Your pet peeve, technical or otherwise.

[IDEngineer]

The nice thing about standards is that there are so many to choose from. Including date formats.

[SiliconWizard]

If you strictly deal with "local" partners, using your local conventions is alright. Well, most often. But otherwise, that's pretty risky. That's what standards are for.
Yes, YYYY-MM-DD and time in 24h format are pretty much standard and should be used if you care about working internationally. Or if you just want to avoid any ambiguity.
Here, it ISO 8601: https://www.ionos.com/digitalguide/websites/web-development/iso-8601/

[DrG]

Enough with the holiday music! I am now two weeks beyond substituting pornographic lyrics as a coping mechanism.

no substitutions required if you grab a copy of this, possible Australia's  greatest export

Just listened to "Kristmas On the Piss" - this guy is a freaking disgrace - I love it!

[mansaxel]

The nice thing about standards is that there are so many to choose from. Including date formats.

There actually only is one, and that's ISO 8601. The rest are not standards, only conventions.

[Siwastaja]

MMDDYY is correct to me, but that's just the convention in my country, other places people have different conventions. You are essentially complaining about the fact that people speak a different language in other countries than your own.

It's not only that, it's because the convention really, actually (from culture-free neutral viewpoint, that is) logically sucks because it's neither little endian nor big endian.

It's like writing one hundred twenty three as 213 "by convention" and then you need to know it actually means 123.

But yeah, I understand your point. When you get mentally stuck into something, no matter how crappy system it is, it's hard to fix.

My pet peeve is the analog clock. Yes, that one which is used literally everywhere, in my culture, and in yours. I still hate it. It's total FUBAR which only works by forcing the convention through the throats of little kids when their minds are flexible enough to learn any arbitrary piece of crap. You might not remember it any more but as a kid, you were taught how to read time from the analog clock, and it was not simple 5-minute lesson, it took real effort to start to really understand it.

As an engineer who likes logical and good solutions and detest all kinds of bullshit, I still struggle reading time from the analog clock.

The problems are:
* Most significant (hours) hand is smaller than less significant (minutes) hand; or they are very similar in size, and form a strange graphical mess where you first need to identify which hand is which before you can start parsing.
* That stupid 12-hour system, it's like a scope that is set to trigger on both falling and rising edges of a sine wave so that the simple periodic signal is impossible to read. Why multiplex two completely different times of day into the same number? Who thought that?
* Despite people making a point that it's this way because of the history of showing the position of sun, this is exactly what it does not do. Astronomy based clock would be 24-hrs clock, and would show directly up at noon (which it does), and directly down at midnight (which it doesn't). It would be a rough approximate instrument without any arbitrary minute hands that mess up the clear physical indication.

Add extremely stupid, probably alcohol-induced ideas like the daylight savings mess.

Digital 24-hrs clock at least solves part of the problems and is orders of magnitude more readable (you can just read the time directly without tedious and unreliable parsing*) but still has the the problem of the absolutely stupid 24-60-60 division (why not 100-100-100 or 1000-1000 or whatever?)

*) which you might not notice because you are so used to it that it runs on "autopilot"

[PlainName]

Digital 24-hrs clock at least solves part of the problems

The analog clock scores over digital because you can easily see 'how close' and 'how far' relative time is. It's kind of like the metric system looks good on paper but doesn't allow for real people, so 15:45 doesn't mean much in relation to 16:00, but look at an analog watch and it means a lot more. Similarly, 16:00 doesn't really carry how far you are through the day.

[Kjelt]

Yes, YYYY-MM-DD and time in 24h format are pretty much standard and should be used if you care about working internationally. Or if you just want to avoid any ambiguity.

Ahhh just use a 64 bit integer in hex notation for all seconds passed since January 1st, 1970 at UTC and cut the civilian interpretation crap 

[Ranayna]

It's like writing one hundred twenty three as 213 "by convention" and then you need to know it actually means 123.

That reminds me of a stupid aspect of german. Those numbers make me crazy sometimes.

We say "(Ein)hundertdreiundzwanzig" literary "(one)hundredthreeandtwenty"

For my job i regularly have to switch between english and german, and i always misplace digits when writing numbers down

[Siwastaja]

The analog clock scores over digital because you can easily see 'how close' and 'how far' relative time is.

Yeah, in theory. It's unclear to me whether this is just because this argument is also taught, or if some people actually are able to eyeball the time from the analog clock effectively and quickly.

I'm not. For me, it takes a significant, and worse, nondeterministic mental load and anything between 500ms - 5 seconds to process what the heck that analog clock is saying. Digital clock is no problem. But this is because I'm good at reading and processing numbers.

Things could be different. Take an alternative reality, take a caveman who can't read. Build a 24-hrs analog clock with only one hand (hours hand), and you have a very well working scenario where even the caveman can eyeball the time and get intuitive idea; while reading the digital clock would be significant work.

But reality is, in addition to the ages-old 12h mistake, modern clockmakers have forgotten that originally there were just one hand, and additional minute hand was supposed to be very very thin, and the result is exactly the graphical mess where the parsing process starts by recognizing which hand to ignore, and only after that step is successful, you can go on and eyeball the time quickly. And even then, it doesn't match with the position of sun or the intuitive "percentage" of the current day, because there is the accidental and arbitrary 2x gain and resulting overlapping of the hours. So now you need to mentally combine this hours information to the approximate expected time of day.

Meanwhile, for those of use who read a lot and work with numbers, the mental load of directly seeing the current time in numbers is totally minuscule, like < 100ms. And totally predictable, there is no graphical cross-coupling of hours and minutes. And, because the hours are expressed first in all digital clocks (even those from the US of A, go figure!), it's trivial to just ignore minutes.

Anyway, Merry Christmas to all!

[Circlotron]

Analog clocks can give a good visual indication of the time remaining. Basic idea was used for traffic signals once.

https://en.wikipedia.org/wiki/Marshalite

[Siwastaja]

Yeah and I had an analog darkroom timer which had 100 seconds per full rotation, and only the second hand.

But such special instruments are not what I was talking about; my comments are regarding the usual timekeeping instruments as seen literally everywhere. Specifically those (today in majority) whose designers failed to understand the original idea of large hours hand and very thin minute hand.

[Ed.Kloonk]

When Mrs. Kloonk and I were trying for the first kid, we soon discovered there was no need to fix the broken thermostat on the kettle.

[Sal Ammoniac]

Yes, YYYY-MM-DD and time in 24h format are pretty much standard and should be used if you care about working internationally. Or if you just want to avoid any ambiguity.

Ahhh just use a 64 bit integer in hex notation for all seconds passed since January 1st, 1970 at UTC and cut the civilian interpretation crap 

Heck, with a 64 bit integer you could keep time in milliseconds since the Big Bang.

[IDEngineer]

Ahhh just use a 64 bit integer in hex notation for all seconds passed since January 1st, 1970 at UTC and cut the civilian interpretation crap 

Heck, with a 64 bit integer you could keep time in milliseconds since the Big Bang.

Make it a signed 64 bit integer and you can do both!

[mansaxel]

Digital 24-hrs clock at least solves part of the problems

The analog clock scores over digital because you can easily see 'how close' and 'how far' relative time is. It's kind of like the metric system looks good on paper but doesn't allow for real people, so 15:45 doesn't mean much in relation to 16:00, but look at an analog watch and it means a lot more. Similarly, 16:00 doesn't really carry how far you are through the day.

You are confusing things; "habitually familiar" and "neutrally correct" are not equal.  Your peeves are not my peeves; mine are opposite! To a person brought up with one system, the other system is always going to be strange; more so if the system is immersive; less so if there is cross-contamination.

I need to think twice for "AM/PM" formatting, but 24h time is a no-brainer. (By the way, Swedish and Danish and Norwegian has this name, "Dygn"/"Døgn" for a complete 24h cycle, that I've not found in other languages. Very convenient. A "dygn" is normally from midnight to midnight, but may also be used as shorthand for "n * 24 hours from <time>".) OTOH, I will readily accept that glancing at the hands of my wrist watch is a very quick and -- to me -- intuitive way to tell how far the current hour has progressed, which is often the most frequently needed information.

Same with measurement systems; I can immediately visualise "19mm" but ³/₄" requires parsing. 

Similarly, and perhaps a better case to describe the problem, the way AWG or SWG or Number Drills or BA threads are named makes me have to think hard about what a value is, instead of being able to deduce that a 1,5mm² cable is 0,6 times a 2,5 mm² cable. With AWG13 and AWG15 I have to know the intriciacies of a wire making method.

[Siwastaja]

Same with measurement systems; I can immediately visualise "19mm" but ³/₄" requires parsing. 

US and their binary system is actually quite clever as arguably binary (as the number system of minimum possible base) is superior to our stupid decimal system based on number of fingers!

It would just work better if it was only one one, and all other zeroes; i.e., 2", 1", 1/2", 1/4", 1/8" and so on. Even more, you could then just denote the place of the sole '1' using a single number, making it a logarithmic scale.

Also, representing the fundamentally binary system with decimal numbers loses most of the strength. This conversion is what makes it look weird.

For some weird reason I can't understand, measurements like 3/8", 3/4", or 15/16" or 7/16" seem to be hugely popular. These are all one-off from sensible values, and cause unnecessary clutter since you really need to look carefully at both numerator and denominator. I tend to mix up 3/8" and 3/4" quite easily, but I have nothing against 1/4", 1/2" and 1" sizes.

[PlainName]

For some weird reason I can't understand, measurements like 3/8", 3/4", or 15/16" or 7/16"

Too many numbers. Given we know how big one of something is, it's easy to realise how big 2 of them are, or half of one. And 3 or 1/4 is doable. But 27 of them? 15/34? Buggered if I know! That's how old systems scored over the metrics - they may be less accurate (for instance, being based on the size of a body part) but it's really easy to relate to a few of them. The trick is to start with a reasonably sized base thing for what you're measuring or describing so only single-digits (and preferably only half of those) are needed.

I use and prefer metric, especially when I am making (or breaking!) stuff. But the numbers are largely meaningless to me and just a maths thing. OTOH, tell me something is 2" long and I can relate to that. Whereas with 50mm (which I mentally convert to 2") I would take a ruler and just check off 5cm, with 2" I just look at my thumb and imagine two of the tops. (In fact, doing that now I note that the entire thumb is a shade over 2". What a coincidence!)

[Siwastaja]

Yes and even here everybody talks about 2x4 lumber and knows intuitively that if you cut that in half, it becomes 2x2, and then everybody knows a 4" or 5" nail is suitable to joining these pieces.

Despite the fact no one normally uses inch for measuring anything else, never has, it's totally weird unit of measurement for most of us.

We still think about lumber in inches even when you can't buy 2"x4" because no one manufactures or sells that, you'd need to import it from the USA! You can buy 96x48 mm but you can also buy 98x48 mm and whatnot for maximized confusion. Somehow 100x50mm wasn't an option, I have no idea why. Then you have to calculate what happens if you cut 96mm in half. 96/2 is not a difficult calculation, but if you want to concentrate on something else, 4/2 sure is easier.

And if you happen to have some old 2"x4" stock, good luck joining it to the significantly smaller "equivalent" metric size. (My house is built in 1952 using actual 2"x4", 2x5" and 2x8" because people back then understood that 1" = 25.4mm, not 24mm.)

Same for nominal pipe sizes. These have no relation to actual measurements, they are just names. Everybody knows what 1" pipe means (it's that roughly 33mm OD pipe), and the fact it does not measure even close to 1". So then lo and behold, we need new metric system. So let's call them DN25 or 25mm nominal pipe. It's still not 25mm, and for added confusion, we do have things like 19mm pipes that actually are 19mm outer diameter. Could have just kept calling 1" pipe a 1" pipe.

Same for SMD footprints, everybody knows what 0402 is and it's just a name, exact measurements are what they are and available in datasheet drawings. Now let's confuse everyone and invent a new naming system which then overlaps with the existing one.

Forced "metrification" makes things worse, especially when you try to apply it in some fully metric part of the world so there are no benefits to be had.

[mansaxel]

For some weird reason I can't understand, measurements like 3/8", 3/4", or 15/16" or 7/16" seem to be hugely popular. These are all one-off from sensible values, and cause unnecessary clutter since you really need to look carefully at both numerator and denominator. I tend to mix up 3/8" and 3/4" quite easily, but I have nothing against 1/4", 1/2" and 1" sizes.

Measurements using those inconvenient fractions happened because a ¹/₄" screw was too weak while there was no room for a ¹/₂" one. Thus ³/₈". Or, ¹/₄"  + ¹/₈"   

The stupid decimal system allows for arbitrary precision using a linear model. That is one of the very good things with it. As Dunkemhigh illustrated with their 2" vs 50mm example vs my 19mm example, the convenience using either is almost exclusively down to habit, excluding precise measurements, when the aggressive proponents of "Imperial" suddenly start using "thou" which, except being based on the inch (which in itself is defined as 25,4mm) very much is a decimal system...

Oh, and lumber-wise, the sawn dimensions are much closer to 2" and 4". The planed dimensions are smaller. As one would expect.

[Siwastaja]

There is no question about the binary system represented as fractional decimals being an utter disaster. I remember a funny Youtube video of two experienced American carpenters trying to figure out some basic geometrical math summing and subtracting fractional measurements and doing conversions. It's hilariously demanding task, even if you are used to it, and even if you are a professional.

But the binary idea behind it, always dividing/multiplying by two, is interesting. This would work perfectly fine if the representation was in binary form as well, and with the units (foot vs. inch and so on) being 2^n multiples (in which case it would be just tagging parts of the number, which you can opt not to do). Then you would have the same "linearity", but doubling and halving would be a simple bitshift.

The decimal system is by far not the only numerical system in use, historically. In this system, it's very easy to multiple and divide by 10, but that is an operation which is very rarely needed.

I wouldn't be surprised to hear that 2000 years in the future everyone's using binary and laugh at our base-10, like we laugh at base-60. Those who have programmed for a long time close to hardware already are pretty fluent in binary and see the benefits.

[james_s]

Digital 24-hrs clock at least solves part of the problems

The analog clock scores over digital because you can easily see 'how close' and 'how far' relative time is. It's kind of like the metric system looks good on paper but doesn't allow for real people, so 15:45 doesn't mean much in relation to 16:00, but look at an analog watch and it means a lot more. Similarly, 16:00 doesn't really carry how far you are through the day.

I don't really see much difference, digital, analog, it's all pretty much the same to me. Once you know that times ending in 15, 30, 45 and 00 are quarter hours it's pretty easy to visualize relative time.

[james_s]

Yes and even here everybody talks about 2x4 lumber and knows intuitively that if you cut that in half, it becomes 2x2, and then everybody knows a 4" or 5" nail is suitable to joining these pieces.

The thing that has always bothered me about lumber dimensions is that a 2x4 isn't actually 2"x4" when you buy it, it has been planed down to significantly smaller dimensions, although the length is what it says it is. It's only mostly tolerable because it's been that way for long enough that people just kind of know.

[IDEngineer]

Oh, you CAN get true 2x4's and 4x6's and the like, but they're "rough". The typical dimensional lumber everyone "knows" is "surfaced four sides", often abbreviated as S4S. Each side that is finished loses 0.25in, hence why they're actually 1.5in x 3.5in.

If you measure posts, like 6x6's and 8x8's, they're usually the full measurement. And they're often rough to the touch because they haven't been surfaced.

The nice thing about standards....

[PlainName]

Once you know that times ending in 15, 30, 45 and 00 are quarter hours it's pretty easy to visualize relative time.

I seem to manage just fine, but seeing the time on an analogue clock surprises me with it's immediate grasp of progress between the hour, or hour through the day. As I say, if I'm due out at 16:00 and it's 15:45 I know I have 15 mins, but I envision it as 3 x 5mins. Glance at a clock and I think, gosh, I am three-quarters of the way there (or just a quarter of an hour left).

Similarly, at 16:00 I know it's nearly tea time, and 4 hours since lunch, but seeing a clock it's obvious that it's late afternoon and not long to evening. Kind of hard to describe. Perhaps like using a mouse when normally you use the trackpad on a laptop.

[SilverSolder]

Once you know that times ending in 15, 30, 45 and 00 are quarter hours it's pretty easy to visualize relative time.

I seem to manage just fine, but seeing the time on an analogue clock surprises me with it's immediate grasp of progress between the hour, or hour through the day. As I say, if I'm due out at 16:00 and it's 15:45 I know I have 15 mins, but I envision it as 3 x 5mins. Glance at a clock and I think, gosh, I am three-quarters of the way there (or just a quarter of an hour left).

Similarly, at 16:00 I know it's nearly tea time, and 4 hours since lunch, but seeing a clock it's obvious that it's late afternoon and not long to evening. Kind of hard to describe. Perhaps like using a mouse when normally you use the trackpad on a laptop.

The face of the clock provides a context...    E.g. seeing an oil pressure of 100psi on a digital gauge doesn't tell you as much as seeing 100psi on a gauge that goes from 0psi to 200psi!