable for a QR code to not be immediately adjacent to a textual representation of the same content.
and fewer to none of those allow you to submit kmesg and “debug information” to the kernel bug tracker
Agreed.
But take a look at computing and UX in general. There is a reason that a common refrain at the college and entry level job levels is “These kids don’t know what a folder is because they are used to everything at the top level”. And… there are very good reasons to not deal with folders in google drive or whatever. Hell, everyone lamented the loss of the start menu but how many of us still just do “winkey, ‘makemkv’” or whatever to launch something? Which is how you get thought processes that hide that until they are outright gone.
And the same with error messages. Hell, I was in a meeting not too long ago where we had a very serious discussion about whether we should even still emit error data to the console for an application when NOBODY ever thinks to copy and paste that. So what are we gaining when the first day of any support ticket is “Okay… can you get me this file from this folder? Okay, open up explorer and click this box and type c colon slash…”
Log it to a text file instead.
If I uderstand it correctly, this feature will be tied to build flags anyway, so server distros can have this turned off in their kernel.
I agree, it’s a great idea. The limitation is that the maximum amount of data a QR code can hold is about 3kB. Assuming the encoded data is compressed - either mostly ASCII and compressed, or a custom serialization format referencing external lookup tables - you could reasonably fit around 10kB of debugging information into a single QR code. That’s not too shabby.