The Daily WTF: Curious Perversions in Information Technology

ParkinT · 2006-01-04 Reply Admin

Anonymous:
It's so simple. The numbers: 0, 1, 2, 4 are all powers of 2, so they correspond to different bits in the machine word. Assuming that int has 32-bits on the target architecture, we have a 32-valued logic system. It is intuitionistic as well (no law of excluded middle), as -1 tags all bits!

Now there's a 'holisitic' approach to the problem. One which would make Douglas Adams proud!

------------------------------

If debugging is the process of removing errors,

programming must be the process of inserting them

ParkinT · 2006-01-04 Reply Admin

This is similar to the choice I give many of my clients (particularly those who are impatiently pushing for shorter deadlines):

You can have it:

. Cheap

. Fast

. Right

Choose any two.

sao · 2006-01-04 Reply Admin

Ytram:
Anonymous:
This is hardly a wtf, just bad engineering.

What the heck do you think a WTF is?

a wtf occurs when good engineering goes bad. not just bad engineering....

sammybaby · 2006-01-04 Reply Admin

Alex Papadimoulis:

> SELECT * FROM [YN_Lookups]
 +----+---------------+
 | ID | Text          |
 +----+---------------+
 | 1  | Yes           |
 | 2  | No            |
 | 3  | Not Specified |
 | 4  | NULL          |
 | 5  | Other         |
 | 7  | Unknown       |
 +----+---------------+

I think they're missing some values.

8. Maybe
9. Could be.
10. Answer hazy. Try again later.
11. Mu.
12. Cmdr Taco.

By the way: what do you suppose happened to 6?

Rank Amateur · 2006-01-04 Reply Admin

Reading through the comments, it's becoming clear that the programmer had remarkable forsight to use INT. It won't be long before one needs to represent over 256 values.

NULL - ?<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o:p></o:p>

0 - Zero<o:p></o:p>

1 - Yes<o:p></o:p>

2 - No<o:p></o:p>

3 - Not Specified<o:p></o:p>

4 - NULL<o:p></o:p>

5 - NUL<o:p></o:p>

6 - There is NO Value #6<o:p></o:p>

7 - Blank<o:p></o:p>

8 - Empty<o:p></o:p>

9 - Nothing<o:p></o:p>

10 - Nil<o:p></o:p>

11 - Zip<o:p></o:p>

12 - Unassigned<o:p></o:p>

13 - Indeterminate<o:p></o:p>

14 - Unknown<o:p></o:p>

15 - Unknowable<o:p></o:p>

16 - Beyond human comprehension<o:p></o:p>

17 - Maybe<o:p></o:p>

18 - Partially<o:p></o:p>

19 - Sometimes<o:p></o:p>

20 - Yes and No<o:p></o:p>

21 - Says “no” but means “yes”<o:p></o:p>

22 - Quantum mechanical<o:p></o:p>

23 - Missing<o:p></o:p>

24 - Lost<o:p></o:p>

25 - File not found<o:p></o:p>

26 - No Carrier<o:p></o:p>

27 - I forgot<o:p></o:p>

28 - Didn't get the email<o:p></o:p>

29 - The dog ate it<o:p></o:p>

30 - Invalid value entered<o:p></o:p>

31 - Unrecognized value entered<o:p></o:p>

32 - Blatantly absurd value entered<o:p></o:p>

33 - User had head up his/her ass<o:p></o:p>

34 - Non-value<o:p></o:p>

35 - Imaginary value<o:p></o:p>

36 - Drug-induced hallucination<o:p></o:p>

37 - It’s a value, Captain, but not as we know it<o:p></o:p>

38 - This is not the value you are looking for<o:p></o:p>

39 - Value corrupted<o:p></o:p>

40 - Value corrupted absolutely<o:p></o:p>

41 - Untrustworthy value <o:p></o:p>

42 - Downright evil value<o:p></o:p>

43 - Worthless value<o:p></o:p>

44 - Invaluable value<o:p></o:p>

45 - Family value<o:p></o:p>

46 - This isn’t really a value<o:p></o:p>

47 - This isn’t really a field<o:p></o:p>

48 - This is not my beautiful house<o:p></o:p>

49 - This is not my beautiful wife<o:p></o:p>

50 - Error 50<o:p></o:p>

51 - Error 51<o:p></o:p>

52 - Etc.<o:p></o:p>

-1 - Not otherwise specified<o:p></o:p>

--Rank

Ytram · 2006-01-04 Reply Admin

Somebodys name I cant remember:
By the way: what do you suppose happened to 6?

It was probably a status so horrible they decided to remove it from their database. Something like "Budget Under Review" or "Flowers in May". So they deleted it, and didn't reseed their identity.

2006-01-04 Reply Admin

GoatCheez:

not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

0 ----> 0000 0000 32 ----> 0001 0000

Just one bit of difference.

2006-01-04 Reply Admin

GoatCheez:

not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

0 ----> 0000 0000 32 ----> 0001 0000

Just one bit of difference.

GoatCheez · 2006-01-04 Reply Admin

GoatCheez:
Anonymous:
masklinn:
hash:
If 4 is NULL, then NULL can't be NULL, but by definition NULL is NULL... I'm confused.

Beware though, "NULL" is not NULL...

This reminds me of my favorite IRC quote:

Person A: You can't tell me that there's one bit of difference between a space and a null.
Person B: That's precisely the difference between a space and a null.

not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

Yeah... was thinking one bit of difference must mean 00000000 to 00000001... Didn't really occur to me that it could be something other than the least significant bit...

bit · 2006-01-04 Reply Admin

GoatCheez:
not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

00 ----> 0000 0000 32 ----> 0010 0000

Just one bit of difference.

2006-01-04 Reply Admin

Anonymous:
GoatCheez:

not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

0 ----> 0000 0000 32 ----> 0001 0000

Just one bit of difference.

ummm, your off by 1

0010 0000

2006-01-04 Reply Admin

3rd times the charm!!

Too bad your quotes don't work for you :(

2006-01-04 Reply Admin

Alexis de Torquemada:
GoatCheez:
not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

ASCII 0 is NUL, not "null". Four letter words were too expensive back then. Come to think about it - four letter words are still quite expensive!

Sorry, I'm a little drunk and seemed to have slurred my typing. I stand corrected. Cheers! [B]

bit · 2006-01-04 Reply Admin

Anonymous:
3rd times the charm!!

Too bad your quotes don't work for you :(

I kept getting errors, but the answers got posted anyway. Even the wrong ones. :)

Did quoting work this time? Lessee.

2006-01-04 Reply Admin

Allow me to play Devil's Advocate, and claim this is actually an amazingly well modelled piece of fuzzy visibility that tells me everything I need to know about the input procedure.

Let's assume we closely modelling a boolean field that may, or may not, have been specified and if specified, may or may not have been specified in a parsable format.

Yes and No are obvious well-parsed cases and the only ones we can really reliably test for.

A case where the field is absent in the input is also modelled -- as Not Specified.

However, what if the parser can't parse the input?

Well, null tells us that the parser was invoked, but there was no data available -- somebody specified the field, but left it blank.

Other informs us we received a declared exception to the boolean case -- "Sometimes" or "Maybe" or some other case that's linguistically related to verity, but has no place in our system.

And Unknown...well, some joker saw the "Is Reportable" field and decided to enter his favorite goalie's name.

Given a reference to this lookup table, I can now report on how many records are true or false...and how many failed verification and why. Not clean, I admit, but powerful.

As for the null and -1 values? Sounds like a braindead contractor underestimating the power of our fuzzy boolean input field!

bit · 2006-01-04 Reply Admin

bit:
Anonymous:
3rd times the charm!!

Too bad your quotes don't work for you :(

I kept getting errors, but the answers got posted anyway. Even the wrong ones. :)

Did quoting work this time? Lessee.

Why oh God? Why?

2006-01-04 Reply Admin

Pardon my alignment, but I hope it's 8 bytes. 64 bit/8 bits per byte...

Stoffel · 2006-01-04 Reply Admin

Not to beat a dead horse (too late) but the difference between null and space in ASCII is 1 if you're referring to the Hamming Distance.

(Nobody had mentioned it by name yet.)

hotzenplotz · 2006-01-04 Reply Admin

Anonymous:
masklinn:
hash:
If 4 is NULL, then NULL can't be NULL, but by definition NULL is NULL... I'm confused.

Beware though, "NULL" is not NULL...

This reminds me of my favorite IRC quote:

Person A: You can't tell me that there's one bit of difference between a space and a null.
Person B: That's precisely the difference between a space and a null.

To all those who didn't get it: Zero (no matter how many bits or whether you mean binary zero or ASCII zero) is not NULL. However, many database systems use "is-NULL" bitfields to mark columns as NULL, so the difference between NULL and non-NULL is always the corresponding bit in this bitfield. Depending on the database the size of the row may shrink too, because some database may not allocate space for fields that are NULL. Others might still allocate the space, because when a (fixed-size) field starts on the same offset for every row, then some queries can be done faster, especially when a table-scan, sort or something similar is necessary.

Databases are fun.

masklinn · 2006-01-04 Reply Admin

OneFactor:
The joins of inner earth 

Three keys (-1, 0, null) for the left-outer-Elves under the sky
Seven for the YN-Dwarf-lookups in their text "unknown"
Null four Dummy Row, doomed to lie
One for the Table Cases on his tps_reportable throne

No Field to rule them all, One Other Query to find them,
One Join Not Specified to bring them all and in the darkness bind them....

Sir, I bow to your greatness, you got me rolling on the floor in real life

KraGiE · 2006-01-04 Reply Admin

0 = no
1 = yes
Null = Unknown.

what's the big deal here? Some one just over engineered it. lol

hotzenplotz · 2006-01-04 Reply Admin

p.s.: the ASCII 'NUL' also is not NULL -- e.g. I can store an ASCII 'NUL' into a char or varchar field. That field would NOT be NULL! And if you think you're smart, just try it.

2006-01-04 Reply Admin

My nomination for #6:

"The truth! You don't want to know the truth!"

2006-01-04 Reply Admin

Also, lots of databases use empty strings to store NULL's and a single space to store empty strings, by trimming trailing spaces whenever the value is retrieved. And of course a string is usually terminated by a 0 byte.

Javelin · 2006-01-04 Reply Admin

Anonymous:
GoatCheez:
Anonymous:
This reminds me of my favorite IRC quote:

Person A: You can't tell me that there's one bit of difference between a space and a null.
Person B: That's precisely the difference between a space and a null.

not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

There are 10 types of people in this world. Those that understand binary jokes, and those who don't.

Actually, there are 10 types of people:

Those who don't understand binary jokes
Those who understand binary jokes, and think they're clever, because they always assume that any numeric string containing only 0s and 1s must be a base-two representation
Those who understand binary jokes, but think they're flawed, because they realize that a numeric string containing digits from 0 through k could be in any base as long as it's greater than k

Membership in the latter 10 types also corresponds pretty closely to these 10 types:

Those who think the "1 + 1 = 3, for very large values of 1" joke is clever
Those who get the whole "treating a numeric string as a variable" joke, and do think that's clever, but are bothered by the "very large values" part. (No: for a very large value of 1, like, say, a million, the expression equals two million, not 3 (unless we're also using a very large value of 3); it only equals three for one specific value of 1.)

I guess the last type could be further broken down into 10 types, making a total of 10 types:

Those who are perfectly aware of the issues I'm complaining about, but think the jokes are clever anyway (and that I'm being annoyingly pedantic)
Those who understand that nerd jokes only work when stated correctly (and that pedantry can itself be one of the highest forms of humor)

Oh yeah -- self-introduction: been reading for a few months and posted anonymously a couple times; registered a few weeks ago, but didn't actually post until now. Is this where I'm supposed to make my obligatory complaint about the crappy forum software and its lack of a preview button?

2006-01-04 Reply Admin

Rank Amateur:
(snipped excellent list)

Great list, with plenty of room for expansion! I'm coopting that for my "InsightfulErrorMessages" module in my next Access project! But I have a couple observations. The value for NULL would be:

NULL - "This value intentionally left blank."

and as for 42,

42 - Downright evil value

How can this possibly be so?

Stoffel:
(Hamming Distance, NIST DADS site)

That's a great site Stoffel!

ammoQ:
48 (ascii code for "0") = 001100000 (also only one bit different)

You got an extra significant zero there buddy, that's a reverse quote (#96). But isn't that two bits different from a NUL?

Michael Bolton:
No, it's just a coincidence!

<o:p></o:p>

Actually this whole table is a WTF! Why is there even a question? You always have to do your TPS Reports!! Didn't you get the memo?

2006-01-04 Reply Admin

A bit overdesigned but not especially stupid. If the answer is not a boolean he could have written the text un the Cases table but this version is more secured because of the verification on the FK.
The only two things I don't understand is WTF there is a "NULL" line in the second table. It is a doubled bloom with the "Not Specified" one, and WTF he doesn't enforce the foreign key, since it cause funny things to happend (like the NULL and -1 line) .
I guess this is a lame trick to store an other information in this field, maybe. In conclusion the idea is not so bad but realisation sucks.

ammoQ · 2006-01-04 Reply Admin

Anonymous:

ammoQ:
48 (ascii code for "0") = 001100000 (also only one bit different)

You got an extra significant zero there buddy, that's a reverse quote (#96).

right... but in this 9-bit-per byte representation, the rightmost bit is the parity bit and therefore not significant (j/k)

But isn't that two bits different from a NUL?

The original quote was:

Person A: You can't tell me that there's one bit of difference between a space and a null.
Person B: That's precisely the difference between a space and a null.

</div></BLOCKQUOTE> 

coincidentally, "Null" is the German word for Zero.

2006-01-04 Reply Admin

ParkinT:
This is similar to the choice I give many of my clients (particularly those who are impatiently pushing for shorter deadlines):

You can have it:

 . Cheap

 . Fast

 . Right

 Choose any two.

I prefer to give them : Expensive. Unhurried. Wrong. Exclude two.

2006-01-04 Reply Admin

let me guess... the contract was for a government agency?

Albatross · 2006-01-04 Reply Admin

And don't forget

| 12 | Yuri |

(http://www.thedailywtf.com/forums/44872/ShowPost.aspx)

rbriem · 2006-01-04 Reply Admin

Otto:

t-bone:

could someone explain the difference between "not specified", NULL and "unknown"?

"Unknown" clearly implies that the value which should be in the field is not known by any mortal agency. While "not specified" implies that the value is known, it is simply not given here, possibly because the user could not possibly comprehend it. The fact that there is a seperate "NULL" from an actual SQL NULL value shows that it is a special sort of "NULL", which implies both that there is no such value and that your query for it is obviously deluded in the extreme. Such a value only showcases the users own ignorance for asking such a silly question in the first place.

These are three easily distinguishable cases, as is obvious to anybody experienced in bureaucracy.

Excellent.

And that's what you get for arguing WTFery with a Sicilian.

(I've gotta watch that again sometime soon)

olddog · 2006-01-04 Reply Admin

I'm betting there's more than one TPS Report... and the field 'Is_TPS_Reportable' is really only *part* of a secret formula that when compared against the lookup, reveals the true business rule of TPS reports.

Also too coincidental - the 'Days Open' field could be resolved to a day-of-the week, which could point to one of options in the lookup table.

Besides, you can always add three to the 'Is_TPS_Reportable' value to fall safely with the range of lookup values, ( albeit you'll never get to 'Yes', probably a ruse ).

There's more here than meets the eye... this smacks of intrigue and secret agents.

rbriem · 2006-01-04 Reply Admin

sao:
Ytram:
Anonymous:
This is hardly a wtf, just bad engineering.

What the heck do you think a WTF is?

a wtf occurs when good engineering goes bad. not just bad engineering....

or when bad engineering goes worse ...

teddlesruss · 2006-01-04 Reply Admin

Anonymous:
GoatCheez:
Anonymous:
masklinn:
hash:
If 4 is NULL, then NULL can't be NULL, but by definition NULL is NULL... I'm confused.

Beware though, "NULL" is not NULL...

This reminds me of my favorite IRC quote:

Person A: You can't tell me that there's one bit of difference between a space and a null.
Person B: That's precisely the difference between a space and a null.

not getting the 1 bit difference thing... in ascii at least (what i thought it was refering to) 0 is null and 32 is space... not a bit, but like a bit and then some. technically the only difference is the one bit that is turned on at that spot...i'm probably thinking too hard on this. anyone care to help me out?

There are 10 types of people in this world. Those that understand binary jokes, and those who don't.

I thought that was "There 11 kinds of people in the world, those that know how to count in binary and those that don't."

olddog · 2006-01-04 Reply Admin

Yup, This could work... Let's first resolve the Days Open value to a day-of-the-week. Let's consider the -1 as a shift between date systems. Some systems starting Sunday, others Monday. With that, Monday(1) is always resolved to Sunday(0). So.. Is_TPS_Reportable will return a true TPS report *rule* when 1 is added at the end.

That could gives us the TPS report business rule:

Monday(1): always report.
Tuesday(2): no report.
Wednesday(3): one of the midweek reports (TBA).
Thursday(4): no report.
Friday(5): always the other report.
Saturday(6): never report. ( nobody does TPS on Saturday)
Sunday(7): waiting for direction from corporate.

chrismcb · 2006-01-04 Reply Admin

hotzenplotz:
p.s.: the ASCII 'NUL' also is not NULL -- e.g. I can store an ASCII 'NUL' into a char or varchar field. That field would NOT be NULL! And if you think you're smart, just try it.

I have a donut hole

Rumor has it quotes work in IE...

EvanED · 2006-01-04 Reply Admin

Otto:
t-bone:

could someone explain the difference between "not specified", NULL and "unknown"?

"Unknown" clearly implies that the value which should be in the field is not known by any mortal agency. While "not specified" implies that the value is known, it is simply not given here, possibly because the user could not possibly comprehend it.

It's not a matter of comprehension, because if you can't comprehend something, it can hardly be said that you know it and, hence, it would be unknown, so that falls into the previous category.

The "not specified" to me clearly means that while the value is known, it's being kept secret. The reasons why this would be is numerous, and I would suggest that the designer refactor the design to include another column in the table (or whatever the "technical" DB term is) with another foregin key that indicates why it's being kept secret, be it national security reasons, being held for blackmail, etc.

(BTW, in addition to the more common WTFs mentioned about this forum software, does it strike anyone as odd that I can post with the name EvanED even when I'm not logged in (purely out of laziness) and despite being registered for a while?)

olddog · 2006-01-04 Reply Admin

EvanED:
Otto:

t-bone:

could someone explain the difference between "not specified", NULL and "unknown"?

"Unknown" clearly implies that the value which should be in the field is not known by any mortal agency. While "not specified" implies that the value is known, it is simply not given here, possibly because the user could not possibly comprehend it.

It's not a matter of comprehension, because if you can't comprehend something, it can hardly be said that you know it and, hence, it would be unknown, so that falls into the previous category.

The "not specified" to me clearly means that while the value is known, it's being kept secret. The reasons why this would be is numerous, and I would suggest that the designer refactor the design to include another column in the table (or whatever the "technical" DB term is) with another foregin key that indicates why it's being kept secret, be it national security reasons, being held for blackmail, etc.

Not Specified - implies an option *is* available.
Unknown - implies an option *may* be available.
Null- implies *no* options are available.

olddog · 2006-01-04 Reply Admin

or this:

Not Specified - implies an option *is* available.
Unknown - implies an option *may* be available.
False - implies *no* options are available.
Null - make no implication.

2006-01-04 Reply Admin

| 6 | Rows that magically disappear when inner-joined with YN_Lookups |

2006-01-04 Reply Admin

OK, this is F'd up, but its not really difficult - just take the rows where the value is 1 ('Yes") and ignore everything else. Sometimes a little judgement goes a long way, Paul! ;-)

olddog · 2006-01-04 Reply Admin

Anonymous:
OK, this is F'd up, but its not really difficult - just take the rows where the value is 1 ('Yes") and ignore everything else. Sometimes a little judgement goes a long way, Paul! ;-)

Don't the other values make you the least bit curious?

2006-01-04 Reply Admin

To me, NULL is No.
i.e. . . . '\0' is the the same as 0
Hasn't failed me yet, in 20 years of C

"Not Specified" -- "Other" -- "Unknown" would all map to the same thing.

Maybe there should have been
| 8 | Remove from all further consideration

2006-01-04 Reply Admin

Anonymous:
As long as he remembers the new cover sheets for the TPS Reports, he'll be fine.

LOL

And a taxi chit for the printer, too

2006-01-05 Reply Admin

Old Printer's Joke.

Bin There. . . not BinLada

2006-01-05 Reply Admin

Yessssss Bit is on the bit.

three's a charm!

2006-01-05 Reply Admin

Oh Yeah? Tell it to the register!

2006-01-05 Reply Admin

Produce a report that will list the Case Number, Officer Name, and Days Open for all remote offices where the case is TPS Reportable. <?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o:p></o:p>

This is just a thought that I came up within 30 seconds when I first looked at the problem.<o:p></o:p>

SELECT * From OfficeDetails where Is_TPS_Reportable = 1; or<o:p></o:p>

SELECT * From OfficeDetails where Is_TPS_Reportable = "yes"; <o:p></o:p>

Depends on value that is actually stored there. Who cares if someone placed Null in text field in some other part of the data abase. If you think there is a problem, mention it to them but I fail to see how this is complicated. <o:p></o:p>

2006-01-05 Reply Admin

Anonymous:
Oh Yeah? Tell it to the register!

This software truly is a WTF

Yessssssss! Noooooooo!

Leave a comment on “Yessssssss! Noooooooo!”