[Kobayashi.K]

Can be seemingly random when the cause-effect relationships aren't known. Once underlying variable is known as a cause, all of the sudden , what seemed to be unpredictable becomes predictable, not random anymore.

That is very true. I have studied this random thing as a sort of hobby because I wanted to know if you can create a true random process in software. But doing so I came to the conclusion that it is plausible that 'random' actually doesn't exist, and that we are only calling a process random because it is too complicated to find out its deterministic properties.

[reh321]

That is very true. I have studied this random thing as a sort of hobby because I wanted to know if you can create a true random process in software. But doing so I came to the conclusion that it is plausible that 'random' actually doesn't exist, and that we are only calling a process random because it is too complicated to find out its deterministic properties.

I took a course in “Modeling and Simulation”, where you begin by studying random number generators, as a Grad Student, then later taught a course in the material at the Undergrduate college level.

Actually, the good generators are fairly simple, but they still meet certain standards.

[Kobayashi.K]

I wonder if the firmware update is also for the random freezes of the camera as some users reported, hidden in
'Improved stability for general performance'.

[MaineNative]

Getting back to the original topic, I did actually get an error message writing to the SD card yesterday. I was just taking some test shots and hit the review button quickly. I got one message that the card could not be accessed followed by a message that a new folder could not be created. I was not writing to both cards simultaneously, I was set up for Raw + JPG writing to the cards sequentially. I had to power down to clear the error but the camera and card were fine. I didn’t lose any pictures.

I was using a pair of new 32 GB Sandisk UHS-II Extreme Pro cards. (I didn’t check them for errors before using them. Sometimes I do but I don’t have the scanning software on my new computer.) I only got the error that one time.

[reh321]

I wonder if the firmware update is also for the random freezes of the camera as some users reported, hidden in
'Improved stability for general performance'.

It is entirely possible that the engineers found paths though the code that either ‘go nowhere’ or go in circles as they investigated reports of strange behavior.

[biz-engineer]

I wonder if the firmware update is also for the random freezes of the camera as some users reported, hidden in'Improved stability for general performance'.

Problem of testing is as systems get more complex, testing efforts increase exponentially (or so). For some systems, it is practically impossible to achieve 100% faults coverage (there is this notion of NP hard problems, https://en.wikipedia.org/wiki/NP-hardness).

[jersey]

In physics an event is random or it is not random. There is no state between those two possibilities. So, as I explained a 'preventable random system glitch' is a contradictio in terminis. I think we are too much accustomed to 'random' processes (in software) that are artificially created and actually semi-random. It is not possible to create a true random process with software. True random signal generators are always using a physical process, for example thermal activity from ions, or particles that are emitted from a radioactive material.

Going back to the subject, a matter of more concern is the reliability of the new firmware itself, because in the past it has often happened that within one or two days Pentax published yet another update to compensate for new errors. So that is a valid reason to delay installing the new firmware.

In programing there are no truly random things. Everything is just a result of code being processed. Stuff may look random to customer, but that's it. So in terms of software there may be preventable random.

As for the update and fear of it I kinda understand it, in complex code each fix usually breaks two other things but in most cases those new errors are more and more "random" and require less and less possible variables. I know, I do it all the time for the money Me, being the programist I would rather update to fix common errors and risk less possible ones then not upgrade and stay with commonly occurring errors.

[reh321]

Problem of testing is as systems get more complex, testing efforts increase exponentially (or so). For some systems, it is practically impossible to achieve 100% faults coverage (there is this notion of NP hard problems.

Modern testing attempts to test every path through the code.
“NP” is irrelevant here because it refers to automata, and testing is done by humans.

[reh321]

In programing there are no truly random things. Everything is just a result of code being processed. Stuff may look random to customer, but that's it. So in terms of software there may be preventable random.

As for the update and fear of it I kinda understand it, in complex code each fix usually breaks two other things but in most cases those new errors are more and more "random" and require less and less possible variables. I know, I do it all the time for the money

“Recursive Regression testing” helps because it insures that those things which worked correctly before the upgrade still work correctly after the upgrade.

[Painter]

There's nothing like being controlled by fear.

Good grief I’m not afraid of updating my camera but if Pentax is churning out firmware updates every two weeks I think I can wait until they fix something of import.

In my job we do firmware and microcode updates on network gear all the time. No one expects things to go wrong but they still do.

[Kunzite]

“Recursive testing” helps because it insures that those things which worked correctly before the upgrade still work correctly after the upgrade.

Regression testing.

[reh321]

You're going too far - people usually don't mean that when they say random. They usually means "it doesn't always happen, and I didn't see an extremely obvious pattern".
OTOH, I'd say that a 'glitch' involving a cryptographic strength pseudo-random number generator would do the trick (can you differentiate from it and true randomness?). Been there, done that. The behavior was completely deterministic, the trigger was for all practical purposes random.

As I said in the comment which precipitated this comment, I have personally created errors which seemed ‘random’ to users, but made sense when I figured out how they got there. A common cause is a mistake in memory management - for example releasing memory in one module that is in use in another one. Modern languages {such as ‘Java’} make such mistakes unusual, but such errors do occur in older ones such as ‘c’; I don’t know what language the K-3iii firmware is developed in.

[biz-engineer]

Modern testing attempts to test every path through the code.“NP” is irrelevant here because it refers to automata, and testing is done by humans.

One testing approach for non-life-critical devices consists of only testing what's possibly used by customer. Software can have for faults in it, as long as those faults don't show-up in real life use.

[Kunzite]

As I said in the comment which precipitated this comment, I have personally created errors which seemed ‘random’ to users, but made sense when I figured out how they got there. A common cause is a mistake in memory management - for example releasing memory in one module that is in use in another one. Modern languages {such as ‘Java’} make such mistakes unusual, but such errors do occur in older ones such as ‘c’; I don’t know what language the K-3iii firmware is developed in.

Clearly not Java, as it's fast
They're using C or C++, or a combination of both (maybe some assembly as well).

[reh321]

Clearly not Java, as it's fast
They're using C or C++, or a combination of both (maybe some assembly as well).

All three have control of memory used {by the code} issues that could appear “randomly”, but an engineer can find by carefully examining the code.