I consider a measurement a quantification of a physical unit.
Measuring the passage of time is... physical?
Depends on the clock. But generally it doesn't sense anything, it just ticks at a predetermined interval.
There maybe special clocks that senses electrons or atoms and check time that way, but thats theoretical/rare.
Any clock, for that matter.
Yes, those rare ones fit the usual definition of sensor, but I did mean it in a broad sense.
In my specific case, a real-time clock unit, the DS1307, it klocks ze real-time, ja.
I have, like, three flags to give, ACTOR, SENSOR and DISPLAY, and yes, I have thought about whether or not a DISPLAY is an actor, and it, philosphically speaking, does concern the same matter, but regardless, it definitely warrants its own category by sheer amount of code alone.
>consider a measurement a quantification of a physical unit
Correction: A quantification of a dimension.
Of course a clock is a sensor and others are morons for saying otherwise.
A clock tells time; it potentially tells time synchronised to some standard (e.g. UTC), but at the very least it tells a change in time. Or to simplify further, it somehow counts the passage of time.
This means that a clock must have a counter that incremented based on the passing of time. To increment a counter something must be measured. So a clock MEASURES something that indicates an interval of time has passed, and it INCREMENTS each time one such measurement is made.
So yes, it does measure something. What it measures depends entirely on the type of the clock. For a simple mechanical clock it may measure the periodic motion of a driven pendulum under the effects of gravity. For a quartz oscillator you measure voltage changes - the quartz in the clock is piezoelectric and oscillates at ~2^15Hz, so when driven by an electric potential (battery) an output circuit sees these oscillations as high/low fluctuations in voltages. These are measured and counted (through a flip/flop array, i.e. a binary counter). And hence you can count the passage of time.
Then in atomic and optical clocks, i.e. frequency standards, you instead controllably excite atoms and measure emissions during controlled transitions between states. From this you get a known frequency standard, which can be counted through various approaches depending on the type of clock. But again, you measure something.
A clock is an instrument that measures the passage of time through incrementing a counter based on the action of a known driven behaviour in the clock system.
>you instead controllably excite atoms and measure emissions during controlled transitions between states
Mmmmh, sounds kinky....
Man, this just keeps getting better and better...
Can a clock not also be a relativistic velocity sensor?
https://en.wikipedia.org/wiki/Twin_paradox
Stupid frog is for sacrificing to Kukúlcan
2 weeks ago
Anonymous
NGL, "AI" is doing a really good job at vectoral graphics, such as anime and comics.
The slop is immediately noticable in pics that try to be live-action, though.
It's also "eerie," somehow, when AI does IRL.
2 weeks ago
Anonymous
>"eerie," somehow
Because you are noticing the jank subconsciously. AI screws up plenty with 2D, both types of art require further editing after being generated if you want to make them look more refined and less "eerie", or for the photorealistic stuff, it takes some trial and error and using the right model plus a little luck to get something very convincing.
I personally don't mind the jank much because,
1. I'm used to it
2. I don't know about you, but my dreams when I sleep are jankier than Stable Diffusion generations.
3. AI gens beat waiting a long time for non-AI artists to make what I like, and it only happens by chance at that unless I were to go and commission stuff. Though for bland and uninspired "1girl standing there staring at you with a neutral expression", there's an overabundance of that already without the need for AI software.
4. For something I really, really like, I use inpaint and Photoshop to de-jank it and refine it more.
Also, in all the negative kneejerk moron reactions to AI gens going on, we are getting all this bullshit about "real artists this and that", non-AI artists pump out plenty of equally bad and worse slop, especially beginners and low-skill folks still in the process of refining their skills and workflows. There is an overabundance of this kind of slop as well, could not be easier to find.
2 weeks ago
Anonymous
I mean, a lot of beginners out there joining in all the time.
2 weeks ago
Anonymous
I mean, a lot of beginners out there joining in all the time.
Crazy shit.
Yeah, I'm on the opposite end of the spectrum, hardware-wise, I mean, 2K bytes SRAM, I feel like 20 bytes to run a task is wasteful, and I'm not even done yet.
Meanwhile, GPUs come in flavors with 24GB.
It's also expensive as frick, and I'd probably have to ..."pause" my NEETdom to finance it.
Meanwhile, you can buy MCUs for literal pennies.
Or at least it used to be so not a decade ago, they seem to have gone up quite a bit recently.
Those numbers are 1xDHT22, 2xdata logging channels (each six channels with 12 data points for showing bar graphs in 1s,12s,1m,12m,1h,4h,12h) intervals for each temperature and humidity, 1xHD44780 20x4 character LCD on a TWI expander as devices.
A task (Process? Application?) base struct has 20 bytes overhead, which controls error handling, timestamping, fault monitoring, synchronization, etc. etc. and there are four of them running, one for rendering the LCD, one for interaction, one for RTC-drift and -synchronization and of course, one running every 1000ms to increment a U32 local real-time counter.
With single buffering of the render "image", it has 6% cpu usage (i.e. 94% of the second interval goes to idle task), with double buffering (i.e. only updating changed chars/pixels), SRAM usage grows by 80 bytes, of course, but CPU usage drops down to 0.3-2%.
Oh man, there's so much more to the display.
You can even set up the TWI interface to call an interrupt when it's done finishing transferring a byte to the LCD and basically render an image non-stop, achieving -- I have no fricking idea -- what kind of framerates, but maybe in the thousands of characters per second.
Meanwhile, the main loop runs on and updates the buffer image.
I wonder if I'll regret it later to have yet touched an STM32 and more "modern" periphery.
Regardless, I tell you, you cannot imagine the fun I'm having writing an operating system! 😀
No? It's just a counter with something that oscillates at regular intervals. It's a sensor, but a kind that doesn't measure any kind external stimulus, making it not really a sensor at all.
Atomic clocks are real, and are used in metrology labs, gps satellites, and datacenters. Not necessarily rare or impossible to have by any stretch of the imagination, so long as you don't need NIST or whatever levels of precision, but it's nonetheless a very specialist piece of equipment that most people don't really need, like most precision/industrial equipment tbh.
Of course a clock is a sensor and others are morons for saying otherwise.
A clock tells time; it potentially tells time synchronised to some standard (e.g. UTC), but at the very least it tells a change in time. Or to simplify further, it somehow counts the passage of time.
This means that a clock must have a counter that incremented based on the passing of time. To increment a counter something must be measured. So a clock MEASURES something that indicates an interval of time has passed, and it INCREMENTS each time one such measurement is made.
So yes, it does measure something. What it measures depends entirely on the type of the clock. For a simple mechanical clock it may measure the periodic motion of a driven pendulum under the effects of gravity. For a quartz oscillator you measure voltage changes - the quartz in the clock is piezoelectric and oscillates at ~2^15Hz, so when driven by an electric potential (battery) an output circuit sees these oscillations as high/low fluctuations in voltages. These are measured and counted (through a flip/flop array, i.e. a binary counter). And hence you can count the passage of time.
Then in atomic and optical clocks, i.e. frequency standards, you instead controllably excite atoms and measure emissions during controlled transitions between states. From this you get a known frequency standard, which can be counted through various approaches depending on the type of clock. But again, you measure something.
A clock is an instrument that measures the passage of time through incrementing a counter based on the action of a known driven behaviour in the clock system.
Yeah, but a sensor usually measures outside phenomena, like a touch, a light, or sound. A clock is neither of these things, since the clock circuitry/mechanism is not beholden to any kind of external stimulus. Maybe if you shook a grandfather style clock, it'll do something but you get my point.
>a sensor usually measures outside phenomena
are you saying that time is not an external variable? that time is the same universally? a clock on the space station is measuring something different than a clock on earth. and you don't necessarily need to have your equipment that reads the clock to be in the same physical location of the clock... if you sent a satellite to orbit, and the only things on the satellite were a clock, and a radio to communicate the time of the clock back to earth... is that not a sensor?
Of course a clock is a sensor and others are morons for saying otherwise.
A clock tells time; it potentially tells time synchronised to some standard (e.g. UTC), but at the very least it tells a change in time. Or to simplify further, it somehow counts the passage of time.
This means that a clock must have a counter that incremented based on the passing of time. To increment a counter something must be measured. So a clock MEASURES something that indicates an interval of time has passed, and it INCREMENTS each time one such measurement is made.
So yes, it does measure something. What it measures depends entirely on the type of the clock. For a simple mechanical clock it may measure the periodic motion of a driven pendulum under the effects of gravity. For a quartz oscillator you measure voltage changes - the quartz in the clock is piezoelectric and oscillates at ~2^15Hz, so when driven by an electric potential (battery) an output circuit sees these oscillations as high/low fluctuations in voltages. These are measured and counted (through a flip/flop array, i.e. a binary counter). And hence you can count the passage of time.
Then in atomic and optical clocks, i.e. frequency standards, you instead controllably excite atoms and measure emissions during controlled transitions between states. From this you get a known frequency standard, which can be counted through various approaches depending on the type of clock. But again, you measure something.
A clock is an instrument that measures the passage of time through incrementing a counter based on the action of a known driven behaviour in the clock system.
Anon, you have multiple sensors in your body :3
It truly is a wonderland! 😀
>kid turning the light switch on and off very fast
32768 times a second, the poor kid. >They made me build one in school
That's cool!
I'm the same anon from 100209925 last night, I woke up this afternoon to find the LCD display full of garbage, there must've been an overflow or something.
The timing circuit from that thread overflows every 65.536 seconds, so about every minute, and it definitely works in that regard, the uptime now is 83m.
I wrote the code that does the error checking for the sensors and everything, but I'm currently discarding them and I need to write the code to do error checking, which basically entails creating tasks, which basically entails creating an OS, and... yeah.
Here I am, apparently writing a general application OS to be run on an ATmega328p.
>Here I am, apparently writing a general application OS to be run on an ATmega328p.
The idea is that you can install it and then plug in LCDs and sensors and other peripherals and there's a device manager, a screen saver, programming PIDs and all that other shit.
Basically, it looks like I'm making a PLC out of the Arduino and currently writing the OS for it...
Sketch uses 7404 bytes (22%) [...] Maximum is 32256 bytes.
Global variables use 676 bytes (33%) [...] Maximum is 2048 bytes.
Those numbers are 1xDHT22, 2xdata logging channels (each six channels with 12 data points for showing bar graphs in 1s,12s,1m,12m,1h,4h,12h) intervals for each temperature and humidity, 1xHD44780 20x4 character LCD on a TWI expander as devices.
A task (Process? Application?) base struct has 20 bytes overhead, which controls error handling, timestamping, fault monitoring, synchronization, etc. etc. and there are four of them running, one for rendering the LCD, one for interaction, one for RTC-drift and -synchronization and of course, one running every 1000ms to increment a U32 local real-time counter.
With single buffering of the render "image", it has 6% cpu usage (i.e. 94% of the second interval goes to idle task), with double buffering (i.e. only updating changed chars/pixels), SRAM usage grows by 80 bytes, of course, but CPU usage drops down to 0.3-2%.
Those numbers are 1xDHT22, 2xdata logging channels (each six channels with 12 data points for showing bar graphs in 1s,12s,1m,12m,1h,4h,12h) intervals for each temperature and humidity, 1xHD44780 20x4 character LCD on a TWI expander as devices.
A task (Process? Application?) base struct has 20 bytes overhead, which controls error handling, timestamping, fault monitoring, synchronization, etc. etc. and there are four of them running, one for rendering the LCD, one for interaction, one for RTC-drift and -synchronization and of course, one running every 1000ms to increment a U32 local real-time counter.
With single buffering of the render "image", it has 6% cpu usage (i.e. 94% of the second interval goes to idle task), with double buffering (i.e. only updating changed chars/pixels), SRAM usage grows by 80 bytes, of course, but CPU usage drops down to 0.3-2%.
7436/756 bytes now, changed a few things.
The sensor in the clip is just symbolic, the actual connected sensor being measured is inside an, uhm... yeah, well, it's legal here, so frick it, my growbox.
The DHT manual stays true to its word, those things actually work with 20m cables.
There's zero external circuitry apart from the 100nF bypass at the sensor, the pullup being used is the ATmega's internal digital on.
>Astable multivibrator, in which the circuit is not stable in either state —it continually switches from one state to the other. It functions as a relaxation oscillator.
Whew! >In electronics a relaxation oscillator is a nonlinear electronic oscillator circuit that produces a nonsinusoidal repetitive output signal, such as a triangle wave or square wave
Hotdamn!
Yeah i shouldn't link wikipedia, lots of big words and categorization just to say it is the equivalent of a kid turning the light switch on and off very fast.
>kid turning the light switch on and off very fast
32768 times a second, the poor kid. >They made me build one in school
That's cool!
I'm the same anon from 100209925 last night, I woke up this afternoon to find the LCD display full of garbage, there must've been an overflow or something.
The timing circuit from that thread overflows every 65.536 seconds, so about every minute, and it definitely works in that regard, the uptime now is 83m.
I wrote the code that does the error checking for the sensors and everything, but I'm currently discarding them and I need to write the code to do error checking, which basically entails creating tasks, which basically entails creating an OS, and... yeah.
Here I am, apparently writing a general application OS to be run on an ATmega328p.
basically the clock tells a counter how fast it is has to count, the counter tells the memory what block must be outputted starting from 0 to infinite, every block is an instruction telling the rest of the architecture what to do.
And that's it, the faster the clock goes the faster everything else is processed, but fast means more energy consumption, more heat and shorter lifespan for your cpu, that's why manufacturers warn you about overclocking the cpu.
It is a unit of measure.
It is a violent act.
It is not a sensor.
I consider a measurement a quantification of a physical unit.
Measuring the passage of time is... physical?
Any clock, for that matter.
Yes, those rare ones fit the usual definition of sensor, but I did mean it in a broad sense.
In my specific case, a real-time clock unit, the DS1307, it klocks ze real-time, ja.
I have, like, three flags to give, ACTOR, SENSOR and DISPLAY, and yes, I have thought about whether or not a DISPLAY is an actor, and it, philosphically speaking, does concern the same matter, but regardless, it definitely warrants its own category by sheer amount of code alone.
>consider a measurement a quantification of a physical unit
Correction: A quantification of a dimension.
>you instead controllably excite atoms and measure emissions during controlled transitions between states
Mmmmh, sounds kinky....
Stupid frog is for sacrificing to Kukúlcan
NGL, "AI" is doing a really good job at vectoral graphics, such as anime and comics.
The slop is immediately noticable in pics that try to be live-action, though.
It's also "eerie," somehow, when AI does IRL.
>"eerie," somehow
Because you are noticing the jank subconsciously. AI screws up plenty with 2D, both types of art require further editing after being generated if you want to make them look more refined and less "eerie", or for the photorealistic stuff, it takes some trial and error and using the right model plus a little luck to get something very convincing.
I personally don't mind the jank much because,
1. I'm used to it
2. I don't know about you, but my dreams when I sleep are jankier than Stable Diffusion generations.
3. AI gens beat waiting a long time for non-AI artists to make what I like, and it only happens by chance at that unless I were to go and commission stuff. Though for bland and uninspired "1girl standing there staring at you with a neutral expression", there's an overabundance of that already without the need for AI software.
4. For something I really, really like, I use inpaint and Photoshop to de-jank it and refine it more.
Also, in all the negative kneejerk moron reactions to AI gens going on, we are getting all this bullshit about "real artists this and that", non-AI artists pump out plenty of equally bad and worse slop, especially beginners and low-skill folks still in the process of refining their skills and workflows. There is an overabundance of this kind of slop as well, could not be easier to find.
I mean, a lot of beginners out there joining in all the time.
Crazy shit.
Yeah, I'm on the opposite end of the spectrum, hardware-wise, I mean, 2K bytes SRAM, I feel like 20 bytes to run a task is wasteful, and I'm not even done yet.
Meanwhile, GPUs come in flavors with 24GB.
It's also expensive as frick, and I'd probably have to ..."pause" my NEETdom to finance it.
Meanwhile, you can buy MCUs for literal pennies.
Or at least it used to be so not a decade ago, they seem to have gone up quite a bit recently.
Oh man, there's so much more to the display.
You can even set up the TWI interface to call an interrupt when it's done finishing transferring a byte to the LCD and basically render an image non-stop, achieving -- I have no fricking idea -- what kind of framerates, but maybe in the thousands of characters per second.
Meanwhile, the main loop runs on and updates the buffer image.
I wonder if I'll regret it later to have yet touched an STM32 and more "modern" periphery.
Regardless, I tell you, you cannot imagine the fun I'm having writing an operating system! 😀
Depends on the clock. But generally it doesn't sense anything, it just ticks at a predetermined interval.
There maybe special clocks that senses electrons or atoms and check time that way, but thats theoretical/rare.
No? It's just a counter with something that oscillates at regular intervals. It's a sensor, but a kind that doesn't measure any kind external stimulus, making it not really a sensor at all.
Atomic clocks are real, and are used in metrology labs, gps satellites, and datacenters. Not necessarily rare or impossible to have by any stretch of the imagination, so long as you don't need NIST or whatever levels of precision, but it's nonetheless a very specialist piece of equipment that most people don't really need, like most precision/industrial equipment tbh.
Yeah, but a sensor usually measures outside phenomena, like a touch, a light, or sound. A clock is neither of these things, since the clock circuitry/mechanism is not beholden to any kind of external stimulus. Maybe if you shook a grandfather style clock, it'll do something but you get my point.
These are all very valid points of the flip side, devil's-advocate-anon.
>a sensor usually measures outside phenomena
are you saying that time is not an external variable? that time is the same universally? a clock on the space station is measuring something different than a clock on earth. and you don't necessarily need to have your equipment that reads the clock to be in the same physical location of the clock... if you sent a satellite to orbit, and the only things on the satellite were a clock, and a radio to communicate the time of the clock back to earth... is that not a sensor?
Man, this just keeps getting better and better...
Can a clock not also be a relativistic velocity sensor?
https://en.wikipedia.org/wiki/Twin_paradox
Of course a clock is a sensor and others are morons for saying otherwise.
A clock tells time; it potentially tells time synchronised to some standard (e.g. UTC), but at the very least it tells a change in time. Or to simplify further, it somehow counts the passage of time.
This means that a clock must have a counter that incremented based on the passing of time. To increment a counter something must be measured. So a clock MEASURES something that indicates an interval of time has passed, and it INCREMENTS each time one such measurement is made.
So yes, it does measure something. What it measures depends entirely on the type of the clock. For a simple mechanical clock it may measure the periodic motion of a driven pendulum under the effects of gravity. For a quartz oscillator you measure voltage changes - the quartz in the clock is piezoelectric and oscillates at ~2^15Hz, so when driven by an electric potential (battery) an output circuit sees these oscillations as high/low fluctuations in voltages. These are measured and counted (through a flip/flop array, i.e. a binary counter). And hence you can count the passage of time.
Then in atomic and optical clocks, i.e. frequency standards, you instead controllably excite atoms and measure emissions during controlled transitions between states. From this you get a known frequency standard, which can be counted through various approaches depending on the type of clock. But again, you measure something.
A clock is an instrument that measures the passage of time through incrementing a counter based on the action of a known driven behaviour in the clock system.
This anon understands clocks and posted what I was going to.
>Of course a clock is a sensor and others are morons for saying otherwise.
I can count seconds out loud. Am I a sensor?
Anon, you have multiple sensors in your body :3
It truly is a wonderland! 😀
>Here I am, apparently writing a general application OS to be run on an ATmega328p.
The idea is that you can install it and then plug in LCDs and sensors and other peripherals and there's a device manager, a screen saver, programming PIDs and all that other shit.
Basically, it looks like I'm making a PLC out of the Arduino and currently writing the OS for it...
Sketch uses 7404 bytes (22%) [...] Maximum is 32256 bytes.
Global variables use 676 bytes (33%) [...] Maximum is 2048 bytes.
Those numbers are 1xDHT22, 2xdata logging channels (each six channels with 12 data points for showing bar graphs in 1s,12s,1m,12m,1h,4h,12h) intervals for each temperature and humidity, 1xHD44780 20x4 character LCD on a TWI expander as devices.
A task (Process? Application?) base struct has 20 bytes overhead, which controls error handling, timestamping, fault monitoring, synchronization, etc. etc. and there are four of them running, one for rendering the LCD, one for interaction, one for RTC-drift and -synchronization and of course, one running every 1000ms to increment a U32 local real-time counter.
With single buffering of the render "image", it has 6% cpu usage (i.e. 94% of the second interval goes to idle task), with double buffering (i.e. only updating changed chars/pixels), SRAM usage grows by 80 bytes, of course, but CPU usage drops down to 0.3-2%.
7436/756 bytes now, changed a few things.
The sensor in the clip is just symbolic, the actual connected sensor being measured is inside an, uhm... yeah, well, it's legal here, so frick it, my growbox.
The DHT manual stays true to its word, those things actually work with 20m cables.
There's zero external circuitry apart from the 100nF bypass at the sensor, the pullup being used is the ATmega's internal digital on.
Coffee break and time to order some parts.
bro has a phd in yappology
Yes it measures vibrations of quartz and schizos pretend it's how time works.
It's an astable multivibrator: https://en.wikipedia.org/wiki/Multivibrator
Like a clock on the wall, it keeps ticking and won't stop until the very end.
>Astable multivibrator, in which the circuit is not stable in either state —it continually switches from one state to the other. It functions as a relaxation oscillator.
Whew!
>In electronics a relaxation oscillator is a nonlinear electronic oscillator circuit that produces a nonsinusoidal repetitive output signal, such as a triangle wave or square wave
Hotdamn!
You son of a b***h!
Yeah i shouldn't link wikipedia, lots of big words and categorization just to say it is the equivalent of a kid turning the light switch on and off very fast.
They made me build one in school.
>kid turning the light switch on and off very fast
32768 times a second, the poor kid.
>They made me build one in school
That's cool!
I'm the same anon from 100209925 last night, I woke up this afternoon to find the LCD display full of garbage, there must've been an overflow or something.
The timing circuit from that thread overflows every 65.536 seconds, so about every minute, and it definitely works in that regard, the uptime now is 83m.
I wrote the code that does the error checking for the sensors and everything, but I'm currently discarding them and I need to write the code to do error checking, which basically entails creating tasks, which basically entails creating an OS, and... yeah.
Here I am, apparently writing a general application OS to be run on an ATmega328p.
The only thing it measures is itself.
basically the clock tells a counter how fast it is has to count, the counter tells the memory what block must be outputted starting from 0 to infinite, every block is an instruction telling the rest of the architecture what to do.
And that's it, the faster the clock goes the faster everything else is processed, but fast means more energy consumption, more heat and shorter lifespan for your cpu, that's why manufacturers warn you about overclocking the cpu.
sneed
Which type? If it’s an RTC, it’s arguable, but if it’s the clock providing clocking to the entire circuit, I don’t think it can be called a sensor.
The circuit does not, and cannot function without a time base, which doesn’t line up with typical sensor behaviour.
My penis is a sensor.
And your mom is a quantity.
no, but anything reading it is a sensor.