Debate user @Napoleon Bonerfart on why the UART needs euthanized

  • 🇵🇦 Nuestro primer dominio localizado está en español en kiwifarms.pa. Our first localized domain is on Spanish on kiwifarms.pa.
  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account

Napoleon Bonerfart

In a Big Chungus dreams stay with you
True & Honest Fan
kiwifarms.net
Registrado
13 de Ago, 2018
We need to take the old-school serial port out back to behind the shed and put it down.

It had its time, but it's not useful anymore.

We have better things. Ethernet, USB, SPI (I2C can go fuck itself)

Shit, even a virtualized USB CDC terminal offers an easy way for someone at a coding bootcamp, learning to "Python into an Arduino," high bandwidth solutions.

What holds us back are the microcontroller developers. And they are nothing but a bunch of easily frightened manchildren afraid to get out of their comfort zones. USB is not that difficult to implement. A few resistors, maybe
a power switch chip, a High Speed 2.0 PHY if you're not a nutless amateur. The HAL isn't a problem either. At my last meeting with NXP, their vendors were not only blowing me, fingering my butthole, and busting my nuts in a balldo, they were also demonstrating out-of-the-box APIs for all of the spooky stuff.

The microcontrollers are also hundreds of megahertz now, why are we wasting that on a slow comm protocol?

And, uh, yes, Greg, you can use that extra MCU horsepower to drive a full curses terminal over CDC. PuTTY has mouse integration, why aren't you incorporating that into your text based UI? I'll tell you why: You're lazy.

tldr.png
 
You're a troll. You have no idea what you're talking about. You've never written or worked on an electronics device with a microcontroller. USB CDC? PuTTY? Are you indian?
 
Dude I've got mx cube open. Adding a CDC serial Port is like a handful of button clicks..
You're a monkey clicking buttons in an IDE with libraries that other, smarter people wrote. You have no idea what happens when you "just" add an USB stack, for the software or the hardware side of things.
USB is a couple orders of magnitude more complicated than UART. Why, just to send and receive a couple bytes at slow speeds? It's massive overkill unless you're actually transferring bulk data or are making a consumer product that justifies USB as convenience. UART will always have a place as a simple console for debugging, logging and consoles.
 
You don't need a UART for async serial. Just bitbang it like a normal person. If I can manage it on a 12F675 then any idiot can do it.
 
UART is much much easier to implement than USB.
I don't need all the USB features when I debug projects.
There is a good reason rs232 is still used. Compatibility and bare bones.
 
Cool.
I have plenty of idea: 500 to 700 KB/sec on a mid spec MCU.

And that's Bytes with a capital B.
As I said, if you need the transfer speed or convenience of hot-plug, you use USB. The tradeoff is the complexity and size of the USB stack library.
If you just want a basic debug console or even a minimal menu system, serial is just much simpler and easier to implement. You won't be able to read faster than maybe 1200 baud. (higher baud rates are good if you're sending control characters or occasionally sending data, likedata logging or uploading new MCU programs to the bootloader)
 
I've just been arranging the rocks in my back yard as an analog computer for as long as I can remember and it's served me well so far. I mean, it took me 6 and a half hours to type this post but if it ain't broke... I just don't see what all the hubbub about this thing I've never heard of and is therefore useless is.
 
Jeet hands typed this.

Let's just kill the one standard in the tech world that lets you cleanly connect to both 50+ year old and brand new hardware with the same tool. Just brilliant.

You're probably one of those autistic losers who will throw a fit about Intel and AMD removing native 16-bit support.
 
Ah so you are the type who would say X11 must be purged and replaced by Wayland.

Edit: yes. Gut and flatten unnecessary clutter from legacy Unix like operating systems.

X11 was written around the time of the Napoleonic wars, there's a lot to be improved in overall architecture.
 
Every tech manufacturer should develop their their own communications protocol (physical and virtual). Copying an existing protocol is basically just a cover song; there's no creative expression. The next generation can be employed hand making custom cables.
 
You're probably one of those autistic losers who will throw a fit about Intel and AMD removing native 16-bit support.
Who cares about x86, it's all bloated, you're never writing assembly for it, everything is going through a compiler, if your program isn't open source it's not good anyway, and if it is open source, it can be fixed up and compiled to whatever processor architecture you want (assuming it doesn't interact with the hardware directly but goes through the kernel or drivers).
If your processor has built in USB hardware and you use an IDE that gives you libraries to use it, fine. You don't know what's going on in those binaries or in the hardware. If something breaks, how are you debugging it? If you run into bugs? Do you know under which license the library is written and if it allows you to use it in a commercial product?
If you don't *need* USB for its hot plug, convenience, or transfer speeds, why build in unnecessary complexity?
 
Atrás
Top Abajo