info
Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
The XY Problem
38 views
Skip to first unread message
Kenny McCormack
Jun 5, 2014, 12:34:49 PM6/5/14
to
Just a quick informational post about "The XY Problem".
According to: http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem
The XY problem is asking about your attempted solution rather than your
actual problem.
That is, you are trying to solve problem X, and you think solution Y would
work, but instead of asking about X when you run into trouble, you ask
about Y.
Editorial note: I think in some of our postings, one or more of us have
exchanged the 'X' and the 'Y'.
--
Just for a change of pace, this sig is *not* an obscure reference to
comp.lang.c...
According to: http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem
The XY problem is asking about your attempted solution rather than your
actual problem.
That is, you are trying to solve problem X, and you think solution Y would
work, but instead of asking about X when you run into trouble, you ask
about Y.
Editorial note: I think in some of our postings, one or more of us have
exchanged the 'X' and the 'Y'.
--
Just for a change of pace, this sig is *not* an obscure reference to
comp.lang.c...
A. Mehoela
Jun 5, 2014, 12:42:08 PM6/5/14
to
Kenny McCormack wrote:
>
> Editorial note: I think in some of our postings, one or more of us have
> exchanged the 'X' and the 'Y'.
>
X?
>
> Editorial note: I think in some of our postings, one or more of us have
> exchanged the 'X' and the 'Y'.
>
Kenny McCormack
Jun 5, 2014, 12:47:30 PM6/5/14
to
In article <53909de0$0$16349$8a7a...@news4.usenet4u.nl>,
'X' is the underlying problem, 'Y' is the proposed solution/tool.
"I'm hungry" is an X. "How do I boil eggs?" is a Y.
--
Modern Conservative: Someone who can take time out
from demanding more flag burning laws, more abortion
laws, more drug laws, more obscenity laws, and more
police authority to make warrantless arrests to remind
us that we need to "get the government off our backs".
"I'm hungry" is an X. "How do I boil eggs?" is a Y.
--
Modern Conservative: Someone who can take time out
from demanding more flag burning laws, more abortion
laws, more drug laws, more obscenity laws, and more
police authority to make warrantless arrests to remind
us that we need to "get the government off our backs".
Frank P. Westlake
Jun 6, 2014, 10:59:35 AM6/6/14
to
On 06/05/2014 09:34 AM, Kenny McCormack wrote:
> The XY problem is asking about your attempted solution rather than your
> actual problem.
I don't think that definition fits very well. Sometimes the problem is
> The XY problem is asking about your attempted solution rather than your
> actual problem.
to solve Y in order to get X. If Y cannot be solved then X is not
desired. The problem then is X->Y, not X.
Frank
Frank P. Westlake
Jun 6, 2014, 11:03:25 AM6/6/14
to
On 06/05/2014 09:47 AM, Kenny McCormack wrote:
> "I'm hungry" is an X. "How do I boil eggs?" is a Y.
And the X->Y problem is "I was to eat eggs I have boiled; how do I boil
> "I'm hungry" is an X. "How do I boil eggs?" is a Y.
eggs".
Frank
Barry Margolin
Jun 6, 2014, 12:45:31 PM6/6/14
to
In article <lmsl0p$tvt$1...@news.albasani.net>,
That's different.
When we call something an "XY problem", it refers to a case where the
asker implicitly assumes that the only way to solve X is via Y, but that
assumption is incorrect. If we knew that their ultimate goal were
actually X, we could give better solutions that don't require Y.
So if someone has a craving for eggs, asking where eggs can be bought is
not an XY problem. Or if they're hungry and they have eggs on hand,
asking how to cook them is not an XY problem.
But if someone is just hungry in general, asking specifically about eggs
would be an XY problem -- there are other, possibly easier, ways to
satisfy hunger than eating eggs. Asking the wrong question unnecessarily
obfuscates the problem, making it harder to give good answers.
http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem
--
Barry Margolin, bar...@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
When we call something an "XY problem", it refers to a case where the
asker implicitly assumes that the only way to solve X is via Y, but that
assumption is incorrect. If we knew that their ultimate goal were
actually X, we could give better solutions that don't require Y.
So if someone has a craving for eggs, asking where eggs can be bought is
not an XY problem. Or if they're hungry and they have eggs on hand,
asking how to cook them is not an XY problem.
But if someone is just hungry in general, asking specifically about eggs
would be an XY problem -- there are other, possibly easier, ways to
satisfy hunger than eating eggs. Asking the wrong question unnecessarily
obfuscates the problem, making it harder to give good answers.
http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem
--
Barry Margolin, bar...@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
Frank P. Westlake
Jun 6, 2014, 1:52:21 PM6/6/14
to
On 06/06/2014 09:45 AM, Barry Margolin wrote:
> When we call something an "XY problem", it refers to a case where the
> asker implicitly assumes that the only way to solve X is via Y, but that
> assumption is incorrect. If we knew that their ultimate goal were
> actually X, we could give better solutions that don't require Y.
Here's a better example. A problem as I state it here is "how can I
> When we call something an "XY problem", it refers to a case where the
> asker implicitly assumes that the only way to solve X is via Y, but that
> assumption is incorrect. If we knew that their ultimate goal were
> actually X, we could give better solutions that don't require Y.
manually enter ASCII-248 into Slackware's Konsole".
Someone might perceive it to be the XY problem:
X= Enter non-keyboard character into a terminal.
Y= Manual keyboard entry.
And that person might insist on only giving me an alternate solution --
a Z (paste it in from some other program) or a W (import it from a
file). But in this case I want only the manual entry, and if I can't
have that then I'm not interested in entering the ASCII-248. This is
what I labelled the (X->Y) problem -- a single desired solution for a
problem.
Frank
Kaz Kylheku
Jun 6, 2014, 2:03:56 PM6/6/14
to
["Followup-To:" header set to comp.unix.shell.]
"enter ASCII-248".
> Someone might perceive it to be the XY problem:
>
> X= Enter non-keyboard character into a terminal.
> Y= Manual keyboard entry.
In anticipation of this, you should make it clear that "I just want to
enter code 248 into the console for the hell of it, just to verify that
entering such a thing directly from the keyboard can be done; I do not wish to
be informed about any other way to get the same effect, like cut and paste.
I already know these obvious workarounds.
> And that person might insist on only giving me an alternate solution --
> a Z (paste it in from some other program) or a W (import it from a
> file). But in this case I want only the manual entry, and if I can't
> have that then I'm not interested in entering the ASCII-248. This is
> what I labelled the (X->Y) problem -- a single desired solution for a
> problem.
This is just an "X" problem. I want to do X, and it's for nothing else;
the chain of requirements begins and ends at X.
On 2014-06-06, Frank P. Westlake <frank.w...@gmail.com> wrote:
> On 06/06/2014 09:45 AM, Barry Margolin wrote:
>> When we call something an "XY problem", it refers to a case where the
>> asker implicitly assumes that the only way to solve X is via Y, but that
>> assumption is incorrect. If we knew that their ultimate goal were
>> actually X, we could give better solutions that don't require Y.
>
> Here's a better example. A problem as I state it here is "how can I
> manually enter ASCII-248 into Slackware's Konsole".
So far, the reason is not give why you need to do this: what is X. The Y is
> On 06/06/2014 09:45 AM, Barry Margolin wrote:
>> When we call something an "XY problem", it refers to a case where the
>> asker implicitly assumes that the only way to solve X is via Y, but that
>> assumption is incorrect. If we knew that their ultimate goal were
>> actually X, we could give better solutions that don't require Y.
>
> Here's a better example. A problem as I state it here is "how can I
> manually enter ASCII-248 into Slackware's Konsole".
"enter ASCII-248".
> Someone might perceive it to be the XY problem:
>
> X= Enter non-keyboard character into a terminal.
> Y= Manual keyboard entry.
enter code 248 into the console for the hell of it, just to verify that
entering such a thing directly from the keyboard can be done; I do not wish to
be informed about any other way to get the same effect, like cut and paste.
I already know these obvious workarounds.
> And that person might insist on only giving me an alternate solution --
> a Z (paste it in from some other program) or a W (import it from a
> file). But in this case I want only the manual entry, and if I can't
> have that then I'm not interested in entering the ASCII-248. This is
> what I labelled the (X->Y) problem -- a single desired solution for a
> problem.
the chain of requirements begins and ends at X.
frank.w...@gmail.com
Jun 6, 2014, 2:11:27 PM6/6/14
to
From "Frank P. Westlake":
hard.
Frank
>Here's a better example. A problem as I state it here is
>"how can I
>manually enter ASCII-248 into Slackware's Konsole".
I do know that 248 is beyond ASCII but old habits die
>"how can I
>manually enter ASCII-248 into Slackware's Konsole".
hard.
Frank
frank.w...@gmail.com
Jun 6, 2014, 2:21:39 PM6/6/14
to
From Kaz Kylheku:
solution clearly enough and I need not already be aware
of other possible methods of character entry. It does no
harm to suggest other methods but it should not be
assumed that my statement is deficient because I am
asking for a specific solution.
My apologies to the two groups for getting involved in
this cross-posted thread; cross-posts normally go to the
bit bucket but I was on another computer which I haven't
set that filter yet.
Frank
>> Here's a better example. A problem as I state it here
>is "how can I
>> manually enter ASCII-248 into Slackware's Konsole".
>So far, the reason is not give why you need to do this:
>what is X. The Y is
>"enter ASCII-248".
>> Someone might perceive it to be the XY problem:
>>
>> X= Enter non-keyboard character into a terminal.
>> Y= Manual keyboard entry.
>In anticipation of this, you should make it clear that
>"I just want to
>enter code 248 into the console for the hell of it, just
>to verify that
>entering such a thing directly from the keyboard can be
>done; I do not wish to
>be informed about any other way to get the same effect,
>like cut and paste.
>I already know these obvious workarounds.
No I need not anticipate that. I've stated my desired
>is "how can I
>> manually enter ASCII-248 into Slackware's Konsole".
>So far, the reason is not give why you need to do this:
>what is X. The Y is
>"enter ASCII-248".
>> Someone might perceive it to be the XY problem:
>>
>> X= Enter non-keyboard character into a terminal.
>> Y= Manual keyboard entry.
>In anticipation of this, you should make it clear that
>"I just want to
>enter code 248 into the console for the hell of it, just
>to verify that
>entering such a thing directly from the keyboard can be
>done; I do not wish to
>be informed about any other way to get the same effect,
>like cut and paste.
>I already know these obvious workarounds.
solution clearly enough and I need not already be aware
of other possible methods of character entry. It does no
harm to suggest other methods but it should not be
assumed that my statement is deficient because I am
asking for a specific solution.
My apologies to the two groups for getting involved in
this cross-posted thread; cross-posts normally go to the
bit bucket but I was on another computer which I haven't
set that filter yet.
Frank
Barry Margolin
Jun 6, 2014, 2:52:52 PM6/6/14
to
In article <14672779674$frank.w...@gmail.com>,
the question is based on a misunderstanding. If you know that there's a
good reason for the weird question, you can save everyone time and
confusion by explaining it.
frank.w...@gmail.com wrote:
> No I need not anticipate that. I've stated my desired
> solution clearly enough and I need not already be aware
> of other possible methods of character entry. It does no
> harm to suggest other methods but it should not be
> assumed that my statement is deficient because I am
> asking for a specific solution.
When people ask weird questions, experienced people can often tell that
> No I need not anticipate that. I've stated my desired
> solution clearly enough and I need not already be aware
> of other possible methods of character entry. It does no
> harm to suggest other methods but it should not be
> assumed that my statement is deficient because I am
> asking for a specific solution.
the question is based on a misunderstanding. If you know that there's a
good reason for the weird question, you can save everyone time and
confusion by explaining it.
Barry Margolin
Jun 6, 2014, 2:24:28 PM6/6/14
to
In article <lmsv4p$kfp$1...@news.albasani.net>,
In that case, the original question wasn't an XY problem.
Just as questioners may make incorrect assumptions that cause them to
ask an XY question, answers may incorrectly assume that the question is
an XY problem. No one is perfect.
When you think something may be an XY problem, a good way to deal with
it isn't to assume a particular overall goal, but to ASK -- "What is it
you're actually trying to accomplish with this?" When you get
clarification, you can then tell whether it was really an XY problem.
Just as questioners may make incorrect assumptions that cause them to
ask an XY question, answers may incorrectly assume that the question is
an XY problem. No one is perfect.
When you think something may be an XY problem, a good way to deal with
it isn't to assume a particular overall goal, but to ASK -- "What is it
you're actually trying to accomplish with this?" When you get
clarification, you can then tell whether it was really an XY problem.
frank.w...@gmail.com
Jun 6, 2014, 5:35:04 PM6/6/14
to
From Barry Margolin:
real difference is the desire of the person who has the
problem. In one case the person is seeking a specific
solution, perhaps thinking it to be the only solution,
but would accept another solution if it were known; in
the other case the person is seeking a specific
solution, perhaps thinking it to be the only solution,
but would not accept another solution if it were known.
Both of these cases fit the definition of 'XY Problem'.
Frank
>> Someone might perceive it to be the XY problem:
>In that case, the original question wasn't an XY
>problem.
Agreed; but it is often perceived as one; and the only
>problem.
real difference is the desire of the person who has the
problem. In one case the person is seeking a specific
solution, perhaps thinking it to be the only solution,
but would accept another solution if it were known; in
the other case the person is seeking a specific
solution, perhaps thinking it to be the only solution,
but would not accept another solution if it were known.
Both of these cases fit the definition of 'XY Problem'.
Frank
frank.w...@gmail.com
Jun 6, 2014, 5:54:20 PM6/6/14
to
From Barry Margolin:
>When people ask weird questions...
"Wierd" isn't part of your XY Problem definition.
>...experienced people can
as I stated previously.
Frank
>When people ask weird questions...
"Wierd" isn't part of your XY Problem definition.
>...experienced people can
>often tell that
>the question is based on a misunderstanding.
"Can often tell" also means that they are often wrong,
>the question is based on a misunderstanding.
as I stated previously.
Frank
Barry Margolin
Jun 6, 2014, 6:16:40 PM6/6/14
to
In article <1467331b25a$frank.w...@gmail.com>,
> >...experienced people can
> >often tell that
> >the question is based on a misunderstanding.
>
> "Can often tell" also means that they are often wrong,
> as I stated previously.
I've been answering questions on Usenet for about 30 years. I think I've
gotten pretty good at this kind of thing. I'll bet I have at least 80%
accuracy.
frank.w...@gmail.com wrote:
> From Barry Margolin:
> >When people ask weird questions...
>
> "Wierd" isn't part of your XY Problem definition.
No, it's a heuristic for suspecting that the question is an XY problem.
> From Barry Margolin:
> >When people ask weird questions...
>
> "Wierd" isn't part of your XY Problem definition.
> >...experienced people can
> >often tell that
> >the question is based on a misunderstanding.
>
> "Can often tell" also means that they are often wrong,
> as I stated previously.
gotten pretty good at this kind of thing. I'll bet I have at least 80%
accuracy.
Kenny McCormack
Jun 6, 2014, 10:51:00 PM6/6/14
to
In article <barmar-C0C1AD....@news.eternal-september.org>,
Barry Margolin <bar...@alum.mit.edu> wrote:
...
The underlying point of this thread is that sometimes it really is f***in'
annoying when you ask a question and you really do need an answer to Y, and
you don't need people second guessing you. That's the 20% that rankles.
I had my own version of this, a couple of million years ago, when I needed
to know how to do a certain thing in Perl (not then nor now one of my
favorite languages) that revolved around having nothing other than the Perl
binary on the system - i.e., no support files.
Well, of course, the answers I got were all economist's answers: First you
assume a can opener. In this case, the answers were all "Install the
support files". I explained that I couldn't install the support files;
that was the whole point of the problem - how to do it w/o the support
files. Back and forth we went, with all the answerers very obnoxiously and
condescendingly explaining that I had to install the support files.
Eventually, I gave up and solved the problem on my own.
That story, which happened long before the term "XY Problem" was coined,
has always shaped my general view of "The XY problem".
--
Those on the right constantly remind us that America is not a
democracy; now they claim that Obama is a threat to democracy.
Barry Margolin <bar...@alum.mit.edu> wrote:
...
>I've been answering questions on Usenet for about 30 years. I think I've
>gotten pretty good at this kind of thing. I'll bet I have at least 80%
>accuracy.
Of course, that means you're wrong 20% of the time.
>gotten pretty good at this kind of thing. I'll bet I have at least 80%
>accuracy.
The underlying point of this thread is that sometimes it really is f***in'
annoying when you ask a question and you really do need an answer to Y, and
you don't need people second guessing you. That's the 20% that rankles.
I had my own version of this, a couple of million years ago, when I needed
to know how to do a certain thing in Perl (not then nor now one of my
favorite languages) that revolved around having nothing other than the Perl
binary on the system - i.e., no support files.
Well, of course, the answers I got were all economist's answers: First you
assume a can opener. In this case, the answers were all "Install the
support files". I explained that I couldn't install the support files;
that was the whole point of the problem - how to do it w/o the support
files. Back and forth we went, with all the answerers very obnoxiously and
condescendingly explaining that I had to install the support files.
Eventually, I gave up and solved the problem on my own.
That story, which happened long before the term "XY Problem" was coined,
has always shaped my general view of "The XY problem".
--
Those on the right constantly remind us that America is not a
democracy; now they claim that Obama is a threat to democracy.
Barry Margolin
Jun 7, 2014, 5:51:48 AM6/7/14
to
In article <lmtumk$ipf$1...@news.xmission.com>,
for free, I think I'm happy with my record.
>
> The underlying point of this thread is that sometimes it really is f***in'
> annoying when you ask a question and you really do need an answer to Y, and
> you don't need people second guessing you. That's the 20% that rankles.
And for the rest of us, the 80% where we waste lots of time trying to
solve the wrong problem rankles.
Since there are more answerers than askers for any particular question,
I think we win. Our 80% is more important than your 20%.
> That story, which happened long before the term "XY Problem" was coined,
> has always shaped my general view of "The XY problem".
The thing is, most questioners are not very expert. They don't know what
they're doing, and they have all sorts of misconceptions. So it
generally doesn't pay to give the questioners the benefit of the doubt
when Occam's Razor suggests that it's an XY problem.
Occasionally there are exceptions. Unfortunately, on the net, everyone
looks the same, unless they've been around a while and their name is
known. If you don't want us to assume the worst, you have to go to a
little more effort to make the question crystal clear.
There are no guarantees in life, and shit sometimes happens. But the
better your question, the more likely you are to get the kind of answer
you need.
gaz...@shell.xmission.com (Kenny McCormack) wrote:
> In article <barmar-C0C1AD....@news.eternal-september.org>,
> Barry Margolin <bar...@alum.mit.edu> wrote:
> ...
> >I've been answering questions on Usenet for about 30 years. I think I've
> >gotten pretty good at this kind of thing. I'll bet I have at least 80%
> >accuracy.
>
> Of course, that means you're wrong 20% of the time.
Better than weathermen, and they get paid big bucks. I'm doing this all
> In article <barmar-C0C1AD....@news.eternal-september.org>,
> Barry Margolin <bar...@alum.mit.edu> wrote:
> ...
> >I've been answering questions on Usenet for about 30 years. I think I've
> >gotten pretty good at this kind of thing. I'll bet I have at least 80%
> >accuracy.
>
> Of course, that means you're wrong 20% of the time.
for free, I think I'm happy with my record.
>
> The underlying point of this thread is that sometimes it really is f***in'
> annoying when you ask a question and you really do need an answer to Y, and
> you don't need people second guessing you. That's the 20% that rankles.
solve the wrong problem rankles.
Since there are more answerers than askers for any particular question,
I think we win. Our 80% is more important than your 20%.
> That story, which happened long before the term "XY Problem" was coined,
> has always shaped my general view of "The XY problem".
they're doing, and they have all sorts of misconceptions. So it
generally doesn't pay to give the questioners the benefit of the doubt
when Occam's Razor suggests that it's an XY problem.
Occasionally there are exceptions. Unfortunately, on the net, everyone
looks the same, unless they've been around a while and their name is
known. If you don't want us to assume the worst, you have to go to a
little more effort to make the question crystal clear.
There are no guarantees in life, and shit sometimes happens. But the
better your question, the more likely you are to get the kind of answer
you need.
Kenny McCormack
Jun 7, 2014, 8:46:01 AM6/7/14
to
In article <barmar-D06B3C....@news.eternal-september.org>,
thread, so clearly, I'm of the opinion that the XY problem is real.
I'm just illustrating the other side of it.
BTW, the "how do I send email?" thread that is going on now in c.u.s. can be
seen as another instance of my Perl problem. It is clear to everybody that
'mpack' is the right solution and that the OP's solution is a
wheel-reinvent at best (and at worst, a collection of bugs).
But, for whatever reason, 'mpack' is not a solution for the OP. As much as
we'd like to convince him to install it himself (which he almost certainly
could do), he is dead set against using it.
Much as installing the Perl support files was not an option for me (but for
different reasons; it was not a PHB problem).
--
They say compassion is a virtue, but I don't have the time!
- David Byrne -
Barry Margolin <bar...@alum.mit.edu> wrote:
>In article <lmtumk$ipf$1...@news.xmission.com>,
> gaz...@shell.xmission.com (Kenny McCormack) wrote:
>
>> In article <barmar-C0C1AD....@news.eternal-september.org>,
>> Barry Margolin <bar...@alum.mit.edu> wrote:
>> ...
>> >I've been answering questions on Usenet for about 30 years. I think I've
>> >gotten pretty good at this kind of thing. I'll bet I have at least 80%
>> >accuracy.
>>
>> Of course, that means you're wrong 20% of the time.
>
>Better than weathermen, and they get paid big bucks. I'm doing this all
>for free, I think I'm happy with my record.
I'm not disagreeing with you - remember, I was the one who started this
>In article <lmtumk$ipf$1...@news.xmission.com>,
> gaz...@shell.xmission.com (Kenny McCormack) wrote:
>
>> In article <barmar-C0C1AD....@news.eternal-september.org>,
>> Barry Margolin <bar...@alum.mit.edu> wrote:
>> ...
>> >I've been answering questions on Usenet for about 30 years. I think I've
>> >gotten pretty good at this kind of thing. I'll bet I have at least 80%
>> >accuracy.
>>
>> Of course, that means you're wrong 20% of the time.
>
>Better than weathermen, and they get paid big bucks. I'm doing this all
>for free, I think I'm happy with my record.
thread, so clearly, I'm of the opinion that the XY problem is real.
I'm just illustrating the other side of it.
BTW, the "how do I send email?" thread that is going on now in c.u.s. can be
seen as another instance of my Perl problem. It is clear to everybody that
'mpack' is the right solution and that the OP's solution is a
wheel-reinvent at best (and at worst, a collection of bugs).
But, for whatever reason, 'mpack' is not a solution for the OP. As much as
we'd like to convince him to install it himself (which he almost certainly
could do), he is dead set against using it.
Much as installing the Perl support files was not an option for me (but for
different reasons; it was not a PHB problem).
--
They say compassion is a virtue, but I don't have the time!
- David Byrne -
Casper H.S. Dik
Jun 8, 2014, 7:39:58 AM6/8/14
to
Barry Margolin <bar...@alum.mit.edu> writes:
>The thing is, most questioners are not very expert. They don't know what
>they're doing, and they have all sorts of misconceptions. So it
>generally doesn't pay to give the questioners the benefit of the doubt
>when Occam's Razor suggests that it's an XY problem.
Often you see that when a question is being formulated, they are
>The thing is, most questioners are not very expert. They don't know what
>they're doing, and they have all sorts of misconceptions. So it
>generally doesn't pay to give the questioners the benefit of the doubt
>when Occam's Razor suggests that it's an XY problem.
not telling you the problem they are facing; instead they believe
they have properly found the issue and want to get that fixed.
But in the end, the analysis wasn't correct and we try to figure
out what the actual problem was.
Casper
0 new messages