## Nerdview

Language Log readers may appreciate the following classic example of writing in technical terms from the perspective of the technician or engineer rather than from a standpoint that would seem useful to the customer or reader. I was engaged in reserving a rental car on the web, and got the date syntax wrong. Instead of having one of those little click-on-the-calendar widgets, the site required entry of the date of the reservation in a blank box using a strict syntax that it did not explain until the reservation failed and an error message was displayed. By strict syntax I mean (i) slash must be used as separator (continuous numbers will not work), (ii) day must be before month (American month-day-year syntax will not work), and (iii) the date must be four digits (two-digit year indications won't work). I'm not worried about having to know the date syntax of the culture I'm in; I can deal with that. And although the 4-digit year is a bit crazy (since time travel is impossible, the first two digits will be 2 and 0 for all reservations for the next 92 years, so they are not carrying much information), that piece of programming stupidness is not my concern here. The classic bit was the error message that popped up in red:

Please select a valid pick up date (DD/MM/YYYY) greater than today.

Have you ever said "I could do lunch any day next week greater than Tuesday"? Or "It would be helpful if you could deliver it greater than the 27th"? Or "I'm younger than my wife because my birthday is eight days greater than hers"? Of course you haven't.

Only for the utter benighted nerd is the greater-than relation, as used in integer mathematics, defined on dates. Sure, in Unix date program date-setting notation, today is 0624 and tomorrow is 0625, so the sort -n command sorts them in that order. A nerd will keep his records for June 2008 in a file called 200806 and the ones for July in 200807 so that the ls program shows everything in chronological order. These facts might cause the mathematically inclined to think of dates as integers so that one can be greater than another. But ordinary people trying to rent cars say "later than" or "after"; they don't say or think "greater than".

And by the way, let me just point out (if I may risk using nerdery to fight nerdery) that under British day-month-year syntax, the greater-than relation will not work correctly: the last day of this month will be 30062008 and the follow day will be 01072008, and 01072008 (= 1,072,008) is not greater than 30062008 (= 30,062,008), so sort -n would not sort the dates correctly into chronological order if they were kept in that form. The same holds if the slashes are retained rather than erased.

The problem I am pointing to, however, is not about web programming or sorting technicalities. It is a simple problem that afflicts us all: people with any kind of technical knowledge of a domain tend to get hopelessly (and unwittingly) stuck in a frame of reference that relates to their view of the issue, and their trade's technical parlance, not that of the ordinary humans with whom they so signally fail to engage. I have written about this before — in Per bus per journey, for example, and probably on several other occasions. The phenomenon — we could call it nerdview — is widespread.

1. ### MattF said,

June 26, 2008 @ 6:44 am

True nerds use the Julian day.

2. ### itai said,

June 26, 2008 @ 7:18 am

When comparing dates, true nerds don't care for days, months and years. They use Unix time (http://en.wikipedia.org/wiki/Unix_time). Unix time being a number, "greater than" is well defined.

If you want date strings to be lexicographically sortable, then you use ISO 8601-formatted dates (http://en.wikipedia.org/wiki/ISO_8601).

3. ### Chris C said,

June 26, 2008 @ 7:42 am

Hello,
you are of course correct about the error message being dreadful but wrong about the "greater-than relation" , I think.
It seems to me that one of the features of mathematics is to spot common features in apparently different domains and unify them with a common abstraction. For instance many people think it is sensible to use different words to describe the ordering relationships between integers (fewer than) and reals (less than). Mathematicians and other nerds do not. Similarly having a different set of words to describe ordering between times and dates is clumsy (in the context of software at any rate). There are many other things that need to be ordered and as a rule we don't invent new words to describe those relationships and stick with greater than and less than.

I agree that the context in which you talk about a date being "greater than" another date should only be to other nerds who it will not confuse or aggravate but I can only agree with the rest of your comment up to a point.

4. ### ed said,

June 26, 2008 @ 7:56 am

The Made to Stick guys call it the "curse of knowledge" or something.

5. ### Mark P said,

June 26, 2008 @ 8:06 am

Actually, I think MattF is right. But there is no excuse for it. But I blame the manager who let the programmer make that kind of mistake without correcting it.

6. ### Doria said,

June 26, 2008 @ 8:16 am

If you want to get the real nerdview try using Unix time to keep track of the date.

7. ### Kellen said,

June 26, 2008 @ 8:37 am

my birthday is greater than everyone's.

8. ### mark (the ideophone) said,

June 26, 2008 @ 9:04 am

I like the word. Many hopeless user interfaces are the result of nerdview. Nerdview means being satisfied when your code does the job for you, instead of thinking for one moment about the user.

Nerds who suffer from nerdview seem to lack certain crucial mindreading capabilities, a fact that renders them quite similar to some of our primate relatives. I see an interesting new field for comparative primatology.

9. ### James Wimberley said,

June 26, 2008 @ 9:08 am

The explanation – not justification – for why "greater than" is a consistent usage for nerds is indeed that many computer programs convert dates into unique integers. Consult your Excel help on the DATE function: dates are stored as the number of days since 1 January 1900. Tough luck if your interests lie in nineteenth-century British financial markets; Excel doesn't produce a negative number for 1/1/1808 but garbage.

10. ### dr pepper said,

June 26, 2008 @ 9:15 am

I always use yymmdd or yy-mm-dd format. As for Excel, no prob, just make that column text.

11. ### JScarry said,

June 26, 2008 @ 9:52 am

Ultimate responsibility for what the users see rests with the business types—not the programmers. They had ample opportunity to change it if they didn't like it. But more likely, they didn't notice any problem since that's how they think about dates when using a computer. Most database programs use mathematical symbols for search operations on dates. The managers are most likely comfortable searching for say—sales>1000 and date>06/01/2008—to find the current month's big spenders. They probably didn't even notice that, when written out, it looks a bit odd.

The real problem is that the programmers/managers were too lazy to explain how they wanted the date formatted or better yet, to provide boxes for day, month, and year (defaulted to 2008) for the input fields.

12. ### Jonathan Lundell said,

June 26, 2008 @ 10:24 am

I've noticed nerdview in advertising, too. I've been enjoying my rTMS a little too much this morning or I'd have more and better examples, but here's one: referring some insurance policy as a "product".

13. ### Craig Russell said,

June 26, 2008 @ 11:04 am

@ Jonathan Lundell

I've noticed this too–I would file referring to creative works as "content" in the same category.

I wonder if there might not exist a kind of corollary to this: the annoyance people feel when someone else tries to usurp their own specialized and technical vocabulary, and end up using it differently?

For example, Language Log's contributors seem particularly annoyed when they see people trying to apply the vocabulary of grammar and linguistics to other fields: e.g. when someone makes a statement like "peace needs to be a verb, not a noun"–a statement whose meaning seems pretty clear to me ("peace is something that has to be achieved through action"), even if it is totally nonsensical if analyzed from a linguistic point of view (how can the word 'peace' function grammatically and syntactically as a verb rather than a noun?).

Perhaps the corollary to members of a specialized field unconsciously slipping into their field's terminology is the same members being unusually unwilling to accept other people's use of this terminology in senses other than the ones they understand.

14. ### Stephen Jones said,

June 26, 2008 @ 11:04 am

—-"Only for the utter benighted nerd is the greater-than relation, as used in integer mathematics, defined on dates."—–

And only for linguists looking for a heresy does is produce any problem in communication. Should we name the syndrome Pullum's Peeve?

The use of a blank text box with limited rules as to data input is attacked in all usabilty manuals, also written by geeks and nerds.

The most logical explanation is they were cheap, and hired somebody to do their web site who took his error messages from a Google search.

15. ### Andy J said,

June 26, 2008 @ 11:12 am

"It is a simple problem that afflicts us all: people with any kind of technical knowledge of a domain tend to get hopelessly (and unwittingly) stuck in a frame of reference that relates to their view of the issue, and their trade's technical parlance, not that of the ordinary humans with whom they so signally fail to engage."
That elloquently describes my thoughts after reading many a LL piece, especially some of Mark Liberman's heavy-on-statistics blogs, LOL. However I would admit that the fault is mine, as I am an ordinary human being in a self-evidently specialist domain. Sometimes it is good to see things from a different perspective.

16. ### David Eddyshaw said,

June 26, 2008 @ 12:12 pm

Barbari!
Hoc scripsi a.d. VI Kal Iul A.U.C. MMDCCLXI.

17. ### Dan T. said,

June 26, 2008 @ 12:51 pm

Re: "how can the word 'peace' function grammatically and syntactically as a verb rather than a noun?"

Perhaps in the slang exclamation "Peace out"?

18. ### Andy Hollandbeck said,

June 26, 2008 @ 12:57 pm

Mark P: "But I blame the manager who let the programmer make that kind of mistake without correcting it."

It's easy to blame the managers for not catching the "error," but then you have a separate problem: The person who is expected to fix the error is the nerdly programmer who made the mistake in the first place. Plus, a mistake like this, which will cause no horrible disasters, would fall pretty far down the list of things to do.

Even worse, if the uber-nerd actually wanted to keep that text that way, it wouldn't be too difficult to put off the change indefinitely with nerdese doublespeak explaining why attempting to change this text creates certain technological hurdles that don't make it cost- or time-effective to jump over.

19. ### Andy Hollandbeck said,

June 26, 2008 @ 1:06 pm

David Eddyshaw,

Get with the times! It's the Computer Age!
It's day 10110010 of year 11111011000!

20. ### Amos said,

June 26, 2008 @ 1:32 pm

Even worse than the "greater than" usage: why are they telling you to make the reservation in the future, not the past? Clearly the result of a test that raises an exception if either the format is wrong or the date is in the past.

21. ### Cavity Lee said,

June 26, 2008 @ 2:16 pm

Hm. I'm not sure I can agree with this post's implicit assumption that the only way for one date to be 'greater than' another from within nerdview is for dates to be considered integers. To a programmer, "greater than" can be defined on all kinds of data– the string "Pullum" is greater than the string "Liberman", for example. I strongly suspect that whoever wrote that error message was using an abstract "date" data type in which the error message was correct and whether one date was "greater than" another did not have anything to do with whether dates were written to the screen in British DD/MM format.

(I agree the error message is stupid; I am disagreeing only with the second-to-last paragraph of the post.)

22. ### Craig Russell said,

June 26, 2008 @ 2:33 pm

Dan,

I had never thought about the origins of or structure behind "peace out". If I had to analyze it, I think I would do it this way (and this is all based on guesswork; it's not backed up by anything):

"Peace out" is shortened combination of two expressions: "Peace!" and "I'm out."

"I'm out" is presumably short for something like "I'm going out". Or maybe "out" is being used here like the radio expression "over and out", to indicate the end of a conversation (perhaps the 'out' in "I'm out" is a kind of predicate adjective(?) that roughly means "finished"?)

But I would connect "peace" to expressions like "peace be with you", for which I would imagine there is continuity going back many centuries. I notice from a quick googling that "peace to you" is common, which I would call an extension of the "peace" that more or less means "goodbye". It seems analogous to "thanks" (which can also be expanded with "to", e.g. "thanks to Mr. Smith for this lovely bouquet"), which I would also categorize as a noun with a missing verb ("thanks be to…" or "let there be thanks").

So I would still call "peace" a noun in the expression "peace out" (although I would be less sure of exactly how to label "out"). But that's just my two cents.

23. ### Sniffnoy said,

June 26, 2008 @ 4:18 pm

I've actually heard "peace out" used as a verb phrase – e.g. "I have to peace out" or "I'm peacing out".

24. ### Frank Oswalt said,

June 26, 2008 @ 4:22 pm

"Greater than" may be unusual in the context of dates, but it's not as if it's particularly difficult to understand. To me, it seems unlikely that this error message will cause problems for any user smart enough to rent a car online in the first place.

@James Wimberley: Just use OpenOffice Calc instead. It is free, and it can handle dates before 1900-01-01 like a charm.

25. ### John Laviolette said,

June 26, 2008 @ 4:22 pm

@ Andy Hollandbeck:

I think you meant: today is September 5413, 1993.

26. ### Nathan Myers said,

June 26, 2008 @ 4:34 pm

It's very rare for anyone testing software, particularly a web page, to spend any time on checking error handling. It's not from laziness; people responsible for testing are notoriously overloaded, and it's very hard to find more of them. Usually it's all they can do to make sure things work when users do everything right. Most managers would be satisfied with the universal response from the original Unix editor "ed" to any error: a simple "?".

27. ### Rubrick said,

June 26, 2008 @ 5:00 pm

I'd also argue that "select" is a poor choice in this context. "Select" usually denotes choosing from a menu or list; for free-form text entry, "enter" is better. (Aruguably "enter" is itself a too-technical term, but I think it's well-understood by most people these days, and it doesn't seem to have a suitable "lay equivalent"– simply "type" isn't quite the same thing.)

28. ### Mark said,

June 26, 2008 @ 5:24 pm

Why do geeks dress up as Santa for Halloween? Because Dec 25 == Oct 31!

Gotta love geek date math humor!

29. ### Craig Russell said,

June 26, 2008 @ 6:59 pm

Sniffoy,

Boy, you are sure right–there are over 10,000 Google entries for "peacing out", and 24,000 for "peacing". I hadn't heard it before, but I guess it makes sense, given the ease with which words move from one category to another in English. I also found 1180 for "Clintoning", 1280 for "McCaining" and 91 for "Obamaing"–is there any word in English that can NOT be used as a verb?

30. ### dr pepper said,

June 26, 2008 @ 7:22 pm

> Barbari!
> Hoc scripsi a.d. VI Kal Iul A.U.C. MMDCCLXI.

Qui Rem Publican amant, debent scribere "Quin", non "Iul".

31. ### Janice Huth Byer said,

June 26, 2008 @ 9:09 pm

Some comments point out "greater than" is clear, but that's only with reference to numbers not with time.. Albeit functionally harmless, it's still an unwanted distraction from the content.

32. ### aph said,

June 26, 2008 @ 10:21 pm

Distracting, sure. But what we nerds are pointing out is that the digression about whether applying "greater than" to British dates is not accurate. As several have noted, in all modern computing dates (as a data type) are defined as the number of some time unit to have elapsed since some agreed-upon time – the number of days since 1900 or the number of seconds since 1970/01/01, etc. As such, more recent dates are unequivocally greater than earlier ones.

In other words, the post is incorrect in suggesting that the "nerd" usage of saying that one date is greater than another does not work across locales.

33. ### jon w said,

June 26, 2008 @ 10:31 pm

By 'British day-month-year syntax' you mean everybody outside the US

34. ### Joe said,

June 27, 2008 @ 2:50 am

> And although the 4-digit year is a bit crazy (since time travel is impossible, the first two digits will be 2 and 0 for all reservations for the next 92 years, so they are not carrying much information), that piece of programming stupidness is not my concern here.

That's the exact thinking that gave us the Y2K bug. While it did not cause massive damage, that was because a lot of people worked hard to fix things beforehand. You can't assume that someone will remember to fix it 92 years from now, nor that it won't still be in use by then. Ugly code is usually the most stable, after all, simply because no one wants to go near it (and they probably lack the time to replace it).

That said, you're correct that they should be using something like a calendar widget. But what they made was a lot easier to produce than one of those, so it may very well be a case of them getting what they pay for.

35. ### Dan H said,

June 27, 2008 @ 3:44 am

If what the article says is true, how come I often hear laypeople use phrases like, "since at least last week?" This use is, I'm sorry to reveal, one of my pet peeves, because you don't know whether the "at least" refers to the period of a week between the event and now, or the date of the event itself. "At least" is a common expression often used even when more specific comparatives like "earlier/later" exist, in a way that "greater than" is not, so the example you give is still very idiosyncratic. I'm afraid I favour the "always favour a specific term over a general term" response to this kind of ambiguity, and always use "at earliest" and "at latest" instead.

Anyway, the message you quote was clearly written by a novice; real computer scientists never say, "greater than," preferring the more precise alternatives, "strictly greater than," and "at least." It's one of those habits you pick up to avoid confusion when speaking in maths and then carry across to everyday speech.

36. ### David Eddyshaw said,

June 27, 2008 @ 7:28 am

doctissime Dr Pepper

Recte scripsisti; non omnes autem amamus istam rem publicam.
Ave Caesar Auguste, qui nobis pacem rettulisti!

37. ### James Wimberley said,

June 27, 2008 @ 10:52 am

Frank Oswalt: I do use OpenOffice not Excel and Word. But Microsoft's only remaining advantage lies in the detail in the help files, which are what I referred to. To geeks these are boring, low-status stoop labour and they will only do it if paid.
OpenOffice Calc starts dates in 1583 not 1900 so it would cope with my hypothetical paper on the 1857 British banking crisis, but not the one on mediaeval Genoese bonds (see Braudel).

38. ### JakeT said,

June 27, 2008 @ 4:12 pm

I always thought "Peace out" was "Peace…I'm out" which, in turn, is "Peace to you…I'm out of here."

If I was guessing, I'd say the software has one function that converts human readable dates to Unix time then compares them.

39. ### Alex said,

June 28, 2008 @ 3:37 am

It's easy to blame the managers for not catching the "error," but then you have a separate problem: The person who is expected to fix the error is the nerdly programmer who made the mistake in the first place. Plus, a mistake like this, which will cause no horrible disasters, would fall pretty far down the list of things to do.

Even worse, if the uber-nerd actually wanted to keep that text that way, it wouldn't be too difficult to put off the change indefinitely with nerdese doublespeak explaining why attempting to change this text creates certain technological hurdles that don't make it cost- or time-effective to jump over.

Depending on the product, language, and company's base, it probably isn't a programmer who wrote the phrase, either — it would be somebody specializing in localization. (Which is to say, somebody paid to make sure the program is easy to use in some collection of languages, largely because it uses well-written text in each of the languages.) Admittedly, this rental car reservation site may not be available in multiple languages, and if it is, there is a pretty good chance the programmers are English-speaking, in which case they probably did write that phrase.

Please select a valid pick up date (DD/MM/YYYY) greater than today.

[…]
And by the way, let me just point out (if I may risk using nerdery to fight nerdery) that under British day-month-year syntax, the greater-than relation will not work correctly

The date is an abstract concept, independent of representation. The parenthetical remark indicates the string format that should be used, but does not change the sentence being about dates instead of date strings. Hence, dates represented in day-month-year syntax should still compare as expected (since the representation doesn't matter), though date strings in day-month-year format would not.

40. ### Mark said,

June 28, 2008 @ 11:52 am

And by the way, let me just point out (if I may risk using nerdery to fight nerdery) that under British day-month-year syntax, the greater-than relation will not work correctly: the last day of this month will be 30062008 and the follow day will be 01072008, and 01072008 (= 1,072,008) is not greater than 30062008 (= 30,062,008), so sort -n would not sort the dates correctly into chronological order if they were kept in that form.

Who said anything about sort -n? It's improbable that a programmer would want to perform a purely textual comparison on the entries. More likely, these dates will end up getting parsed and passed as arguments to an SQL query somewhere, or represented as a date object in whichever language the web site is written in, and the comparison will be valid.

Anyway, it's silly to get worked up about a little nerd terminology slipping through to the user interface. You're renting a car over the frickin' Internet on a frickin' computer; of course you'll be exposed to some of the native jargon here and there. This is akin to going on a tour boat ride and getting indignant when the captain brings your attention to some scene on starboard, rather than "on the right."

41. ### Mr. Shiny & New said,

June 28, 2008 @ 2:41 pm

I thought I would have to point out that the formatting of a date doesn't affect which one is earlier and thus "less than" another date, but Alex beat me to it. But I did want to mention to Joe that even though 2-digit years lead to y2k problems, the problem isn't in the user interface as much as in the storage of years. So it's ok if your user interface lets you type in a two digit year (heck, even a 1 digit year!) as long as it A) can guess what year you mean, and B) stores the whole year in the long-term storage (file, database, whatever).

That said, I agree with the post that the error message is wrong. Just because in database programming you compare dates using doesn't mean you should read those symbols as "less than" and "greater than", rather they are "before" and "after". Java gets this right in its Date programming library, where the function used to compare one date to another is before() and after().

42. ### Andrew Johnson said,

June 29, 2008 @ 9:07 am

Although the error message is quaint, it is easy to understand. "Select" is entirely appropriate if you mean "choose a date from the set of possible dates" – that is, not 99/99/2008. Using "greater than" with dates like this is indeed nerdy, but the operator > makes perfect sense, even if reading it as "greater than" is not idiomatic.

But the real danger is that an American would type 8 July 2008 as 07/08/2008, which would be silently accepted as 7 August 2008.

And according to the error message, you can't make a booking for later on the same day.

June 29, 2008 @ 1:55 pm

maxime doctissimi Dr. Pepper et David Eddyshaw:

vero, secundum Romanorum usum in quo uterque primus ultimusque annus includuntur, nos opportet scribere A.U.C. MMDCCLXVII.

quippe scribere possimus "praesidente(pro consule aut consulibus) octiens Bushense filio Bushensis". consentitisne?

June 29, 2008 @ 6:01 pm

eheu, mea culpa: MMDCCLXII

45. ### Cris said,

June 29, 2008 @ 8:48 pm

I've always heard "Peace out" used exactly as one would "Chill out".

46. ### Drew Kime said,

June 30, 2008 @ 12:13 pm

Even worse than the "greater than" usage: why are they telling you to make the reservation in the future, not the past? Clearly the result of a test that raises an exception if either the format is wrong or the date is in the past.

A little usability would be nice here. Tell the user what they actually did wrong. If it was a bad date format, show them the correct format. If they selected a (properly-formatted) date that was in the past, tell them that.

And if they selected today, and the system truly won't accept a same-day reservation, I'd handle that as a special case. But I wouldn't argue too hard if the high muckety-mucks wanted a single error message for all dates <= \$today.

47. ### Rich said,

August 3, 2008 @ 9:59 am

At my workplace it is a matter of policy that ALL years always be expressed with four digits, and stored as such. After all that spending on Y2K only a few years ago the last thing anybody wants is to have to rewrite all the company's software again in 100 years when it happens all over again.

Sure, this particular application might seem unlikely to persist so long, but you never know. When you start talking about enterprise-scale software the lifetime goes up significantly. Also – since some dates pertain to future events the issues can crop up well before the actual Y2.1k date switchover.

Of course, one could point out that all we're doing is postponing problems for 8000 years this way, but I guess you have to draw the line somewhere.

However, I agree with the other sentiments expressed here – error messages should be user friendly. Perhaps if a user really did try to enter a date in the past the software could give the message "the date you entered has already past – please enter a date in the future."

48. ### x said,

August 3, 2008 @ 12:17 pm

What everyone has missed so far is there doesn't seem to be a comfortable way to state this error message! Consider: "Please select a date after today." Could be interpreted two ways, an absolute usability no-no. The intended meaning is one, but the other meaning is "Please come back tomorrow and try this form again."

"Please select a date later than today", to my mind, seems to suffer from both problems (ambiguity and nerd-ism).

"Please choose a different date. Today is too soon." is too long, and reads like a zen koan.

And so on. The programmer here went with the phrasing that seemed least awkward and got the message across.

49. ### Michiel van der Blonk said,

August 3, 2008 @ 10:46 pm

Just change the sentence structure, x, I see you are trying to keep the same format. There is no reason to write 'technical' to a user, not even in the structure of the sentence. And you can use more words, error messages don't have to be this short.

"We have not vehicles available for today, so please enter a later date. Also, please use the format DD/MM/YYYY, e.g. 07/08/2008 (for 7th of August 2008)".

50. ### rkillings said,

August 30, 2008 @ 3:03 pm

The compelling reason for four-digit year is that you want a single standard date format everywhere in the user interface. If elsewhere you may need to have the customer enter a past date (birth date, passport issue date, etc.), you don't present a different format for exclusively future dates.

51. ### Marcio Rocha Pereira said,

October 31, 2008 @ 12:29 am

Geez, so nerds (let's say specialists to maintain pride here) are unable to communicate because they are too deeply immersed in their own world-view to notice semantic differences. Seems correct, methinks, but that leads me into a parallel question: is communication possible AT ALL? Maybe we are all nerd about ourselves, such that we are in effect unable to communicate except in approximations. For example, i am so aware of the connotations that, say, "cup of water" have to me that i don't even accept the possibility that those connotations are subjective. From that my idea extends that we never actually do achieve perfect communication. For example, there is no cup of water as a semantic unit anywhere, and despite the fact that i have been conditioned to bring you water if you say "can you please get me a cup of water?" there is not a meanings-field where "cup" or "water" or "to bring" can be really disambiguated, so that in the end there is no sharing of meaning whatsoever but only language acts without a specific substance? Does that even make sense? And if communication is impossible but this does not render language useless (as would be a very, very, very naive reading of my hypothesis), where does that lead us? Can we develop protocols to negotiate unspeakable concepts? Just wondering…

52. ### Christian R. Conrad said,

June 22, 2009 @ 9:42 am

Mark said:

"Who said anything about sort -n? It's improbable that a programmer would want to perform a purely textual comparison on the entries."

Not really all that improbable: That's exactly how it used to be done on mainframes programmed in COBOL, where it seems numbers where more or less interchangeable with strings (at least string-format-specifiers ["masks"] were an integral and necessary part of the specification of numeric variables, as I understand it).

Which is exactly why the international standard date format was developed, and why it was specified as "YYYYMMDD" (or "YYMMDD" [though that is of course deprecated nowadays], or "YYYY-MM-DD"): So it would sort chronologically correctly, even when stored and sorted as strings.

Also see the comment by "dr pepper" above:

I always use yymmdd or yy-mm-dd format. As for Excel, no prob, just make that column text.

Again, "Dr P" is doing exactly what the old MF COBOL coders did: Making the text-format representation and the actual numeric content behave equivalently. (Rear Admiral Grace Hopper must be cheering, somewhere.)

(Yes, BTW, there is indeed such a thing as an International Standard Date Format. Never mind that Sweden is the absolutely only place in the world that actually *uses* it; it's still an International Standard…)

53. ### Peter Taylor said,

September 2, 2009 @ 9:55 am

"Only for the utter benighted nerd is the greater-than relation, as used in integer mathematics, defined on dates."

Any real nerd knows that the greater-than relation defined on dates is different to the greater-than relation defined on integers. In fact each poset has its own greater-than relation.

PS The use of "poset" there wasn't nerdview but brevity. Look it up on Wikipedia if you want to know what it means.

54. ### Michael Ellis said,

August 4, 2011 @ 3:37 pm

Sounds to me as if they had the wrong time of specialist designing the user interface. If they had employed User Experience Design (UXD) specialists (also nerds?) the problem should not have occurred.

55. ### Michael Ellis said,

August 6, 2011 @ 3:18 am

That should read "type of specialist".

56. ### Elonkareon said,

September 28, 2014 @ 4:03 pm