David NestorApr 05, 2005 at 10:31AM

I don't have a solution, but I can add some food for thought. Perhaps we create a pick two cliche because it's the simplest form besides pick one. Pick one implies a multiple choice, where each choice is exclusive. Once you introduce pick two, it becomes complex, allowing for any combination of the elements involved, except the combination of all elements. And the reason it's pick two instead of pick three out of four (healthy, fast, tasty, low carb) is that the latter could also be pick two out of four... and that makes it too complex. A good cliche has to be as simple as possible while making its point. So pick two is the simplest request allowing for a user-defined combination (instead of just yes or no) while refusing to grant all of the options.

Phew... that was a mouthful. I had too much coffee, forgive me.

cushie butterfieldApr 05, 2005 at 10:31AM

This is a trilemma . I'm not sure where the term originates but it is often used in political economy especially as the Mundell-Fleming trilemma about currency flows and the openness of an economy.

[I'm not convinced that all of your examples are true trilemmas, eg rice and beans are cheap, healthy and tasty]

cushie butterfieldApr 05, 2005 at 10:33AM

Wikipedia trilemma: http://en.wikipedia.org/wiki/Trilemma

ZelnoxApr 05, 2005 at 10:41AM

From project management, I learned that projects have three major areas of concern: time, budget, and features. The manager has to juggle all three.

Ryan AbramsApr 05, 2005 at 10:42AM

I've had a senior manager at my company use the "Fast, Cheap, and Good" example around me several times. He contends that you always pick fast and cheap, because money and time are definite values, but you can always go back and make it good. I completely disagree with this assesment.

The most interesting thing I find about these "trilemmas" is that we assume they are true in the first place. It seems to me that its very big corporate thinking. Most successful startups are successful because they worked fast, cheap, and good. If you take one of the three away, you end up with just another failed startup. Too long to market, ran out of capital, or a product no one wants.

The next time someone tells me to pick two, I'm going to pick all three. If they can't do it, I'll find someone who can.

Mobil'HommeApr 05, 2005 at 10:46AM

I'll venture a stab as to why three is a cognitively important number, which might explain why "three" shows up in lots of contexts (including the pick two scenario).

Three is the smallest number of things (ideas, objects, etc.) you need to form an abstraction. You need two things that are the same and one thing that is different. If you have two things that are different (say, a blue ball and a red brick) your mind treats them as particulars. If you have two things that are the same (say a blue ball and a blue hat), your mind still treats them as particulars. BUT, if you have all three--a blue hat, a blue ball, and a red brick--your mind can isolate the aspect on which the hat and the ball share a sameness. You need two things possessing the same trait in order to see that the trait is applicable to things generally, but you need a third thing to serve as foil in order to differentiate that trait (as opposed to all the other possible traits).

I'm cribbing copiously from this work which is brilliant.

Mobil'HommeApr 05, 2005 at 10:57AM

I should have stated about my example--distinguishing the color "blue"--that it would only apply to someone who doesn't already possess the abstraction "blue," i.e., someone really young. It's the same process for higher level abstractions, but an adult doesn't have to go through that process for "blue"--an adult already possesses that abstraction.

The colloquial form in which we are all familiar with this process is "one of of these things is not like the others" from Electric Company.

PeterApr 05, 2005 at 11:00AM

so if i want something good and cheap it will take a long time? i don't see how that combination works. wouldn't more time result in more cost?

Eloy AnzolaApr 05, 2005 at 11:04AM

Bueno, Bonito y Barato.
Good, Pretty and Cheap.

When it comes to professional projects/web developement, it is basically imposible to have all three. In any language or culture.

Cheap might not just include money.

It could be argued that a good poem is cheap in $ to produce. But it probably required some pain and effort.

In Ryan's example, a succesful startup does not come cheap. It usually comes at a very high price: Effort, dedication and sacrifice.

Jason ColemanApr 05, 2005 at 11:05AM

I have to disagree with several posters referring to these rules of thumb as "trilemas." By definition, a trilema requires a single choice out of three options. Here, we are talking about getting two out of three, and why just two out of three. I think this is a little memorable way of stating the different kinds of costs we pay (or businesses pay, since this seems to be a business adage). I don't really agree with all of Jason's examples, but they're great observations none-the-less.
As far as why three? Three is just a great number. Two represents a simple choice of this or that. Three adds complexity and multi-level decision making. I don't think anyone thought deeply on that when they first stated this rule, it just comes natural. It the same reason so many great jokes have three lines... two to set a pattern and the third to make a punch line. Science or mathematics may have an answer for this, but I'd say it's really no more complex than that.

Rob MientjesApr 05, 2005 at 11:07AM

I'd say it's like with the Venn diagram, but then without the three overlapping. The only perfect spots are where two circles overlap. Three overlapping circles seems like pretty much the biggest impossibility, at least in this case.

But don't ask much more, I don't even fully get it myself ;)

BrianApr 05, 2005 at 11:12AM

On the one hand.
On the other hand.
On the gripping hand.

theMikeApr 05, 2005 at 11:12AM

It seems to me that of the many elements that could possibly go into completing a project, you would have to encompass more than one, which would disclue 1 of 2. Also, the more choices one has, the longer it takes to actually get to the problem at hand. I think 2 of 3 works because it speaks to our need to choose, but doesn't confuse the brain too much with information, and let's us get into solving the problem at hand, which is all we as human beings want to do anyway.

Darius KazemiApr 05, 2005 at 11:16AM

This makes me think of graph theory--the branch of mathematics that deals with discrete elements and relations between these elements. In graph theory, we often start with a base case of three elements and the relations between these elements, simply because this is the first nontrivial case. The "pick two of three" construction is definitely a reflection of the simplest interesting construction in graph theory, which is three nodes in size.

For more on graph theory, see this free, complete online textbook.

R J KeefeApr 05, 2005 at 11:20AM

The answer to "Why Three" may well be discovered by neurologists. I think that we all experience a certain satisfaction with trios; as a writer, I find that a lineup of three descriptive adjectives has a conclusive power that pairs rarely reach.

"Fast, cheap, and out of control" - this movie title isn't an example of what you're talking about, is it. It's the "pick two" version of "Fast, cheap, and disciplined."

TedApr 05, 2005 at 11:20AM

Seems like two out of three is the smallest tradeoff that's interesting. One out of two is boring. One out of three doesn't satisfy. Two out of three allows the chooser to feel like s/he is getting something out of the tradeoff (not just 50/50).

RobApr 05, 2005 at 11:36AM

The Universal Rule of Thirds.
Avoiding symmetric choices (where no trade off exists, and no clean winning choice). A variation of the Golden Mean.

TonyApr 05, 2005 at 11:41AM

Seems to me that it's all about forcing a decision-making process that can be completed in a short timeframe (and therefore is more likely to be undertaken).

If you present someone with a list of 20 choices and have them pick 14, most people will probably just scan down the list and choose things that look good until they have reached 14 items. However, limiting the choices to 2 out of 3 allows the subject to carefully consider the options. It's not the options that are important here, it's the process.

BryceApr 05, 2005 at 11:48AM

Some associations that spring to mind... make of them what you will:

1. My friend Brad used to be fond of pointing out that 'anyone can make a straight line with only two points.' His context for this observation was in the world of writing, where he believed that a bare minimum of THREE supporting facts was mandatory to prove any thesis.

2. Scott McCloud is fond of this observation about Ernie Bushmiller (of 'Nancy' fame):

Much has been made of the "three rocks." Art Spiegelman explains how a drawing of three rocks in a background scene was Ernie's way of showing us there were some rocks in the background. It was always three. Why? Because two rocks wouldn't be "some rocks." Two rocks would be a pair of rocks. And four rocks was unacceptable because four rocks would indicate "some rocks" BUT IT WOULD BE ONE ROCK MORE THAN WAS NECESSARY TO CONVEY THE IDEA OF "SOME ROCKS."

Again, these are just some odd triad-related things that come to mind...

jason rossittoApr 05, 2005 at 11:51AM

firmitas, utilitas, and venustas is a rule of three set down by the ancient Roman architect and first codifier of architectural theory Vitruvius. It means Firmness, Utility, and Delight.
Like some of the other examples here, good buildings, that is elegant solutions to the problem of building, must have all three.
When we build we most often choose two and pay lipservice to the third. Are there elegant solutions to all of these '2 of threes?' Is it as Ryan says about startups?
Yes and no. The point that startups are paid for with sweat is well made. This is true of architectural design too. A building which is such an elegant solution is the product of careful and thoughtful research and a long series of unsuccessful attempts (we call them iterations). But then time and money are not elements in Vitruvius's three-some (although they could be read into the concept of utility).

There is an example of a four-some in Heidegger's concept of the fourfold, although there is no choice implied. The fourfold are the things which we make manifest in that which we build in order to authentically dwell. They are earth and sky, mortals and divinities. Dwelling means simply being or our way of being. "we are on the earth, under the sky, as mortals, before the divinities."
That was sort of a tangent but its Interesting that the the title of the work in which this is found is a three-some. Building, Dwelling, Thinking, which was published, among many other places, in a book called Poetry, Language, Thought.

Jacob MartinApr 05, 2005 at 11:51AM

Yes, everyone is indeed missing an underlying mathematical relationship here about the relative number of possible sets of mutually exclusive sub-sets of items.

In other words, it's much easier to come up with 3 things such that choosing any 2 of them mutually excludes the remaining item than it is to come up with 4 things such that choosing any 3 of them mutually excludes the remaining item (and similarly for even larger numbers).

This means that the "solution space" of possible n-element sets is dominated by 3-element trickery. (Homework: Try to create a "here's ten things, you can have any 9" question that works!)

I agree with Ted as to why we like "pick (n-1) out of n" questions, but should add that I think it's also about making the minimum tradeoff to obtain the maximum number of desired outputs.

David NestorApr 05, 2005 at 11:58AM

I forgot to mention the obvious conclusion... Three Is a Magic Number.

JochenApr 05, 2005 at 12:09PM

To me the pick-two-problem seems to be rather a problem of our own resources than of the resources themselves. Take your first example:

Elegant, documented, on time

I theoretically can deliver elegant, documented software on time - unfortunatley I am not brilliant / cool / engaged enough.

So the reason why the pick-two-problem is not a pick-17-problem may lie in our brains: We are not built to deal with too many variables in parallel. (Remember some brain research on this, can't find it right now.)

DnashApr 05, 2005 at 12:15PM

While I can't provide more insight into the "why" - I thought I'd offer another example. From Armistead Maupin's Tales of the City books: "Mona's Law" - "You can have a hot job, a hot lover and a hot apartment, but you can’t have all three at the same time."

AnthonyApr 05, 2005 at 12:16PM

There definitely is something about a trinity. As the jokes go, three people always go into a bar.

Bob BarsantiApr 05, 2005 at 12:26PM

While there is some great commentary, I think this "pick two" is a great mental shorthand and applaud you for fleshing it out.

I don't think the two is important, as is the three. In other words, the assumption of "healthy, tasty, and cheap" is that all three would be a trifecta-perfect. Perhaps you could come up with four or five qualities and muck up the idea: "Good, fast, cheap, or healthy" so, you can make a Good product cheap, but the workers won't be healthy or...

Language further games the system. What do you mean by fast or good? In a limited set of people, these words could have an assumed definition. That definition falls apart in a larger group. For example, I may see GTAIII as good, while others wouldn't.

The wandering knife of Quality (I see you Phaedrus) asks you to prioritize and accept less than perfection. They beauty of this formulation is that it forces you to use the knife. Father, Son, or Holy Ghost?

AlexApr 05, 2005 at 12:33PM

On the math side, it's a triangle.
This is a very interesting proposition, because time and money are definite values, and if you assign those values to the sides of a triangle, then you have the value of "quality", arbitrary or not, you have a number.
You run into all sorts of semantics, trying to interpret what the angles mean, what the number you get for "quality" actually means something, and what units it'd be in. It's neato.

Jeff EgnaczykApr 05, 2005 at 12:36PM

Elegant, documented, on time

HA! Well I guess it better be elegant and on time!

bernadetteApr 05, 2005 at 12:48PM

how about approaching this by thinking about its opposite? i.e., where "two gets you three".

specifically, this makes me think of an fun adage i read a while ago, regarding men:

money, sex, power:
any two will get you all three.

bllApr 05, 2005 at 12:52PM

I'm joining this late and no one has brought up "Rock, Paper, Scissors" as an example?

As others have mentioned, just about everything has some sort of "rules of 3," from comedy to painting. Why? Dunno, probably the way we wired as it can be traced all the way back to Aristotle and his "Three" Unities" describing how plays should be written.

RolandApr 05, 2005 at 12:56PM

Cheap, healthy, tasty.

Assuming this is talking about food, I have to disagree. Plenty of fruit and vegetables are all three.

BruceApr 05, 2005 at 1:20PM

Another trilemma:

Do what you love, make lots of money, stay within the law (pick two)

John BApr 05, 2005 at 1:24PM

Peter says:
so if i want something good and cheap it will take a long time? i don't see how that combination works. wouldn't more time result in more cost?

If we're developing a web application "good" generally means more features. And if you want it cheap and good it will take a long time because we probably won't be able to throw as many "resources" at the effort. Hence the time is longer, but the actual manpower (personpower?) used will be less.

jkottkeApr 05, 2005 at 1:28PM

Cheap, healthy, tasty.

Assuming this is talking about food, I have to disagree. Plenty of fruit and vegetables are all three.

Right, but in the aggregate, apples that taste better than most apples (because they're organic or homegrown or more fresh, etc.) are generally going to be more expensive and healthier than apples that are mass-farmed/produced (if you pardon the pun). Again, these are obviously cliches and there are all sorts of examples/counterexamples, but this one holds up well I think.

Mike DApr 05, 2005 at 1:28PM

Peter says: so if i want something good and cheap it will take a long time? i don't see how that combination works. wouldn't more time result in more cost?

That scenerio says that you can give a client something good for cheap, so long as you can give it lower priority than your other projects and don't have deadlines or milestones. For instance, I built a nice site for my friends' band for free, but I took 3 months to complete it.

The time factor mostly applies to freelancers. At my day job, web projects are billed by time, so doing it over 1 month vs. 3 months doesn't make a difference. In that situation, it's "Good and Cheap: Pick One."

Lara MurphyApr 05, 2005 at 1:30PM

There is also the NYC two-thirds theory, wherein you can like your job, like your apartment, and like someone romantically, but rarely/never all three at once. It's the Holy Trinity of NYC, the true American Dream. And if you get this three out of three here, and also manage to have your own washer/dryer, well, you've blown my mind and I hate you.

barnesApr 05, 2005 at 1:31PM

The "good, fast, cheap" idea is related to the supply-demand model. The intersection of the two curves (supply and demand) identify a third variable: price. With the "gfc" scenario, the saddlepoint between the "time" and "quality" curves change the "cost" variable. Shifting one of the curves will affect the cost saddlepoint and the other curve, just as changing the cost target wil also affect one or the other of the curves.

In gfc, picking two ranges of values for the time and quality variables ("good" and "fast," for instance) will raise the saddle point outside the realm of "cheap."

David KornahrensApr 05, 2005 at 1:35PM

Could be something Kottke.org started? I was doing some reasearch on it as well, after reading this, and I haven't found another system alike.

Dave TomlinsonApr 05, 2005 at 1:39PM

One word: Meatloaf.

Rajan PatelApr 05, 2005 at 1:49PM

i would like to say you're all losers and i want all 3.

Kris PayneApr 05, 2005 at 1:52PM

I learned about the pick-two principle from reading Design Observer. I think it's the Triangle theory that fit's it so well. However, I am totally with you in wodnering why it's two out of three, although it completely makes sense when you think about "good, fast or cheap."

Jason C.Apr 05, 2005 at 2:02PM

Here's one to think about when looking for a partner:

Attractive, intelligent, sane.

Pick two.

Steph MineartApr 05, 2005 at 2:16PM

When I first heard this in a lecture (time? place? I have no idea) there was a mathematically diagram attached to it that illustrated why achieving only two of the three was possible. Of course, I couldn't draw the thing to save my life now, so I'm no damn help at all, am I?

diddyApr 05, 2005 at 2:25PM

I think Lloyd Dobler had something to say about picking two (or none)out of three.

"I don't want to sell anything, buy anything, or process anything as a career. I don't want to sell anything bought or processed, or buy anything sold or processed, or process anything sold, bought, or processed, or repair anything sold, bought, or processed. You know, as a career, I don't want to do that."

HCApr 05, 2005 at 2:35PM

Choosing one of two is a choppy, digital act that does not lend itself well to many situations. Project Management is as much art as it a science. You really tend to delineate needs based on must have, nice to have and total fluff when the choice is two of three. Usually a better and more analog way of making a decision, IMHO

funtooshApr 05, 2005 at 2:46PM

puh, this goes way back to the time when i studied philosophy, but there is a rhetorical topos of using a triade ... classically, you would have a binary opposition, from good vs. evil to the modern ideas of dualism.

i think it was already in greek philosophy that triadic models were used to describe more complex setups that a mere dualism could not explain.

now i should go & search through my notes, where the modern use of triadic models comes from, shouldn't i? i think the frankfurter schule, adorno etc did it a lot, and it might have something to to with hegelian dialectics, thesis > antithesis > synthesis - this is rather embarassing, how much you forget in so short a time ...

embarassed, and willing to look things up:
-martin

PS by the way, thanx for reminding me ;~}

liaApr 05, 2005 at 2:59PM

Assuming this is talking about food, I have to disagree. Plenty of fruit and vegetables are all three.

Depends on where you live, this. In poorer parts of NYC, fresh fruits and vegetables are more expensive than fast food.

funtooshApr 05, 2005 at 3:07PM

you might want to have a look at, as always, the wikipedia:
http://en.wikipedia.org/wiki/Dualism
and more importantly:
http://en.wikipedia.org/wiki/Dualism_%28philosophy_of_mind%29

the so-called GUT makes use of a triade as well:
http://en.wikipedia.org/wiki/Grand_unification_theory

for those who speak german, the article on dualism is rather different:
http://de.wikipedia.org/wiki/Dualismus
and there s even an article on "trialism" tracing it from aristotle to hegel's dialectics:
http://de.wikipedia.org/wiki/Trialismus
http://de.wikipedia.org/wiki/Triade_%28Philosophie%29
and even:
http://de.wikipedia.org/wiki/Quadrialismus

but there is the idea of the triade as a (mere) rhetoric figure, too - i just don't have a respectful link at hand, right now ...

will that suffice? ;-)

HernanApr 05, 2005 at 4:02PM

It is not really about choosing 2 out of 3, because the same choice can be redefined as choosig one out of three: the one least preferred.

Was it Einstein that said (in the early 30s, way before the more terrible events) that you must be two out of:

1. Intelligent
2. Honest
3. Member of the National-Socialist Party

but not all three?

Scott StowellApr 05, 2005 at 4:10PM

I have a better suggestion (and it's a "one out of two"):

Here are two choices:
1. good
2. not good

Pick one.

"Good" can't usually happen fast or cheap.
"Not good" is easy to do fast or cheap or both.

LennyApr 05, 2005 at 4:15PM

"Why is 'pick two out of three' the rule? Why not 'one out of two' or 'four out of six'? Or is 'pick two out of three' just a cultural assumption?"

One picks two out of three instead of four out of six because in all of the example you gave there were three options. :-)

richardApr 05, 2005 at 4:40PM

I actually agree with Kottke, two out of three. You present the options and stick by them. Just as easily it can be three out of five, or ten out of thirteen, but three to me sounds simple and to the point. Giving clients more and more options is only going to complicate things anyway. Generally speaking I tell mine to tell me what they want, anything after that comes as a redesign, new pricing and new structure etc.,

obliviousApr 05, 2005 at 4:42PM

The pick two out of three rule cannot be understood as a mathematical formula. It is a verbal instrument that allows people to explain that in gaining a greater good one needs to compromise on another front.

Two out of three, because that is the least complex relationship, that allows us to visualize the lesser loss we Have to suffer for the Greater good. Two is enough to make our point. [I agree in part with David Nestor above]

We can't analyze this as math, since none of the three elements involved can be weighted against each other. Each person might have his own reasoning as to why one is heavier that the other two. This is merely a manner of speech, that can only be used when both parties involved understand the relationship and the relative weight of the elements involved.

spsApr 05, 2005 at 4:59PM

As a senior in HS looking to run at the D1 level, I was told by a friend that coach gave her this type of advice:

academics, athletics, social life...you can do two well

so yet another variation on why one would pick two out of the three. In this case, those are pretty much the defining characteristics of campus life, in broad, LCD terms. From my experience, the coach was right. Attempting to do all three well drives one completely insane and produces mediocrity in all three.. Of course, there's a hidden 4th variable here which is sleep.

I think in this case it has to do with deciding what you want to do in life and balancing it. Doing something well to me doesn't mean average, or getting it done, or managing at least the status quo. It's rising above all that. Unfortunately, at least for me, the temptation has always been there to try pulling it all off, no compromises style...but when I have, I realize that by doing it, compromising is exactly what I'm doing.

Anyway, so yeah, no real answers here, but I couldn't find this sort of reasoning for 2 out of 3 (you can do two well, but not all three) anywhere else in your list or in the comments.

MartinApr 05, 2005 at 5:30PM

It's very similar to Shirky's Power Law:

You can have popularity, but not quality.

Or

You can have quality, but not popularity.

Don't fight the Power Law.

yiApr 05, 2005 at 5:46PM

i've learned that too many choices is a bad thing. 3 seems to be just about right. maybe the adage is 2 out of 3 because designers prefer speedy decisions from our clients. any more causes confusion.

but sometimes 3 choices is still too many. after reading this this morning, i went to lunch today at Texas French Bread. i got their lunch special which cost $7.95 and you get two of the three: half a sandwich (ham and swiss, turkey, chicken salad, nicoise, pimento cheese or vegetarian), salad (green, pasta or fruit), soup (usually they only have one soup, but sometimes they will make two soups), plus chips, salsa, a cookie, and a drink. once i decided i wanted a sandwich and a salad, i then had to choose WHICH sandwich and salad i wanted. then i had to choose WHICH cookie and which drink i wanted. tired!

ZacApr 05, 2005 at 6:02PM

I agree with Anthony: it's a joke and jokes use threes.

There are probably whole theses on the semiotics of 3 out there. We need some full-time netizen without a day job to find them *cough*Jason*cough*. Semiotics-lite: Umberto Eco's _Foucault's Pendulum_ contains a character, Lia, who (too late!) breaks the news to the protagonist about the mundane realities of numerology and high symbolism (her speech quoted in full here). Lia reckons three has significance because it is the first number not represented in our bilaterally symmetrical physiology and also because it is linked to the greatest mystery of all for pre-scientific societies: two parents and...pow, baby makes three!

Meanwhile in cognitive anthropology, it's a human universal to comprehend one, two, and "more than two". This fits with Mobil'Homme and Bryce's comments above, that 3 is the bare minimum shorthand for "some". Our brains understand three intrinsically, it's cognitively compatible. Four is one too many.

Which is why jokes use threes.

Jason Coleman: I think it is a trilemma, albeit a cryptic one. The choice is neither to pick two out of three nor to pick the lesser of three evils. The choice is: of XYZ, choose XY, or XZ, or YZ. That's a trilemma. QED. (And that's a TLA.)

But mostly, it's a joke and jokes use threes.

jfwlucyApr 05, 2005 at 6:16PM

One other late addition -- this also reminds me of the game -- Which One Would you Marry, Have Sex with Once, and Throw Off a Cliff.

The questioner gives the questionee three roughly equal choices, according to her/his particular tastes (i.e. Brad Pitt, George Clooney, and Harrison Ford; OR Dick Cheney, PeeWee Herman, and Carrot Top.) The questionee has to choose which among them to marry, have sex with, etc.) Invariably three questions, though there are some variations on what the questions are.

jfwlucyApr 05, 2005 at 6:20PM

See link for more examples:

http://www.scraplove.com/forum/archive/index.php/t-265.html

Justin JonkerApr 05, 2005 at 6:21PM

As an engineering student, the two of three system was pounded into my brain somewhere during freshman year and repeated when necessary by professors who have lots of experience in these situations.

For me it always makes sense as a gradient on a coordinate system with three nodes, taking the shape of an equilateral triangle. You can have lots of one thing, and less of the other two, but I don't really agree with just 2/3. It's more like you have a set number of credits to distribute over three areas, like "production" in a turn based strategy game (population/production/technology) or picking a player's abilities in a sports video game (speed, agility, accuracy).

I have always felt that even in design or engineering you can pick the middle point and try to stay close to there. It is harder, like walking the narrow path next to the slippery slope, but I have always felt that a good design team should be able to walk that tight rope together. Why not? Maybe some long hours, maybe some dedication to a project, looking out for each other, knowing when to say "good enough." It requires discipline which can only be found with a variety of personalities making up the team.

essApr 05, 2005 at 6:25PM

As Zac and the rock poster said, three is the best way to think about things. It's just what we do. Vivid contrasts and narrative tention fall apart when you have more than 3. (Much as my mind falls apart with the flu.)

That said, this is just a joke explaining that even when we think we've done all our analysis and pre-planning and scoped a project, we're doomed as doomed can be. Brevity is required to make this point with wit.

Engineers were making this joke before the end of the 19th century - and possibly earlier.

That suggest it has resonance.

Some of the others on the list seem contrived, to me.

miesApr 05, 2005 at 7:15PM

I think we're looking at this from the wrong end: it's not so much choosing the two out of three we want, its choosing the one which we are willing to tradeoff, i.e. sacrifice. The hard decision is not what we want, but what we are willing to give up. So in that sense, a two out of four would also work, but the tradeoffs are more complicated. In the two-out-of-three scenario, we can see how the one value/goal/whatever that we are sacrificing impacts the two values/goals/whatevers that we do want. So it is an issue of being able to see causality.

Peter O.Apr 05, 2005 at 7:24PM

Although it may not look it, each of these is pulling at the same resource. Good, fast, and cheap, which I think is the hardest to identify of the list, each pull at time-money. Sort of like how there is space, and there is time, and there is space-time. So you have this much time-money:

|---------------------------|

You gotta divide it up somehow. People go for giving 50% to two traits because giving it away in thirds doesn't leave you with the feeling that you are actually making any decisions. It's also easiest to think about solving a problem with two goals, rather than three. But there's no reason you can't divide this resource up equally. I mean, do you really think it would be impossible to do a project that was kind of okay, delivered at a moderate speed, and costs you a fair amount of money? No. It's just more fun to have something that's awesome and took you a couple years.

dan hartungApr 05, 2005 at 7:52PM

I'm with mies and others who've explained this in the résumé-->circular file sense: you've gotta decide which factor to ditch. It's definitely more of a linguistic or cultural "equation" than anything formally mathematical.

Along those lines, there's a forgotten analog device (I can't even remember its name, but it's associated with some French mathematician) which allows you to balance resource costs on a graph, or even more usefully, a map. You take a ring, attach three lines to it, and attach weights to the lines representing the delivery cost to you of three major resources -- for instance, coal, steel, and wood. You drop these weighted lines through holes in the map where the resource originates. The different weights pull the ring across the map to a point of equilibrium, which mathematically corresponds to the optimal location for your factory.

This example illustrates a more flexible approach -- instead of 100-100-0, you can balance your three things like 80-60-20. This is probably more realistic, but harder to express mathematically -- let alone in a simple English blurb.

druApr 05, 2005 at 9:59PM

Career, Meaningful Relationships, Children

Jason ColemanApr 05, 2005 at 11:24PM

Zac: Okay, I'm sold. It's a negative trilemma... you must choose the one you can live without. I was just seeing the glass half full. Oh, speaking of jokes and threes:
Optimists see the glass as half full. Pessimists see the glass as half empty. Engineers see the glass as having a safety factor of two. See, how I used three one, two, three thing there?

MarcApr 06, 2005 at 12:10AM

From my days - ages ago - when I was a bank teller: "Fast, Accurate, Courteous". They actually talked about this in our training, and explained we'll never manage to be all three, so pick the two that are "right for you". To which I replied, "WTF?" But it was so true. I was fast and courteous, and at the end of the day, my cash drawer never balanced. Customers fucking loved me (wonder why?) but the manager hated my guts. Ironically, month after month I'd win the customer service award. Hmmm...

ricardo vacapintaApr 06, 2005 at 12:18AM

(Homework: Try to create a "here's ten things, you can have any 9" question that works!)

1. Alice can have 1/9th of the cake
2. Bob can have 1/9th of the cake
.
.
.
10. Joe can have 1/9th of the cake

Martin S.Apr 06, 2005 at 1:54AM

Jon Favreau on his TV show "Dinner for Five" has several times mentioned a lesson he's learned in show business. The way he put it was, if you want to work on a good movie, then two of the following three elements had better be good: director, script, cast. That is, if you have a good script and cast, you can take a chance on the director, and so on around the triangle. If you have only ONE good element, though, and are taking a chance on the two other elements -- this generally leads to bad movies.

I think it's an interesting variant because it kind of implies that you WANT to take a chance on one of the elements -- you actually do want to take a chance on a young director who has a chance of being really good -- but if you do, you better do it with a solid cast and script (and so on around the triangle).

jasApr 06, 2005 at 3:28AM

another one i've been dealing with lately in trying to buy a house: size/ price/ location
pretty obvious, but this search is killing me and had to share.

BruceApr 06, 2005 at 3:31AM

Jason Coleman:

The glass is completely full, half with water, half with air.

Kyle HaleApr 06, 2005 at 3:53AM

I think that Hegel's dialectic theory explains the significance of the triangle in problem-solving and decisionmaking very well.

Hegel's dialectic theory, in short, suggests that everything in our world has an opposite. Our goal is to combine one aspect of the world (the thesis) with its opposite (the antithesis) into a new, more ideal system (the synthesis.)

Mentally, we begin the dialectic process at a very early age. If babies cannot see an object, they assume it is not there. The game of peek-a-boo synthesizes the thesis idea of "here"-ness with the antithesis of visibility.

In short, all of our thoughts, all of our syllogisms and negations and positions, must be synthesized out of two other ideas. This is actually even more elegant as viewed through the "pick two out of three" because our thought process only sees two ideas as valid at any moment, and we can visualize in our mind what "fast and good" vs. "fast and cheap" vs. "good and cheap" might mean for our project.

Hegel's Phenomonelogy of Spirit. An excellent read, friends.

Tim MartinApr 06, 2005 at 6:55AM

Meatloaf laid it all to rest with Two Out of Three Ain't Bad

MartinApr 06, 2005 at 7:10AM

Jason Coleman said:

The glass is completely full, half with water, half with air.

Reminds me of a classic Gary Larsson Far Side cartoon.

Split into four panels, the first panel had a woman staring at a glass saying; "The glass is half empty".

The next had a guy staring at a glass saying: "The glass is half-full".

The third had a guy staring at a glass saying "The Glass is half-empty... half-full... half-empty.... half-full".

The fourth panel had a guy staring at a glass saying: "Hey, I ordered a cheeseburger".

The caption beneath read: "The Four Basic Personality Types".

AKApr 06, 2005 at 8:37AM

good, fast, or cheap

We use this not to reduce our responsabilities but to provide our clients a sense of limits. As a company, we need to improve all three to survive. For a client, they need all three to engage in a project. For a company-client relationship, you can't have all three because it will never have a defined "ending". When is it cheap enough? When is it good enough? When is it fast enough?

It helps us and the client make judgement calls on what's important to them so you can justify the choices. We've had the problem where we're nice to clients and overlook the diregarded choice, eg good "i'll just work a little longer to get it perfect", fast "I can work all night", cheap "sure, I'll do it because I like the client"; and the project and client relationship spiral out of control. Soon you get the client asking why it was free yesterday and why they can't get the same today.

ParagonApr 06, 2005 at 10:37AM

My favourite.

Woman - beautiful, smart, kind

Pick up any two.

Ronan H.Apr 06, 2005 at 12:35PM

responding to Peter O.

each pull at time-money.

No, cheap and fast pull at time-money, good pulls at a third 'quality' variable.

do you really think it would be impossible to do a project that was kind of okay, delivered at a moderate speed, and costs you a fair amount of money? No. It's just more fun to have something that's awesome and took you a couple years.

No, in software there is no such thing as 'kind of okay'. You have 'absolute crap' and 'quite good'. There is no perfect, and no 'good enough' (any project that is able to achieve good enough is going to have to also be quite good or even excellent). Also, the concept of 'good' in software encompases attributes such as maintainability (so that you don't have to spend 9 months adding a 'trivial' feature to a project that took 6 months to code in the first place) and its cousin: modularity (helps maintainability, and if a re-write is in order, it might only be a re-write of 5% of the project, not the whole thing).

If you don't believe what I am saying, look at the history of software development. How many projects were never delivered? How many projects were delivered (late and at great expense), but had to be abandoned because they: didn't work (e.g. fulfill original requirements) or didn't fulfill the real requirements (because the customer gave bad requirements, or the requirements changed). Software development is fraught with un-containable and inestimable risks to a greater extent than almost any other profession (except for neurosurgery?).

I am not bragging about my skills as a developer, nor am I being pessimistic. I believe that by the time humans have been developing software for 1,000 years (less time than we have been building bridges or houses), we will have a science of software engineering as mature as that of civil engineering.

Pete FreitagApr 06, 2005 at 2:20PM

For it to be a trilemma you can only pick one of the options.

ricardo vacapintaApr 06, 2005 at 3:04PM

This thread has reached the "repetition point"

That is, it is no longer of a size manageable enough to read all the way through and so comments from here forward will, in 2 out of 3 cases, simply repeat something that was said before.

Later, the thread reaches the "duplication point" in which ALL comments are repeats of what has been said before because it has, in one form or another, all been said.

So, a thread can be:

Non-Repetitive
UnModerated
Has a "best" answer somewhere in it

Choose two. :)

jkottkeApr 06, 2005 at 5:26PM

So, a thread can be:

Non-Repetitive
UnModerated
Has a "best" answer somewhere in it

Choose two. :)

At which point it becomes self-referential. Excellent. I expect the thread to become self-aware sometime in the next 2-3 months...will keep everyone posted.

rupaApr 07, 2005 at 1:11AM

Physiological proof of three: the heart beats lub-dup, lub-dup, lub-dup ... if you think that's two you weren't counting the commas.

Gary-OApr 07, 2005 at 9:12AM

It's a cop out. A way of shifting blame to someone else. If I tell you that you can have 2 out of Cheap, Light, and Strong, then I can blame YOU for any errors or faults in the end product. "Hey - you said you wanted it CHEAP and Light. So that's what I gave you." This is just a fancy way of saying "Passing the Buck." If I turn out to be a homicidal lunatic then it's my wife's fault for choosing Attractive and Intelligent, but leaving out Sane when shopping for a mate. Though pop culture would tell us differently, Attractive, Intelligent and Sane shouldn't be too much to expect from people.

All this really is is a new way of describing the human need to blame someone else for our own mistakes.

angelaApr 07, 2005 at 3:18PM

This thread is good, pretty and cheap.

L. PriceApr 07, 2005 at 4:26PM

There's an interesting variant of the pick two out of three that I ran across in an old book on systems analysis. It's called Accept, Constrain, Optimize.
It appplies to Schedule, Budget, Quality. So you can Optimize one axis, as long as you can constrain another and accept whatever the third is.

I've found it a useful tool in setting out the tradeoffs for a particular project.
It's a bit fuzzy since Quality is a very subjective metric, and you can invert the meaning of the qualifiers (Optimising the budget is spending the least).

For myself I find it gives me guidance, in that usually I constrain quality, accept schedule and optimise budget. Even though my usual instinct is to optimise schedule, constrain budget and accept quality (I've found that this ultimately morphs into optimising quality, constraining budget and accepting schedule which is even more pathological).