# How awkward is it for you to read 24h time stamps?

edited May 2020 in Random

I grew up in a culture where we say "two o'clock" in the evening hours, but all digital clocks display "14:00". I only know the AM/PM dance from school textbooks and the Internet.

So ISO date-time stamps with 24h-formatted hours came natural to me: 2020-05-19 15:03, which is easy to shorten into 202005191503.

How awkward is it for you to read 24-hour time stamps?

I was thinking about this when Sascha mentioned something about encoding the IDs with a different base than 10, and noticed how neatly date-time IDs increment strictly monotonic. The first 4 digits are for the year. (The year 9999 would be 7pr in base36). The second 4 digits are month and day, and that combination always increments monotonic as well, but with gaps (after 0131 aka Jan 31st comes 0201 aka Feb 01st). The largest number possible in that chunk is 1231 ( y7 in base36). And the last 4 digits for the time increments, too, in 24-hour-format. (23:59, the largest value, is 1tj in base36) 12-hour-formatted timed would repeat and need a AM/PM flag that can make the number non-monotonic if you append the flag to the time (say a/p or 0/1 as a suffix, e.g. 12000 for AM, 12001 for PM, would not form a monotonic function values.

Edit: Ah well, this also applies to the whole number 🙄 So the monotonic property of the 24h format is given even then.
Edit 2: I came to this sudden realization due to/after ridicule by @sfast 😛

Post edited by ctietze on

Author at Zettelkasten.de • https://christiantietze.de/

• This is interesting. As you know, I'm personally using the note's creation date in its base32-encoded form (after Crockford). This may be also worth a look.

For manual testing, I'm using a Perl module:

#!/usr/bin/perl
use Encode::Base32::Crockford qw(base32_decode);
print base32_decode(<>);
__END__

#!/usr/bin/perl
use Encode::Base32::Crockford qw(base32_encode);
print base32_encode(<>);
__END__


There's also a Ruby library (base32-crockford), although that seems to produce slightly different output for numbers with >16 digits.

• @ctietze said:
How awkward is it for you to read 24-hour time stamps?

That's pretty much the only way I think about time. And I would be so bold to say that at almost every Swede below the age of fifty do the same.

Do you think that you could do a measurable optimization by using another base for date and time? The only other bases that I'm somewhat comfortable with is binary and hexadecimal, from my introductory course to computer engineering.

• I don't like it at all because I don't naturally just look at one and see the the time ID. I have to parse it out and it serves as a distraction. But I do see the utility in them. My perfect scenario would be having them used in file naming and then show up as a metadata option in the file explorer for the note.

But I've also seen people say they look at their notes in a temporal aspect and are able to find stuff by thinking "oh what I am looking for I created around this time period". Which I do not do at all, makes me think my memory might be broken... I am pretty sure I have "Severely Deficient Autobiographical Memory" and wonder if this is related.

• @inquisitiv3 said:
Do you think that you could do a measurable optimization by using another base for date and time? The only other bases that I'm somewhat comfortable with is binary and hexadecimal, from my introductory course to computer engineering.

To be honest, there's not that much to the actual encoding part -- the beauty of the 24h format making date-time-stamps increase monotonically came to me while I showered. I cannot do base-anything-but-10-encoding in my head, so I thought about making this more accessible by splitting the whole it into 3 parts: year, month+day, and time; and if that would help, and then I wondered how folks from 12-hour-time formats feel about this whole thing anyway

Author at Zettelkasten.de • https://christiantietze.de/

• Hmmm. Well, I'm English, so I would say that 14:00 is in the afternoon, not the evening! "Evening" comes from an Old English word meaning the time around dusk or sunset, so it usually means from around 18:00 onwards. As to the more important question, I would think that you would be hard pressed to find many people anywhere in Western Europe who are not fairly comfortable with the 24-hour clock, unless they are well into old age. I'm heading towards the age of seventy, and it has been in my personal usage for so long that I can't remember when I first saw it. Granted, almost everyone in the UK would (I believe) say "two o'clock" in conversation, rather than "fourteen hundred", but they might well write "14:00". And let's face it, writing "half past three" is a little lengthy when you can type "15:30". Though it might depend on the keyboard you are using. As for writing "a quarter past three" -- forget it!

• Of course, on the other side of the Atlantic most people looking at 15:30 would be scratching their heads and saying "What?"

:wq

• In the US people refer to it as “military time”. I know some off the top of my head. But I have to do the math to convert others. It’s clearly the superior way to communicate time. Just as the metric system is clearly the superior way to measure. I really wish we’d get our act together and join the 19th century. Though we’ve got some bigger cultural issues to deal with at the moment. So common sense measurement might have to wait until after we’ve righted the ship.

• @ctietze said:

@inquisitiv3 said:
Do you think that you could do a measurable optimization by using another base for date and time? The only other bases that I'm somewhat comfortable with is binary and hexadecimal, from my introductory course to computer engineering.

To be honest, there's not that much to the actual encoding part -- the beauty of the 24h format making date-time-stamps increase monotonically came to me while I showered. I cannot do base-anything-but-10-encoding in my head, so I thought about making this more accessible by splitting the whole it into 3 parts: year, month+day, and time; and if that would help, and then I wondered how folks from 12-hour-time formats feel about this whole thing anyway

Do you mean, you want to switch from a "202005241743" format to something like "2020-0524-1743" to make it more readable for users?

It is easier to decode, but its still something the user has to decode.
Why not allow to display the date in the users preferred format?

From the UX perspective, I don't think there is a way around a "live preview" (I don't know the exact term) of the base plain-text to something that hides the cumbersome timestamp-links and converts it to the preferred date-format (or to something different).
Isn't something like that needed anyway when you implement the display of pictures from within a note?

• I've seen in Japan some stores open hours go something like 8:00 ~ 26:00.

For timestamp I use a23 for the fist note I wrote at 23:00. Then e23 is the fifth note written at 23:00. My complete timestamp is 20200601a23. The limitation is 56 notes in one hour (a,b,c...,z,A,B,C,...,Z) and alphabetical ordering but I read the note's title.

• If you're asking how easy it is to read 24 hour time as opposed to 12 hour with an am / pm indicator, my answer would be: very easily. In day to day stuff, I use 24 hour and 12 hour interchangeably: 12 hour if I'm casually referring to a time in (for example) a Facebook post or conversational email, 24 hour if I'm jotting down an appointment. (Interestingly, at the age of nearly 50, I still don't know which out of 12am / 12pm is midnight and noon 😂😂)

If you're asking how easy is it to read 24 hour timestamps as we have them for zettel IDs, the 24 hour is no problem (as above) but it takes me time to inspect the ID and 'see' the year, month, date etc.

• @GBC

Interestingly, at the age of nearly 50, I still don't know which out of 12am / 12pm is midnight and noon

For every other hour on the clock (1 through 11) the first minute past the top of the hour uses the same designator (AM or PM) as the top of the hour itself. It would be strange if the 12 o'clock hour were any different.

12:01 AM is always one minute past midnight and 12:01 PM is always one minute past noon. 12:00 AM is therefore midnight and 12:00 PM is noon.

In 24-hour time 0000 and 2400 are both midnight, but what date is to be assigned to that time: the end of the day before or the beginning of a new day? The ISO 8601 standard answers this question by assigning the time 0000 to the following day and the time 2400 to the previous day, but both refer to the same instant in time.

:wq

• I'm an American. Despite having preferred and used 24-hour timestamps for years, it's a constant, effortful translation. I have my digital clocks set to 24 hours to try to train my mind, but it's still an uphill battle.