When ever i see "never implement your own...", i know i want to implement it myself. People say that about hard things, and I only want to do hard things. Nobody wants people who can do easy things, people want people who can do hard things. The only way to learn how to do hard things, is to do hard things, so do the hardest things.
So go ahead, write your own date library, your own Unicode font rendering, compiler, OS, game engine or what ever else people tell you to never do because its hard.
Some things are good hard, the kind of hard that's driven by an interesting domain, going deep with well-architected tools or systems, learning lots of cool stuff.
I expect datetime-adjacent code is basically the opposite of all of this. All the hard parts are driven by fiddly adherence to real-world geography, politics, physics/astronomy, etc. There's no underlying consistency from which a sane model can be extracted, it's just special cases and arbitrary parameters all the way down.
I'm up for a challenge of course, but all else being equal, I'm happy to leave work that is the "bad hard" to others.
Reminds me of this passage from Postgres documentation:
”As an example, 2014-06-04 12:00 America/New_York represents noon local time in New York, which for this particular date was Eastern Daylight Time (UTC-4). So 2014-06-04 12:00 EDT specifies that same time instant. But 2014-06-04 12:00 EST specifies noon Eastern Standard Time (UTC-5), regardless of whether daylight savings was nominally in effect on that date.
…
To complicate matters, some jurisdictions have used the same timezone abbreviation to mean different UTC offsets at different times; for example, in Moscow MSK has meant UTC+3 in some years and UTC+4 in others.”
Parsing datetimes indeed sounds like a challenge in collecting, knowing and maintaining all these warped out standards and compromises. ”Bad hard” is a great description
Correct. It's not hard, just stupidly time consuming to the point of being unable to ever produce anything that works 70% of the time.
I hate anyone who will attempt to craft their own 10-lines line parser and then ignore that it fails 4 times a day. Just use the damn library. Thank you.
Write it for fun, but don't ship it. You're wasting everyone's time with your craft.
By all means, write it. Just don't use it. These warnings are almost always in the context of code you're going to release, not exercises in learning on your own.
Hard disagree here. Use it. Of course, if you running code that drive a pacemaker or a train maybe be careful, but in general, do things. We don't want a world where only three old bearded guys can write a compiler or a physic engine. Do the same errors again and you'll learn, eventually you'll do better than those who were here before you.
Well don't do it and instead of using an off the shelf library that is known to work while the rest of the development team isn't reinventing the wheel.
Doing it for fun and education is fine of course.
What IS the right way to model dates in a pacemaker ...? I hope the answer is "just don't do it" -- but I don't know what reasons there might be for a pacemaker to need to depend on calendar dates in order to best do its job ...
Well naturally it will need to connect to your phone via Bluetooth for the app to proxy update downloads and historic location data uploads. But in order to do anything on the network securely you need an accurate clock and the ability to parse datetimes because the PKI implementation depends on that.
Then the app pings you to remind you that your premium subscription will be expiring soon after which your heart rate will be limited to 100 bpm or less.
In the case of date libraries, I think if I ported the tests from a few well-known libraries to my own, I'd have reasonable confidence in my own.
Having said that, I don't think date libraries are hard, I think they're messy. Mostly because humans keep introducing convenience fudges - adding a second here, taking eleven days off there, that kind of thing.
You might be right, I haven't checked. It just seems on the face of it such an easy thing to test. Scalars go in, scalars come out. (This could just be me doing the Dunning-Kruger thing).
You could run a fuzzer against two libraries at the same time to find discrepancies....... hmm. That might actually be a good exercise.
Most well-known date library systems have failed in places. Quite a few, still do. So whilst you might get some known regression to test against, nothing can give you a foolproof guide.
You can have reasonable confidence that here there be dragons, but not so much that your assumptions about something will hold.
I'd say write it, probably don't use it, and don't share it unless it's substantially better than the alternative.
This way, you'll learn about it, but you'll more likely stay with something standard that everyone else is using, and you don't share yet another library that wastes others' time and your own (having to read about it, evaluate it, use it, and the migrate off of it when it's abandoned).
This is such nonsense. All the stuff that we use, someone wrote. If nobody makes them, then how is that going to work?
The messaging here is that you should be careful about using what you build on your own because it:
- hasn't been battle tested
- likely has bugs
- isn't mature
The only way that it will be all of those things is if someone invests time and energy in them.
From an ecosystem perspective this is absolutely the right thing. You want duplicate projects. You want choice. You want critical knowledge to be spread around.
> If nobody makes them, then how is that going to work?
I see it as “Dont write your own X, unless you want to maintain X. Here be dragons, this problem is deeper than it appears, the first 80% will be easy, the next 15% will annoy you, and the last 5% will consume your life for weeks, months, or even years. Or you could use a library”
The latter, if you want to get it completely right. I occasionally read the commits in the Qt framework, and from that, I can tell you that date-time stuff is complicated, and not in an instructive way, but in a super tedious way.
I think there is missing point in this discussion.
Most of the time you build something else.
Like if you build a todo app and have to deal with scheduling you don’t spend time making date library because it’s not your goal. But people would do that.
Heck most developers instead of starting blog on a blog platform start writing code for their own blogging engine.
This assumes, that the practices/methods used to create a working library are suitable for solving the problems. They might be ill-advised and include tons of workarounds for bad design decisions. Too often following the advice of never reinventing anything (and possibly doing better), is how we ended up with stacking broken stuff on top of other broken stuff, limiting us at every turn with leaking abstractions and bad designs.
It is very possible to have a clean implementation with good design choices overtake an established in time, enabling more extensibility, modularity and maintainability.
An example: People still way over-use regexes for all kinds of stuff. Even in code editors people for syntax recognition, where people really should know better.
In order to have these mature libraries, someone hat to start building them. They all had to to be incomplete, immature and horribly buggy early in their lifetime, too.
You assume that you always have a mature option available. That's (a) definitely not a totally generalizable assumption and (b) my point is that mature options only exist because the people that developed them just did it when confronted with the task.
We are specifically talking about something that does have a mature option available. That’s why it’s stupid to try and implement your own version of something complex.
If you change the story such that the product is actually needed and universally immature, of course building it is a valid argument.
Regarding b: Right, and the point of this article is that for those types of things, go for the already-mature thing. You’re arguing a point nobody is making.
Yeah, so do you want to go through that process of shipping broken crap and fixing user complaints one at a time until it happens to work for everyone, which is a mandatory process for all new libraries in one of these areas to go through, or would you rather stand on the shoulders of someone who's already done it?
I think that advice makes sense in the context of cryptography, where the consequences for getting it wrong can be quite serious indeed. I don't think it holds true for something as unimportant as a date parsing library.
Correct date handling (including parsing) can be monumentally important. Imagine an app that reminds people when to take their medications, for example
1) Dates are often stored as strings, so parsing them correctly is a necessary component of storing them. Also, those dates need not be simple app state. They could come from an API provided by your doctor/pharmacy
2) Many people (especially the elderly) take enough medications on different schedules that managing them all would be a significant cognitive load for anyone
It’s just an illustrative example, though. My point is getting dates right (including parsing their string representations) often matters quite a bit. If you disagree, let’s argue about that rather than quibble about the minutiae of the example
I get wanting to do hard things, but do you write in binary? Do you crank your own electricity?
My most valuable resource is time. Sure, I could learn more low-level aspects of my craft ... and sometimes I find it useful to do so.
When I focus on doing the hardest, already solved things by re-implementing them my own way, what value am I adding?
I've never met a client who cared about a library or how I did something in code - until it broke. Then, they didn't care who wrote it, they just cared it started working again.
People have built tables but I still build tables myself. Not as many people will use them as people who use IKEA tables, but that’s okay, I’m still going to build them.
I mean, a table is as hard as you make it. I work with rough construction lumber, and make nice finished goods, my point was that people still do stuff that isn’t worth their time financially.
The entire process is "the thing". In the case of a table by adjusting the inputs to the process you can cover quite a wide range of difficulty and required time.
For example, start from a felled tree, use only hand tools, and assemble using medieval joinery techniques. Building a table that way is quite hard by modern standards.
Now if you'll excuse me I need to get back to writing this date parsing library in assembly.
Not even doing it from scratch is the hard thing, the hard thing is getting the experience to know how to fit stuff together the best way to achieve your design and utilitarian goals, what wood to use, etc.
The hard thing isn’t building the date parsing library in assembly, it’s learning assembly well enough to do it in the first place.
I’m not sure where this discussion began but I was rebelling against everyone who says “just buy it” in regards to anything hard to do.
> The only way to learn how to do hard things, is to do hard things, so do the hardest things.
and i don't want to pay my employees to learn, i want to pay them to produce output i can sell.
Doing hard things are good, if this hard thing has never been done before - like going to the moon.
Doing hard things which has been done, but just not by you, is not good unless it's for "entertainment" and personal development purposes - which is fine and i encourage people to do it, on their own dime. Like climbing Mount Everest, or going to the south pole.
But if you are doing a project for someone else, you don't get to piggy back your personal wants and desires unrelated to the project on to it.
Except making employers do only easy things will make them stagnate. People who do nothing but simple CRUD apps over and over won't even be particularly good at making CRUD apps... whereas the guy who builds an Unicode font renderer in his free time always seems to write better code for some reason.
Getting better at your job is not just a "personal want" but very much something that the employer appreciates aswell.
Of course reinventing the wheel isn't good in corporate because the reinvented wheel is buggier than the ready made npm package but employers should go out of their way to find hard problems to solve that they can pass to their employees. It's called a growth opportunity.
Unless you work for enterprise consulting where employers appreciate replaceable cogs that they randomly drop into any project, and nicely out project budget regardless of delivery quality.
You can’t convince an employer with that attitude. They’re gonna keep exploiting their employees and “encourage” them to do their “personal development” in their free time.
> and i don't want to pay my employees to learn, i want to pay them to produce output i can sell.
This can be a bad local optimum. It probably depends on what exactly your business does, but it can make sense to pay an employee to acquire knowledge and skills that are needed in the business. You can't buy this off the shelf in all circumstances. Of course, it also has to make economic sense and be viable for the company. Unfortunately, I often see employees doing things quite badly that they don't really understand because they are not given the opportunity to learn properly. I can't imagine that this burns less money in the medium and long term than giving paid employees adequate space to learn.
I am in a work environment where I actually get to do hard shit for fun, learn a ton, and also "get stuff done" and my employer is happy.
For some of the stuff that has been done already, it might still make sense to do your own implementation, for example if you want to be able to experiment without having to navigate and learn a huge codebase and then have to maintain a fork just to have your own stuff in.
Another project we are starting now involves replacing software which is outright crappy and wastes our time. Thankfully my employer was able to see and understand this after talking it through with them.
Let's be a little charitable and assume they mean just learn. There are hard tasks you can learn from that also provide something you can't just get off the shelf, rather than just reimplementing the wheel.
> Your customers will pay more for things that are hard to do. Ask ASML.
What a silly example. ASML is valuable because it does something no one else does. It's not because it's hard, it's because they have the know-how and infrastructure to do it whereas others don't.
Juggling is hard. Do you know any millionaire jugglers?
No one else does it, because it is hard I thought? Hard to get all the steps and processes aligned to produce what they do. It is so hard, that there is no rich guy that wants to throw money in the hat and do it himself.
> People say that about hard things, and I only want to do hard things.
That's perfectly fine. Your time, your hobbies.
> Nobody wants people who can do easy things, people want people who can do hard things.
No, not really. People want people who do easy things, because they are clever enough to avoid needlessly wasting their time having to do hard things when they could have easily avoided it.
It's your blend of foolish mindset that brought us so many accidental complexity and overdue projects. There's a saying: working smart instead of working hard.
> So go ahead, write your own date library, your own Unicode font rendering, compiler, OS, game engine or what ever else people tell you to never do because its hard.
I can't believe this is such a controversial take. Solving hard things by yourself is growth. I 100% agree, rather solve a hard solved problem yourself than learning yet another JS framework or launching yet another revenue losing SaaS ("successful" because of VC). Or whatever. Push hard boundaries.
Nobody is really saying not to build these things. They’re saying the problem is exceedingly annoying to solve—and often not in a technically interesting way but in a way that is just massively tedious—and a better alternative almost certainly already exists.
If you want to build it to scratch an itch, go ahead. If you want to build it for fun, go ahead. If you want to build it because an existing solution gets something wrong and you can do better, go ahead (but know that it is a way bigger undertaking than you might assume at first glance).
The real advice is “don’t casually build your own X”, but that’s less punchy.
An exemplary one is "don't build your own timezone database"
It's not interesting, it's not fun, it's just a process of getting complaints it's wrong in edge cases and then fixing them, over and over until no one can find another broken edge case.
You can start by writing down England is +0, Germany is +2, etc... someone's going to mention DST and you'll put in a field for switching on the Nth Sunday of month X... later you'll run into a country that uses a different rule and you'll add a bunch of spaghetti code or write a Turing-complete DSL, etc... One day someone tells you about a village where they count 17 hour days on seashells and then you'll give up.
And if your DB doesn't produce identical results to the Olson DB in all cases then you created an incompatibility anyway. Might as well just use the Olson DB.
I think it's a spectrum and most fall somewhere on the line, hopefully dependent on the project.
My personal limit is rolling my own crypto, but I'm definitely more on the DIY scale because I agree. It's a fantastic way to grow and learn, and it's likely you might not have the energy to do it outside of work.
It's controversial because 1) good on someone for wanting to do something difficult and 2) I cannot think of a worse thing to try to implement. Maybe trying to parse the world's postal and street addresses is a close second?
> When ever i see "never implement your own...", i know i want to implement it myself.
Doing stuff for learning is useful, and the intent behind this general phrase is to not ‘implement your own’ something which is both hard and critical in a production environment. I work in cryptography (for security purposes) and have implemented quite a few things myself to learn, but I still use stable, field tested, and scrutinized crypto for any actual use.
> People say that about hard things, and I only want to do hard things. Nobody wants people who can do easy things, people want people who can do hard things.
Only wanting to do hard things limits yourself quite a bit: what about things which seem easy but could be improved? I worked in a non-tech related medical manufacturing job for a bit and took time to learn the process and tools. Afterward, I implemented a few tools (using what my coworkers (who have no programming or IT experience) have available to them: Excel and the VBA on the lab computers) to help them prep inventory lists which they have been doing by hand. Doing it by hand took them 3 hours as a group (and the first shift had to do this every morning), which my tool did in 5 seconds with a single button click. They still use it to this day, about a decade later.
This wasn’t something ‘hard:’ I glued a few files together, grouped a list by a filter, sorted the groups by a column, and made a printout which was easy to read and mark on as they went about their day. However, my coworkers didn’t even know this was possible until someone came in with a different skill set, learned what they did (by doing the job well for months) and then made a solution.
You must be careful with doing only ‘hard’ things. It requires other people to identify what is hard! In addition: crackpots do only hard things and believe they find better solutions than what exists so far (without consulting or learning about what has been done). Interesting people learn about things as they are (with the humility of knowing that they are not experts in most things) and tries to improve them using the knowledge they already have.
Don’t waste your time rolling your own crypto when you could do the _actual_ hard thing and identify unaddressed space to make careful and considered improvements.
It's all about the nuisance created by human behavior. Calendar, DST, timezone, all the problems you never imagined can happen and can only be met in real life scenarios, and you will meet same problem again, struggle then found out the same problem have been solved long time ago by mature library, and the solution doesn't require any smart or advanced technique, just another corner case.
Firstly because I have a great imagination, but secondly because I am old and have a lot of real life scenarios to think about.
State-of-the-art here has changed a few times in my professional career: Once upon a time most time/date libraries used a single integral type and try to make it do double-duty by being both interval and absolute (whatever that means) time by taking the interval from an epoch.
Relatively recently however, that's started to change, and that change has been made possible by people using languages with better type systems reinventing the date/time approach. This has led to fewer bugs, and more predictability with regards to calendar operations in different programs.
But bugs still happen, so this approach is still unsatisfying. One thing I keep having to worry about is distance; I record RTT as part of my events, since when I am looking for contemporaneous events, the speed-of-light actually tends to be a real factor for me.
So I don't think this is solved simply because my problems aren't solved by existing libraries, and I keep getting into arguments with people who think GMT=TAI or something dumb like that.
It's not "all about" anything: Nobody knows shit about what's happening in the next room over, and if there are 12 different date/time libraries now, I guarantee there'll be a 13th that solves problems in all of them, and is still incomplete.
I think in the case of the article the date library isn't necessarily hard but tedious. They mention most date libraries suffer from supporting too many standards or allow ambiguity.
I agree with you though, do the hard things even if it doesn't work 100% right you will have learned a lot. In university I had to implement all of the standard template library data structures and their features, it wasn't as robust as the actual STL but the knowledge of how those work under the covers still comes up in my day to day job.
I find this a perplexing comment in view of the fact that almost all of the linked article is in fact about how the author wrote his own date parsing library; the "never do it" bit is just a couple of lines at the start and so far as I can tell is mostly there for fun.
(In particular, at no point does the article actually argue for not writing your own date parsing library. It just says, in essence, "Never do it. I did it. Here's what I did and why.")
There things which was a result will make your mind click to an other way to comprehend a problem space and how to navigate through it.
And there are things which are hard due to pure accumulation of concurrent conventions, because of reasons like coordinating the whole humanity toward harmony with full happy peaceful agreement of everyone is tricky.
Handling date is rather the latter. If you dig in the lucky direction, you might also fall into cosmological consideration which is a rabbit hole of its own, but basically that's it: calendars are a mess.
Writing an OS. I've learned more about computers, hardware, CPU design, compilers, etc. that have translated into literally every other facet of my IT world than I could have done without this project.
> So go ahead, write your own date library, your own Unicode font rendering, compiler, OS, game engine or what ever else people tell you to never do because its hard.
You can absolutely do these things. What you need to be aware of is that in most cases maintaining these things to a production quality level is full-time job for a talented engineer. So you shouldn't attempt these IF:
- You have a higher-level aim you are also trying to achieve
- You need a production quality implementation
If one of those isn't the case then knock yourself out.
Missing context is - there is always something else you work on like the guy was making Eleventy so it was waste of his time.
If you work for a company and build todo app most likely it will not be beneficial for you to implement in-house library because there will be stuff that will bring much more value.
Like you don't have now 2 years to cover for all hard stuff because you have to make synchronization of tasks between devices and your boss most likely won't appreciate that.
"Never roll your own cryptography" is always used in context of building another application it is never "don't become a cryptography specialist".
In a scenario where a programmer has to do this for work and might naively think that date handling is simple, the title is invaluable advice. It is one of those things that can cause real trouble.
OTOH writing, e.g., your own renderer could cause some funny display at worst and maybe some unnecessary effort.
The only way you understand X is by making your own X and trying to support it for a few decades, and our industry needs more people who understand X; fewer who just ask chatgpt/stackoverflow/google for "the answer".
Software companies make money by providing value to their customers via the software they provide. How does reimplementing a hard but already well-solved problem align with their goals? How does that compare with solving a hard problem for which there are no good solutions yet?
Due to my work I rely on web scraped data for cybersecurity incidents. For Amazon Linux, they are disclosed with the fvcked up US datetime format (Pacific Time) and not in ISO8601 formatted strings which could imply Juliet/Local time.
In 2007 there was a new law that changed when Pacific Time enters/leaves Daylight Saving Time. Instead of making this fixed by a specific Day of a specific Month in numbered form like say "YYYY-03-01 to YYYY-10-01", they literally wrote the law quoting "first Sunday of April" to "last Sunday in October". Before 2007 it was "Second Sunday in March" to "first Sunday in November".
I'm not making this shit up, go ahead and read the law, come back and realize it's even more complex for other timezones, because some nations seem to make fun of this by going to +14:00 hours and -11:30 hours depending on the president's mood on Christmas or something.
In order to find out the Day of a specific calendar date, there's this cool article about Determination of the day of the week [1] which is quite insane on its own already. There is no failsafe algorithm to do that, each method of determining the day of the week has its own tradeoffs (and computational complexity that is implied).
Then you need to get all Sundays of a month, count the right one depending on the year, map back the date to ISO8601 and then you know whether or not this was daylight saving time they're talking about. Also make sure you use the correct local time to shift the time, because that changed too in the law (from 02:00LST to 03:00LDT and 02:00 LDT to 01:00LST before, to 02:00LST to 03:00LDT and 02:00LDT to 01:00LST after the changes).
Took me over 4 fvcking weeks to implement this in Go (due to lack of parsers), and I hate Amazon for this to this date.
PS: Write your own Datetime parser, this will help you realize how psychotic the human species is when it comes to "standards". After all this I'm in huge favor of the Moon Phase based International Fixed Calendar [2]
Reporting of cybersecurity incidents are easily late by a month or more, time zones are well below the rounding error. You will be more accurate to display it as YYYY±6month.
You seem to be not aware that there are a lot of legal obligations which come with providing this kind of inaccurate data. Especially if things go wrong because of it.
Every time someone mentioned "days" or "months" or "years" in Andor I had to mentally zap my brain not to think about how it doesn't make a sense across a galaxy.
Consider it a translation convention. There's a time and a place for "cycles" or "rels" or whatever, but it gets into "Calling a Rabbit a 'Smeerp'" [1] territory pretty quickly. The payoff isn't really all that great.
Stargate SG-1 is one of my favorite instances of this. The first couple of episodes address the fact that the Earth characters do not speak the same languages as everyone else in the galaxy. Then, having established the point that A: the show runners understand this is an issue and B: it makes for a rather tedious watch, they moved on to "everyone speaks English" and we all breathed a sigh of relief. I just think of it as part of the "camera" now. It turns out that we don't necessarily want a truly literal recording of what such things would look like.
"The Hunt for Red October" had an interesting way of handling this with the Russian speakers. The movie starts with them speaking Russian with English subtitles, does a slow zoom into the Russian-speaker's lips, and switches to English mid-sentence.
With some elegance, too; iirc they pivot languages on the word "Golgotha" as he reads from the bible, the Latin word for a location near Jerusalem, but having a non-English/non-Russian word be when they switch made it a lot less jarring. Plus, having it be during a read-out-loud-from-book portion allowed for more measured cadence that smoothed the switch but probably would have felt jarring if the audience were parsing multi-character dialogue when it happened.
I found that incredibly clunky when I saw it. Also, it's a little bit extra jarring that Sean Connery goes from speaking Russian to speaking English with a Scottish accent.
That trick has been used in movies before that too. "Judgment at Nuremberg" does something similar. A character is speaking German, slow zoom, then a switch to English.
I think that these fundamental things can be turned into an interesting topic, but you have to try for it.
Like, in a story background I'm pushing around, there's a coalition of a large amount of species developed on different planets. And you're a military officer, and you need to coordinate shifts, but - assuming some collectively normalized number of hours - some of your tiny dudes are tuned to 3 hours of sleep, 3 hours of leisure and 3 hours of work, others weird dudes with 2 arms and 2 legs are tuned to 38 hour cycles, and some huge dudes with a trunk in their face are tuned to 356 hour cycles.
Even if you could train and adjust this by an hour or two (which, for the 3 hour dudes would compare to an 8 earth-hour extension of duty for us), how the heck would you coordinate any kind of shifts across this? Or does every species have their own schedule? Good look finding crossover meetings then. Some of the small guys would have to do overtime for longer meetings even.
But you have to make it a point of the story and the challenges if you want to include it. If it is just a weird side note, just say that they figured out a conversion and that's it.
Some Star Trek books took the opportunity to work multiple species into the Enterprise's roster, when you don't have special effects problems with doing so.
But some others took the approach that Starfleet has a lot of vessels, and they're still somewhat segregated by species just because of those issues, and while the TV series don't corroborate that very well, I think it's better fanon overall. Peace and harmony among the species is great and all but trying to work 17 hour shifts in 2.5 Gs is going to get really old for the humans. And who wants to wear complicated breathing apparatuses for years at a time?
It would be an interesting direction to take a book series in... why do we see so much about the Klingons and Cardassians and Vulcans on TV? It's not because they're the only important species, it's because they're the species that breath our atmosphere at more-or-less our gravity and solar cycles. The Federation could be a whole bunch of parallel Federations-within-a-Federation where there's an entire set of species who also crew with each other but breath methane, need .7G, and work around 14-hour day/night cycles, and they just don't interact much with each other, not because they hate each other but just because it's so tedious to have prolonged contact.
If you’ve read David Weber’s Safehold series, this point gets super clear. It's written with names like "Zherald Ahdymsyn" (Gerald Adamson), but that makes it quite the slog for many.
I could not get through Banks’ “Feersum Endjinn” for this sole reason. English isn't my first language, though I’m fluent in it and read lots of hard sci-fi in it. But half a book using English “spelling rules” applied haphazardly just to make a minor point about one character was well beyond my capacity. I quoted “spelling rules” on purpose because let’s be honest, English doesn’t really have any.
You couldn’t translate that novel to Italian or Finnish, or any language with proper phonetical spelling.
Like in Game of Thrones when Davos was trying to learn to read and incorrectly pronounced the word "knight" the way it was spelled - somehow I could accept that everybody in a fictional universe spoke English except for all the ones who spoke other fictional languages, but I drew the line at words being spelled the same as well.
Coming from an educational background of imperial units, I sometimes catch flak from ... most of the world about this.
I take joy in exuberantly pushing back on their insistence of clinging to such archaic time units as "minutes", "hours", and "days", telling them to come back when they embrace kiloseconds. It is telling that most of my friends accept this with equal joy and laughter (:
It probably doesn't hurt that I've also spent time drilling metric conversions so that I can code-switch pretty seamlessly among units. Neurotic tendencies can have payoffs.
Unfortunately, the Second is measured for purposes of our timekeeping standards at sea-level on Earth which is ~1PPB slower than it would be in free space, as opposed to having a correction factor built into our time standards and so, for example, interplanetary ping times would be slightly shorter (in UTC/TIA nanoseconds) than expected.
I mean, than nucleus is much heavier and much smaller than electron, so it will be much less affected by external forces. We may see no difference between sea level and space based Thorium-229 clocks, or difference will be much smaller.
9,192,631,770 is clearly a sensible number and not something that's blatantly chosen to match some arbitrary pre-existing geocentric standard like 10,000,000,000 would have been.
It's retrofitted to what we already defined as a second, sure.
But you can tell an alien species our units are expressed in multiples of that, and they can translate it into how theirs works. (Vinge, for example, has space-faring humans talk about "megaseconds" and "gigaseconds" rather than days/years.)
> "megaseconds" and "gigaseconds" rather than days/years.
More like weeks and decades. Arranging to meet someone in a megasecond is like meeting them on the weekend; a megasecond is ~11.5 Earth days. A kilosecond is short enough to be used for moment–to–moment planning. They’re about a quarter of an hour each so they’re good for scheduling work tasks, scheduling time to meet people for a meal, etc etc.
Gigaseconds are more rarely used, since each one is ~32 Earth years.
Diaspora by Greg Egan has some fun with this too. The main character is a software emulation; called a citizen rather than a flesher. Most emulations live at a subjective rate 800× faster than the flesher normal. The second chapter is three flesher days after the first but 256 megatau, or ~8 years, have passed for the main characters. The fourth chapter is two thirds of a teratau later, over 20k subjective years. For the fleshers of Earth a mere 21 years have passed. The main character has actually forgotten the events of the third chapter; one of his friends brings it up and he has to search his archived memories to figure out what his friend is talking about.
Case-in-point, you are mistaken. The duration of a day changes due to many things, both logically and also physically due to the nature of Earth. Also just because you can call a second a second doesn't mean that is helpful making datetime software usable or easy on a different planet.
I think he's speaking historically. Obviously now a second is a fundamental SI unit defined in terms of physics experiments, but the origin of it was as the amount of time that was 1/3600th of an hour of which there are 24 in the day.
Similar to how almost-pi-squared meters-per-second shows up in the constant for gravitational acceleration near Earth's surface because the meter was originally "the length of pendulum that ticks once a second" and there's a pi in the pendulum motion equation.
(... it's not exactly pi-squared because the French yanked it around a bit before settling into the modern number based on light in a vacuum and cesium atoms).
The issue is that planetary locales will each have their own days and years (and possibly hours), so it would be confusing to adopt that same nomenclature for an interplanetary/interstellar time unit. And since the latter will be inconsistent with local time systems anyway, it’s easier to just have it use powers of ten. At least until we meet aliens that may prefer a different base.
But this makes no sense, humans can't just change their circadian rhythm to match an arbitrary daylight cycle, and clocks aren't necessarily reconfigurable. And with a good enough artificial lighting you don't need to depend on star. Daylight is just weather, it has nothing to do with how calendar works.
I was 100% thinking of use by humans living on other worlds. Pretty sure Mars will use seconds and hours. Handling dates will awkward whatever they decide on.
Currently, a Mars days is called "sol", FWIW.
If we find other species out there I won't speculate on how they think about time.
The Galactic Standard Calendar or Galactic Standard Time was the standard measurement of time in the galaxy. It was based on the Coruscant solar cycle. The Coruscant solar cycle was 368 days long with a day consisting of 24 standard hours.
60 standard minutes = 1 standard hour
24 standard hours = 1 standard day
5 standard days = 1 standard week
7 standard weeks = 1 standard month
10 standard months + 3 festival weeks + 3 holidays = 368 standard days = 1 standard year
That makes me think of the foreword Isaac Asimov wrote for Nightfall, explaining his choice of terms:
> The essence of this story doesn't lie in the quantity of bizarre terms we might have invented; it lies, rather, in the reaction of a group of people somewhat like ourselves, living on a world that is somewhat like ours in all but one highly significant detail, as they react to a challenging situation that is completely different from anything the people of Earth have ever had to deal with. Under the circumstances, it seemed to us better to tell you that someone put on his hiking boots before setting out on a seven-mile walk than to clutter the book with quonglishes, vorks, and gleebishes.
Well, in entirety of SW (or at least in mainline movies) it is kinda strange, that day and night happens basically on the same 24h period as on our Earth, given that all the planets are different. Could make a much more interesting story without this crutch for the audience.
It's funny to reason why we must go to bed when the clock has a certain number, since modern technology could easily be programmed to adjust as needed. No technical reasons the Martians can't go to bed at 9:00am today and 9:40am tomorrow. This mirrors my thoughts on why farmers caring about daylight savings time is farcical, farmers I know use the timekeeping of "crack of dawn" and "sunset".
iirc DST was never about farmers, and always about energy usage (lighting, etc) in the evening
Agree I've never met a farmer who cares about DST. Though also, for non-ag farmers, sometimes "crack of dawn" isn't early enough lol. Cow: "Dairy barn has electric lights, why aren't you awake at 4am tending to my needs, Human? Vacation? Lol no, stay here. Every morning. 4am."
DST is not about energy use. If it actually saves energy is debated and depends a lot on the local climate and air conditioning usage etc.
What it is about is moving an extra hour of daylight from say 5:30am-6:30am (when it is only of use to farmers and a few other early risers) to say 7pm-8pm when 95% of the population is still awake.
People want to understand when things are open/reasonable without having to do a lookup every time. A conversion has to happen somewhere - either I can say "X is in timezone foo. It's 1300 there so they're awake and I can call them" or "It's 1900 UTC, X is awake from....err....2200 to 1400, so I can call now".
The first is significantly easier as it requires remembering only a single offset and then going with societal conventions.
> "It's 1900 UTC, X is awake from....err....2200 to 1400, so I can call now"
Assuming it's that simple of course. Like, you can do mental math about what UTC "officially means" for someone many miles away, but people coordinate with others even if it means their local schedule is not aligned with that ideal case. Time zones account for this by being wider or narrower than their Platonic 15° of longitude ideal in some places.
And farmers are annoyed that one day the shop is open just after dawn, and the next not for another hour. Farmers are building their life around sunrise/sunset (or sometimes the dew cycles which is tied to the sun), and then fit the rest of the world in between.
We're mostly still a diurnal species. We go to bed at 9PM instead of 9Am for evolutionary reasons. We can fight against it, but the reasons are as arbitrary as biology is.
Likewise Daylight savings is a concept that had its uses, but makes less sense as technology progresses. I don't think even farmers care much about 7AM approximating to sunrise and 6PM as sunset.
Time zones are less about having uniform names for the times for waking up or going to work or mealtimes, and more about when your calendar changes from one date to the next.
I ran into date heck recently in a medical setting for storing birthdates. Eventually I settled on the idea that a birthdate isn’t a physical time, it’s just a string. We can force the user to enter it in the format 02/18/1993 leading zeroes and all, and operations on it other than string equality are invalid. We’ll see if this survives contact with the enemy but it’s already going better than storing and reasoning about it as a point or interval in time and people’s birthdays changing when they move timezones.
I like how Temporal[0] does this. What you were dealing with is Temporal.PlainDate[1], i.e. a date with a calendar associated but no time or timezone (might be due to being implied but also might be irrelevant, like in birthdates).
Temporal has other cool types, each with distinct semantics:
- Instant: a fixed point in time with no calendar or location. Think e.g. "the user logged in at X date and time" but valid across the world for any timezone or calendar system. This is what we usually use "Unix UTC timestamps" for.
- ZonedDateTime: like an Instant but associated with a particular calendar and location. Think an Instant but rendered "real" into a calendar system and timezone so the user can see a meaningful time for them.
- PlainDate: already discussed. Think e.g. birthdates.
- PlainTime: think "run task every day at 6:30pm".
- PlainDateTime: like an Instant but associated with a calendar system, but no timezone. Think e.g. what a user would insert in a datetime picker, where the timezone is implied instead of explicitly selected.
- PlainYearMonth: think e.g. "we'll run our reports during October 2025".
- PlainMonthDay: think e.g. "my birthday is June 13".
- Duration: think e.g. "the task ran for 3hrs 30min".
yeah, it was a long (and painful) time coming, but i think the temporal api finally basically nailed it. you know a library is good when you learn something about how to think about the problem just from how the code/api is structured.
Relatedly, std::chrono isnt exactly a beauty, but it did get people thinking about time points and durations and clocks and which operations are valid ways to move among them. Stuff like this is good.
> a date with a calendar associated but no time or timezone (might be due to being implied but also might be irrelevant, like in birthdates).
It might also be relevant: Ever ask an older Korean person their age?
> Instant: a fixed point in time with no calendar or location. Think e.g. "the user logged in at X date and time" but valid across the world for any timezone or calendar system. This is what we usually use "Unix UTC timestamps" for.
This is not a thing. Those are intervals to some Epoch, maybe taking into account leap-seconds and maybe not. They are not very useful except grossly over long ranges.
> - ZonedDateTime: like an Instant but associated with a particular calendar and location. Think an Instant but rendered "real" into a calendar system and timezone so the user can see a meaningful time for them.
Like when a user logged in at X date and time. They don't do this from no location, but from some location.
> - PlainDate: already discussed. Think e.g. birthdates.
And already wrong.
> - PlainTime: think "run task every day at 6:30pm".
Erm no. You can say 18:30 hours after midnight, or you can say when the calendar says 6:30pm, but these are different things. Imagine the poor fool who wants to run the task every day at "1:30am" and has it run twice on some days.
Bars close in some parts of the world at 30h (30時) to mean 6am the following day.
> - PlainDateTime: like an Instant but associated with a calendar system, but no timezone. Think e.g. what a user would insert in a datetime picker, where the timezone is implied instead of explicitly selected.
No, like a string.
> - PlainYearMonth: think e.g. "we'll run our reports during October 2025".
Nonsense. Also a string.
> - PlainMonthDay: think e.g. "my birthday is June 13".
Your birthday might be on the 29th of January. You cannot do reasonable arithmetic with
such things, so it might as well be a string like many of these others.
> I like how Temporal[0] does this.
I don't if you can't tell. This stuff is complicated and I'd like more people exploring it because I don't know The Right Answer™ either, but I know enough to know that every existing solution is wrong in some way that can cause real harm.
Just FYI you were downvoted with no explanation because you missed the point in all of these and you're using a smug and off-putting tone which makes it look like only care about "being right" and not finding what "is right".
Also you obviously didn't bother reading the "important concepts" link ([3]).
I was going to assume good faith and reply to each of your comments but it'd probably be a waste of time. As a summary: most of your concerns are wrong due to (1) confusing "timezones" with "location" or "internationalization" (2) confusing internal representations (like Epochs) with what these objects represent as described in the "important concepts" link and (3) just being completely wrong like saying you cannot do reasonable arithmetic with PlainMonthDay or even understand that not every relevant operation is arithmetic (good luck calling `.toPlainDate(2025)` with your string representation).
> FYI you were downvoted with no explanation because you missed the point in all of these
I think people downvote things they disagree with, and I'm not surprised people who think handling dates and times is beyond them are mad at someone exists who doesn't, because I'm calling out their impotence in a way.
And I understand you don't want to feel impotent, so telling me that you're not going to respond to me is a way for you to reclaim some of that.
> just being completely wrong like saying you cannot do reasonable arithmetic with PlainMonthDay
Then prove it: How exactly do you think you can meaningfully add two PlainMonthDays? What does such a thing mean?
I think you don't know what you are talking about, and you somehow think that means I don't either.
Is this DD/MM/YYYY or MM/DD/YYYY? I can tell from the 18 that it's the latter, but that convention isn't universal. I'd recommend YYYY/MM/DD as a less ambiguous format, but I don't have a perfect answer.
Furthermore, require dashes over slashes to signal that you are expecting ISO-8601 compatible dates, i.e., YYYY-MM-DD. Most users does not know the standard even exists, but it serves as an affordance that it is different from dd/mm/yyyy, etc.
Those ISO-8601 dates are great for date processing with primitive tools such as GNU sort or awk since they sort lexically, at least if you're not comparing dates in different time zones.
Yes, except for all the completely unhinged stuff in ISO 8601. You probably do not want to deal with durations, or repeating intervals, or even week dates and ordinal dates.
or a text field with some code that converts vernacular dates to a structured format. I don't think users are going to be too weirded out at seeing "1997-04-15" and will probably learn to use that natively.
The hard part is that a lot of devs aren't aware that there's a standard and that standard is superior to the alternatives.
Every time I see an input date string in XX/XX/YYYY format I get a micro PTSD flashback. This cannot be parsed reliably and is locale dependent. The standard date format is YYYY-MM-DD (it's also the date part of the ISO time format). Raw text inputs should be avoided as much as possible, date/time pickers should be preferred.
Even worse with just two digits for the year! 01/02/03 could be 1 Feb 2003, or 2 Jan 2003, or 3 Feb 2001. Let’s just be thankful no one ever uses any of remaining three permutations.
> The standard date format is YYYY-MM-DD (it's also the date part of the ISO time format)
Strictly, it is the extended form of the ISO 8601 calendar date format. (The basic format has no separators.)
ISO 8601 allows any of its date formats (calendar date, week date, or ordinal date) to be combined with a time representation for a combined date/time representation, it is inaccurate both to call any of the date formats part of the time format, and to call the calendar date format the format that is part of the combined date/time format.
(There's a reason why people who want to refer to a simple and consistent standard tend to choose RFC-3339 over ISO 8601.)
This isn't really relevant to the parent's topic though, aside from UX. The UI can tell the user which is the day and which is the month. The logic layer knows the format explicitly.
I guess in your case you're never doing date arithmetic or greater than or less than, but only doing equality testing, right? That is, it's part of a composite key.
I faced a similar problem with a form where people were supposed to submit a date and probably not aware of what timezone was involved. I figured that so long as they selected "02/28/1993" and people always saw "02/28/1993" that was correct and if they ever saw it differently it was wrong. So I used non-TZ aware dates throughout the whole system.
If you store it just as a string it means that you cannot do anything useful with it like age-dependent logic or you just pass on parsing logic to users of the field.
FHIR in my opinion has a pretty good system for dates (including birthdates):
YYYY, YYYY-MM, or YYYY-MM-DD. (Not knowing your exact birthday is common for some countries).
What environment are you in where you have to work with birthdates, you have timezone aware dates, times, and intervals, but you don't have a naive/plain/local date type that already exists forcing you to use strings in place of date-without-timezone?
You seem to have a reasonably expedient solution for that problem, but it is surprising to have the combination of things you have to have and things you have to be missing to have that problem in the first place.
They may have something that's just not as easy to work with as strings. E.g. in Swift, you have DateComponents, but that's too dynamic (and Date is sometimes referred to as naive, but that's a misunderstanding, since they are timestamps, not date+time).
> people’s birthdays changing when they move timezones
That's because the developers use datetimes (aka timestamps) to store a single date. Just pick an arbitrary epoch date (such as January 1, 1900 as used by Excel, or my favorite January 1, 1600 since 1600 is a multiple of 400 making leap year calculations even simpler) and store the number of days elapsed since then. The rules involving leap years are much much simpler than rules involving timezones and timezone databases. The translation from/to this representation to a broken-down y/m/d takes only ~50 lines of code anyways.
Of course if you don't need to do arithmetic on dates, just store three numbers, year, month, and day.
No, don't do that. Use a date datatype (not date/time). You aren't the first person to ever need to handle dates without times/timezones in a computer program. Use what your database/language/libraries already have to support that.
Specifically, a "local date", codified as LocalDate in every date library worth a damn, except for Javascript which chose "PlainDate" just to be different.
which are the moral equivalent of Unix timestamps with a different offset and multiplier. These work OK for human history but will break if you go far enough into the past or the future because uncertainty in the earth's rotation adds up over time.
If you don't care about timezones timezones may still care about you, if you want to minimize trouble it makes sense to properly use timezone-aware Zulu (GMT) dates for everything if you can.
In certain cases you might be doing data analysis or building an operational database for throttling access to an API or something and you know there are 16-bits worth of days, hours, 5-minute periods or something it can make sense to work relative to your own epoch.
In my humble opinion, this is not good advice unless you demonstrably need it for query performance or something. It is very easy for the logic layer to accidentally mess that up, either in reading or, worse, in writing back.
In this case, I'd suggest storing what you mean (the user wasn't born 9,487 days after Jan 1 1970. They were born Dec 23, 1995.)
Storing the literal units (and ONLY the relevant units), as the parent has, is robust and logically+semantically correct (they could add a translation layer for UX so the user doesn't have to be particular, but that's beside the point). Whether you use a string or a struct or some date-only type is moot, as long as you're literally storing the year, month, and day, and only those three things. You can ephemerally convert it to your platform's date type if you need to.
That's exactly why I propose a multiple of 400, not a multiple of 100. The proleptic Gregorian cycle is a 400-year cycle. There are 97 leap years in it. What's tricky about it? Just take a look at my code: https://github.com/kccqzy/smartcal/blob/9cfddf7e85c2c65aa6de...
Why not just make a map of dates to days since 0001-01-01 in a plain text, then compress it at build time? We are not constrained by memory anymore.
Just use simple database as source of truth with all days passed since a start of human history (e.g. 6000 years ago) with labels such as "this day 12345678 was known as day XXXX-xx-xx in those regions, also known as YYYY-yy-yy in those regions, also known as ZZZZZ in this specific region". It's not a hard task to automatically compress such database into a compact representation.
"Years are not leap years, unless % 4, unless % 100, unless % 400."
It's a wacky rule for sure.
2000 was fun. Everyone knows about "unless % 4", but there was also an interesting and very vocal set of people who knew about the "unless % 100" but somehow knew that without knowing about the "unless % 400" part. A very specific level of knowledge.
IIUC why medical cares at all, this is really insightful. Because as far as I'm aware, the medical industry basically uses birthdate as a key; it helps to (a) tell two patients with other primary keys (like name or address) apart and (b) do a quick mental-acuity check on the patient by just having them regurgitate the value and doing a human-brain string= on it.
IMHO, ISO 8601 as a standard is way too broad and unspecific. ISO 8601 is way too messy. Telling somebody that they need to parse an ISO 8601 date time is not enough information to do the job. Which variant is it? Does it include the time part. IMHO allowing the full range of ISO 8601 dates and times in a data format is usually a mistake. You want to be more specific.
There's a need for a standard that locks down the commonly used variants of it and gets rid of all the ambiguity.
For me, timestamps following this pattern 'YYYY-MM-DDThh:mm:ss.xxxxxZ' is all that I use and all my APIs will accept/produce. It's nice that other legacy systems are available that don't normalize their timestamps to UTC for whatever reason, that consider seconds and fractions of a second optional, etc. But for unambiguous timestamps, all I want is this. It's fairly easy to write a parser for it. A simple regular expression will do the job. Of course add unit tests. Technically the Z is redundant information if we can all agree to normalize to UTC. Which IMHO we should. In the same way the T part and separators are redundant too. But they are nice for human readability.
You can use datetime libraries that are widely available to localize timestamps as needed in whatever way is required locally. But timestamps should get stored and transmitted in a normalized and 100% unambiguous way.
It's only when you get into supporting all the pointless and highly ambiguous but valid variants of ISO 8601 that parsing becomes a problem. There's actually no such thing as a parser that can parse all valid variants with no knowledge of which variant is being used. There are lots of libraries with complex APIs that support some or all of the major and minor variants of course. But not with just one function called parse().
I think the main challenge with ISO 8601 is that it never called out this variant as a separate thing that you should be using. This really should be its own standard. And not using that would be a mistake. ISO 8601 is what happens when you do design by committee.
> You can use datetime libraries that are widely available to localize timestamps as needed in whatever way is required locally. But timestamps should get stored and transmitted in a normalized and 100% unambiguous way.
If by "timestamp" you mean past dates and deterministic future dates, then agreed. (Although I prefer unix epoch in ms for those, to be able to use integers and skip string parsing steps completely.)
But if your unlucky enough to need to handle future dates, especially "clock on the wall" ("let's meet in Frankfurt on July 26th 2029 at 1pm"), then you just can't know the timezone. The reasons can be many political ones, but especially in this case there's a high probability that EU will remove daylight saving time by then.
So in those cases, if you want to be correct, you'd need to include the geolocation in the stored timestamp.
That's a localized date and time. And it shouldn't be what you store but what you present based on the user's preferences. The same calendar event would be completely different in the calendar of a tourist visiting Frankfurt. And calendar invites of course can have multiple participants that each want their date times localized to their time zone and locale. So the best way to do that is to store the time zone in a 100% normalized, unambiguous way that is then transformed into whatever matches the user's preferences as part of the presentation logic.
In the same way, place names are localized. The French would spell Frankfurt as Francfort, for example. Location should be a separate field. And it implies nothing about the locale that should be used for presenting the timestamp or the timezone that should be used. Because Frankfurt is a rather international town and some might prefer MM DD YY instead of DD MM YYYY. Not to mention the use of am/pm vs. 24 hour time. And of course it has a big airport that people use to travel to pretty much all time zones on the planet. Hard coding all that in a timestamp is a mistake. Because now you need needlessly complex parsing logic and transformation logic to fix that mistake. Which is what this article is about.
Note that I'm not arguing against storing the numeric part in a consistent format, of course you should. My point is that right mow in 2025 you don't know the timezone.
E.g. if you write that as the zones are now, it would be "2029-07-26T11:00:00.000Z", but if the timezones change, then when a person is looking at their calendar in 2029, they will be an hour early or late to the meeting.
So it's not about presentation of the timestamp, it's that the timestamp does not match what the user meant.
> In the same way, place names are localized.
I beg to disagree. All spellings refer to the same phyysical place, but you can't guarantee my example timestamp description and a simple ISO time representation without a geolocation will refer to the same singular point in time.
I agree with the gist of your comment but I think you’re using the wrong words.
A time zone, as the name implies, is a zone, a place. The most common way to indicate a timezone is a continent+city combo that is both politically and geographically close, so it’s unlikely to change. Therefore, I very much know the timezone of Frankfurt, both today and in the reasonably close future, namely “Europe/Berlin”.
You’re talking about the timezone offset. You’re right, don’t store that for future local dates. Store the timezone.
(Or just the exact location, but then you need a way to map location -> offset if you ever want to do time math on it, and the IANA Continent/City format strikes a nice pragmatic balance there)
Well, not really. :) Timezones are the whole area that has the same time. So "Europe/Berlin" and "Europe/Rome" right now can be used in user-facing help texts to make the timezone understandable, but those cities might not be in the help text in the future. That's because:
That's why I stand with my original point: if you want to support "clock on the wall" future dates, a geolocation is needed. And given that Germany was two countries in the recent past and Frankfurt and Berlin were in different countries, I wouldn't solve this by picking a city from the help text that's currently closest to the timezone, but instead use the actual geolocation of the event.
P.s. All this is, of course, for 99.99% of the cases a complete overkill, and as said, I haven't had issues in using unix epoch for all dates. But just wanted to say that timestamps aren't as simple as I felt was suggested in the first comment.
The way I understand it, Continent/City names aren't for pretty dropdowns or help texts (in fact, I'd wager they're kinda meh for that). They're the least bad way people have come up with to store and communicate timezones with future local dates, that doesn't require online geolocation. You're right that geolocation is the best. But offsets aren't a close second, like I feel your comments suggest - they're just a bad idea. To suggest that a timezone "is" an offset, while technically correct if you disregard future changes, is actively harmful to helping people write code with fewer timezone bugs.
Use IANA Continent/City names. They're great, all proper datetime libs support them, your OS supports them, they're reasonably human readable and completely machine readable. They're better for storage than offsets (or terrible future-incompatible terms like CET) in every way.
Really the entire timezone debacle can IMO be summarized as:
- for datetimes in the past, store UTC. convert from/to local in the UI; just assume the system/browser locale is the one unless you know otherwise.
- for global datetimes in the future, eg comet passings and the likes, same thing.
- for local datetimes in the future, at some "wall clock time in a place", store localtime + an IANA Continent/City combo, and keep your servers' tzdata files reasonably updated.
It never makes sense to store or communicate offsets. It hardly ever makes sense to do geolocation unless you think you think the chance that Berlin moves to a different country than Frankfurt is bigger than the chance that you lose access to up-to-date geolocation information.
Note that the above implies that the option for ISO date strings to include offsets (or words like "CET") is, IMO, nearly always the wrong choice. For past dates (ie 99.9% of dates stored by computer systems) just use UTC, and for future dates, IANA timezone names are strictly better.
Or in 2027 the Sith Empire could conquer Earth and install imperial time system. Good luck with your timestamps. Just make your scheduler as a walled garden SaaS - no need to exchange timestamps.
This article doesn't get into some of the special fun of ISO 8601, including relative values, non-Gregorian values, durations...
Some of the things in the standard are surprising, like maybe were a special request. At the time, I commented, something like, Somewhere, in the French countryside, there is a person who runs an old family vineyard, that is still stamping their barrels with the timepoint information [...]. And that person's lover was on the ISO 8601 committee.
(I once wrote an time library in Scheme that supported everything in ISO 8601. It did parsing, representation, printing, calendar conversion, and arithmetic. Including arithmetic for mixed precision and for relative values. It was an exercise in really solving the problem the first time, for a core library, rather than cascading kludges and API breakage later. I don't recall offhand whether I tried to implement arithmetic between different calendar systems, without converting them to the same system.)
I used to work at a company that stored all dates as ints in a YYYYMMDD format. When I asked why, I was told it was so we could subtract 2 dates to get the difference.
I asked them why they couldn’t use DATEDIFF since this was in a sql db.
They said they hadn’t heard of it and that it must be new.
Which sort of works, but then you also have to deal with the cases where the days or months go negative (at the month and year bounds), and that also involves knowing how many days there are in each month. It's pretty difficult for me to imagine how this could be easier than just converting to e.g. unix time and subtracting seconds.
No other programming concept has caused me more grief than dealing with time and timezones. It starts to get really mind-bendingly complex once you start thinking about it deeply. That is even before you start encountering the quirks (some places have timezone changes that depend not only on the time of year but also on the actual year). Lesson learnt - choose a library (moment is great) and never think about time again.
Except you do still have to think about time, no matter what… Libraries will help with the really messy details, but even the high level requirements have a lot of pitfalls.
“Simple” example that anyone who’s ever worked on a scheduling application will probably be familiar with:
“Get a list with all of today’s events.”
Well, whose “today” (timezone) are we talking about? Server, client, setting in the user account? Or none of the above, and actually timezone at the physical location of the event, if there is one?
And what does “today” mean, anyway? Truncate the date? 00:00-23:59? Business hours?
And what does “today’s event” even mean? Events can cross midnight… Does an event need to start today? End today? Both? Can events span multiple days?
I understand the complexity of time, but this scenario doesn't seem all that difficult. Any user expects (I won't get into what they want) to be shown _their_ day, and they expect it to be local calendar time. You _might_ want to show wee hours events, from midnight to just short of 5am. Apps like Teams do a good job of spacially illustrating time as it matters to you _and_ as it matters to others.
Libraries can't paper over the hard parts of dealing with timezones. There are many "right answers" for most date problems. And, they're all what you want for some context. So, the library can't be opinionated about most things.
You just need to understand how time works if you write code handling time.
I agree with everything other than "moment is great". Even the devs say don't use it any more, and accidentally mutated datetimes have been the source of bugs in apps I've worked on multiple times. Luxon is great though.
Unfortunately, not many people in our industry really understand the common pitfalls of timezone management, including the people who set the requirements. The classic "I want this to run at 2am each weekday for the user", and then proceed to simply store that as a UTC offset and call it a day (pun intended).
It's also a beautiful maze of nerd snipes. There's apparently some shifting going on inside the earth, combining with the slight variability of the moon's distance, which means various days over a month can be more than a millisecond shorter than average. Good luck integrating that into your date.addDays()
The really important thing to remember about timezones is they're not a mathematical construct, or a physics construct, or an astronomy construct... They're a political construct. Framed in that light, they have every bit the complexity of having some piece of your code depend on the law (except it's the law of every nation you expect to be running your code in).
Do your users type in such dates? No? Problem solved.
The benefit of DIY parsing is to make the problem simple by restricting it to the set of plausible inputs your users will want your code to handle, not to make a highly general library. The right takeaway for juniors is to stop over-complicating things.
> Do your users type in such dates? No? Problem solved.
This is spot on. So many of the "X is really hard and your intuition is wrong" takes ignore the fact that most people are not building something which needs to be usable in every country, language, and culture on this earth. Yes, human behavior is insanely complex, but for any given application you can probably ignore huge swathes of it.
I put these into an ISO8601 parser and it didn't work, I'm going to tell ISO off for obviously DIYing their solution instead of doing the proper thing.
I requested an ISO 8601 date parser in the Python "datetime" library in 2012.[1] "datetime" could format into ISO 8601, but not parse strings. There were five ISO 8601 parsers available, all bad. After six years of bikeshedding, it was was fixed in 2018.
That's what it took to not write my own date parsing library.
moment is far smaller if you include it without locales you don't need.
I don't care how much they talk themselves down on their homepage, begging me to choose a different library - I like it and I'll continue using it.
> We now generally consider Moment to be a legacy project in maintenance mode. It is not dead, but it is indeed done.
> We will not be adding new features or capabilities.
> We will not be changing Moment's API to be immutable.
> We will not be addressing tree shaking or bundle size issues.
> We will not be making any major changes (no version 3).
> We may choose to not fix bugs or behavioral quirks, especially if they are long-standing known issues.
I consider this a strength, not a weakness. I love a library that's "done" so I can just learn it once and not deal with frivolous breaking changes later. Extra bonus that they plan to continue making appropriate maintenance:
> We will address critical security concerns as they arise.
> We will release data updates for Moment-Timezone following IANA time zone database releases.
>Consider "200". Is this the year 200? Is this the 200th day of the current year? Surprise, in ISO 8601 it’s neither — it’s a decade, spanning from the year 2000 to the year 2010. And "20" is the century from the year 2000 to the year 2100.
there is so much wrong with this paragraph, it's a nest of people who shouldn't work on date parsing. there is no way 200 is any kind of date, but if you're going to insist it is, 2000 to 2010 is 11 years unless "to" means "up to but not including" in which case it should say 2001 to 2011 if you want to refer to the 200th decade, since decade 1 was 1AD through 10AD...
> ... billed as “an extension to the ISO 8601 / RFC 3339” formats
The quote comes from MDN[1], but is wrong on so many levels:
1. RFC 3339 is not a format; it is a meta-format. It specifies properties that any timestring used in internet applications should have. One application could require that "T" be used to separate the date and the time, while another could require a space to be used, and both applications would conform to 3339. Using the same parser to parse times and dates from the two applications would be incorrect, as "1234-12-12 12:34 ..." unambiguously represents the date of December 12, 1234 in the former but the time of 12:34 on the same day in the latter.
2. RFC 3339 is thus not a subset of ISO-8601; there exists a subset of ISO-8601 which satisfies RFC 3339, but there also exist many timestrings that are not subsets of ISO-8601 which also satisfy it (most famously replacing "T" with a space, but that is merely an example for which RFC 3339 conforming applications may deviate from the ABNF in the RFC for readability).
3. It is obvious from both #2 and the table in TFA that RFC 9557 can't be an extension of ISO-8601 given that there are valid ISO-8601 timestrings in the table that are invalid RFC 9557 timestrings.
4. This is a minor nitpick, bu RFC 9557 also alters the semantics of some RFC 3339 timestrings with regards to offsets. RFC 3339 specified "Z" (or "z") for the offset being the same as +00:00, while 9557 has "Z" being the same as -00:00; the meat of RFC 9557 is adding suffix-tags to RFC 3339, so I wouldn't quibble with calling 9557 an extension of 3339.
> Most date parsing woes (in my opinion) come from ambiguity: from supporting too many formats or attempting maximum flexibility in parsing. ... There is a maintenance freedom and simplicity in strict parsing requirements ...
Is this the endgame of "be liberal in what you accept, but strict in what you output"? You end up supporting an untold amount of usage patterns, not even chosen by yourself if you use a library like `luxon` to handle the "be liberal" part for you.
In UIs prefer date/time pickers instead of raw text inputs which will give the date/time in standard ISO format such as ("2025-07-25" or "2025-07-25T18:47:26.022Z"). Prefer ISO formats everywhere where possible.
The thing that bothers me about the firefox date picker is it display the date in YYYY/DD/MM format.
While I am fairly sure this a a locale defined thing. locales are this huge pile of worms and I have never figured out how to change it to show YYYY-MM-DD format
Great writeup! Your journey perfectly captures the universal developer dilemma: "Never roll your own X... until you absolutely must."
The bundle size reductions are impressive (230kB client-side savings!), and your RFC 9557 alignment is a smart forward-looking move. Two questions:
Edge cases: How does your parser handle leap seconds or pre-1582 Julian dates? (e.g., astronomical data)
Temporal readiness: Will @11ty/parse-date-strings become a temporary polyfill until Temporal API stabilizes, or a long-term solution?
Minor observation: Your comparison table shows Luxon supports YYYY-MM-DD HH (space separator) while RFC 9557 doesn’t – this might break existing Eleventy setups using space-delimited dates. Maybe worth an explicit migration note?
Regardless, fantastic work balancing pragmatism and standards. The web needs more focused libraries like this!
I wrote the ethlo ITU library because I was annoyed with the lack of performance and large amount of ceremony to parse and format standard timestamps in RFC-3339 format in Java. It is somewhat more extensive now, and is used in other libraries. Ask me anything!
Obviously there is a bit of nuance here, but the best rule of thumb is use a regex and then warn the user that it doesn't look valid but still accept it if it doesn't pass, and then send an email to verify it.
The table of listed date formats doesn't look too difficult to implement. A quick look at the RFC tells me, that the RFC even specifies a grammar, though very incomplete. It would be prudent to specify a complete grammar in the RFC of course. Then it would be even simpler taking that grammar and translating it to whatever library one uses for describing grammars for parsing stuff. I really hope all these libraries didn't make silly things with regexes ...
Good general rule of thumb, but desperate scenarios call for desperate measures. I would never do this in Python or Rust for example, but it's necessary in Javascript; `Date` and `Moment`, are so full of traps that the ends justify the means: Especially if you have use for a `Date` or `Time` type.
moment's given me no trouble at all. I certainly haven't found it to be full of traps. Addressing the most common complaint: a moment object is mutable, sure - that's a valid design choice, not a trap. Follow the docs and everything works perfectly well IME.
As a specific point, I have not safe found a way to represent a date or time in Moment. When I point this out, I generally get agreement from people who are more used to other languages, and the claim that "You should never be representing a date or time; everything should be a datetime" by JS devs.
Now i want to make a date format, combined with other data that is the ultimate challenger of date parsing.
IntroDuceThing: The ip:port,date/time,longlat string. Oh, yes its format is also dependant on the language you encode it in and what parts you leave out to be defaulted. .:, is now a valid locationdateip
I'm not even a programmer, but I can tell that dates are ambiguous a lot of the time.
e.g. dd/mm/yyyy (British) and mm/dd/yyyy (USA) can be confused for the first twelve days of every month.
So, given the high volume of international communication, I think we should hand-write months in full, or at least as the first three letters (Jan, Feb, Mar, ..., Dec)
We should also abandon three-letter acronyms (but that's another story).
I wrote one in Erlang years ago for Riak’s time series implementation. I don’t remember all of the motivations, but most of all I wanted the ability to encode incomplete date/time objects.
But subtle plug of something I made long ago for when you find your data pipelines are running hot parsing timestamp strings etc: https://github.com/williame/TimeMillis
I’m still pumped by the performance of the thing! :)
I once had to maintain a CalDAV server that was developed in house, computing the "free busy" with recurring events, exceptions, different timezone than the organizer + some DST is a bug source that keeps on giving.
Maybe the title should be “it is difficult to write a date parsing library”
“Never write your own x” kind of titles come off as arrogant and demotivating.
Maybe some other person will write an excellent date parsing library that will be better than current ones? Maybe they think it is worth to spend some time on it?
These kinds of hard things tend to have libraries that are extremely bloated because everyone uses one library, and that one library has to work for everyone’s use case.
You can see this in the post too, not everyone needs to be able to parse every single date format.
Multiple times in my career I've had a good laugh when a non-technical manager says something along the lines of "it's just the date, how hard can it be?"
Agreed! My team is constantly humbled by the mess of user data: names, birthdays, addresses, people dying or living abroad etc.
Honestly, sometimes I think about the linear algebra, AI, or robotics I learned in school and get this feeling of, "Is this what I'm doing? Stuff that feels like it should be simple?"
It's funny, even our product manager - who is a great guy - can fall into that "come on, this should be easy" mode, and I'll admit I sometimes get lulled into it too. But to his credit, every time I walk him through the actual edge cases, he totally gets it and admits it's easy to forget the on-the-ground complexity when you're in 'planning mode'.
So yeah, seeing your comment is incredibly validating.
Which is not relevant unless you are trying to support every culture on the planet. Which most people aren't, and for them it'll be just fine to have first+last name.
Well, I did that. And it works flawlessly for a decade now. The thing is just that I know and control the context from which the dates are being parsed. If you're now like: "Yeah ok if you're the one who sends the data being parsed it might be okay", the claim of "never do X" is proven wrong if there are specific situations where soing X is not only okay, but might be the sensible option.
Which is why you should never use the word "never" unless you're really sure you can't come up with a situation that is an exception.
Like for example, lets say you are writing very performance sensitive code where nanoseconds count. All of the date parsing libraries available for the language you are writing are too slow for your requirements. So you might roll your own lighter weight faster one.
I wrote my own, so had to click, but mine was for a very different use case: converting extremely varied date strings into date ranges,
where a significant % of cases are large number are human-entered human-readable date and date range specifiers, as used in periodicals and other material dating back a century or two.
I.e. I had correctly interpret not just ISO dates, but, ambiguous dates and date ranges as accepted in (library catalog) MARC records, which allows uncertain dates such as "[19--]" and "19??", and, natural language descriptors such such as "Winter/Spring 1917" and "Third Quarter '43" and "Easter 2001." In as many languages as possible for the corpus being digitized.
Output was a date range, where precision was converted into range. I'd like to someday enhance things to formalize the distinction between ambiguity and precision, but, that's a someday.
When schema is totally uncontrolled, many cases are ambiguous without other context (e.g. XX-ZZ-YYYY could be month-day-year or day-month-year for a large overlap); and some require fun heuristics (looking up Easter in a given year... but did they mean orthodox or...) and arbitrary standards (when do seasons start? what if a publication is from the southern hemisphere?) and policies for illegal dates (Feburary 29 on non-leap-years being a surprisingly common value)...
In a dull moment I should clean up the project (in Python) and package it for general use...
So go ahead, write your own date library, your own Unicode font rendering, compiler, OS, game engine or what ever else people tell you to never do because its hard.
reply