According to the NewHackersDictionary
], science fiction author TheodoreSturgeon
once said "Sure, 90% of science fiction is crud. That's because 90% of everything is crud."
The law is almost always quoted with the word "crap", so that can be considered to be the common usage, but Sturgeon himself said "crud", so it depends on whether you want an exact quote or the common usage.
Of course, Sturgeon may very well have been thinking 'crap', but wrote crud to make sure it got published. And remember, 90% of all self-censoring is cra..err crud.
By itself, the quote is not much more than a glib remark. But...
- In its original context, it was far more, because Sturgeon was replying to someone who was dissing science fiction, saying 90% of it was crud, implying that this was something unique to, and damning of, the whole field of science fiction. Thus Sturgeon's response, which is more enlightened than glib in that context.
- Of course, some in the ScienceFiction community try to flout SturgeonsLaw by redefining ScienceFiction in order to exclude the crud. Of course, that leads nowhere but to the NoTrueScotsman fallacy. (See SturgeonsLawDisproved)
- Yet Sturgeon's Law generalizes to that, as well: people in the art community often try to redefine "real art" to exclude the crud. People do that in every community. Yet crud still exists.
An important corollary to SturgeonsLaw
... but ninety percent of everybody thinks they are part of the ten percent that's not
People are far better at recognizing incompetence in others than themselves - see UnskilledAndUnawareOfIt
There is a sort of converse to SturgeonsLaw
, in that it has been said that when 10% or more of a given population reach enlightenment, society as a whole will dramatically improve. Perhaps this can be extended to mean that when at least 10% of any given (whatever) has high quality, then the (whatever) as such becomes interesting. Sort of a rule-of-thumb minimum signal-to-noise ratio for information...? Applicable to code/design/patterns/methodology/? -- BoLeuf
DuaneElgin was on NPR the other night -- he said it was 20%. Who knows....
- The founder of Transcendental Meditation said this, giving the number as 1%, decades ago (and before his organization began to exhibit the symptoms of a corrupt cult, so never mind the ad hominems).
Obviously if 90% of everything is crud then there is a 10% distillate of non-crud, be it truth or wisdom. SturgeonsLaw
reminds me of the ParetoPrinciple
which is that 80% of the (cost, significance, property) is produced by 20% of the (parts, content, owners) ... Ultimately what this may mean is that there is a lot of noise in ordinary processes and data and one must filter to find that which is really worth knowing, etc. -- RaySchneider
Not necessarily. If you assume that 90 percent of everything is crud (or crap) then that doesn't necessarily imply that
only 90 percent of everything is crap. Even assuming that 90 percent of everything is crap,
95 percent of a specific something could be crap. And also, it's likely that more than 90 percent of everything is crap - based on the thesis that, once 10 percent or more of everything is NOT crap, then society will greatly improve.
Additionally, crap is a very relative word. What one person views as 'crap,' another may view in a much improved light. Crap isn't very specific, either. If you say something is crap, and something much worse is also crap, what distinction is there between the two? So if those two items were placed into a set in which one was at the higher end, and the other at the lower end, then the superior object would no longer be considered crap.
Hopefully that makes sense. I'm not too sure what I'm talking about right now.
Then there's Larry Wall's Corollary: 111% of crud is everything
These days, "crud" is invariably replaced with "crap" if for no other reason than you don't dare produce a movie that's rated "G"....
If you did make "G" rated movies, some people will watch it now and 30 years from now, which is not the intent of film-makers today, it is to make a fast buck.
That was their intent back then, too. You don't go into business for any other reason. BTW, many of the classic movies from the golden age would not be rated G if they were released today.
[Well, frankly, dear, I just don't give a damn.]
if 10% of crud is super crud, & 10% of life is super life, then 10% of my brain stops me from suicide ;)
"90% of everything is crap. Except crap. 100% of crap is crap."
-- Too Much Coffee Man http://www.tmcm.com/
This sounds a lot like the ParetoPrinciple
: 20% of the effort achieve 80% of the result. Ie. start with the 20% (or 10%) non-crap first to achieve most of the result. For a perfect solution, you will have to dig through a lot of crap. And of course, 80%/90% of your perfect solution are going to be crap... -- ChristianRenz
There's a similarity, but of course the difference is that the ParetoPrinciple
can apply to work by a single author, whereas SturgeonsLaw
might not; it depends on a big literary gene pool to achieve its 10% non-crap. Most individual authors produce 100% crap.
To put it another way, SturgeonsLaw
discusses the makeup of heterogenous collections, while ParetoPrinciple
perhaps applies to both heterogenous and homogenous collections alike.
But the one discusses effort and the other quality, so there's no proper subset/superset law there.
Today, after considering the foolishness of Sun's JavaServerPages
and the lack of a Controller object in Microsoft's AspDotNet
, I concluded:
"Sturgeon's Law, revised: 90% of everything is crap. The other 10% was written in Smalltalk." -- RobMyers
- 10% of everything was written in Smalltalk? Were that true, SmugSmalltalkWeenies would all be jumping for joy (and would all have well-paying jobs in their favorite language)? Unfortunately, Smalltalk's market share is well below 10%. And not to dis Smalltalk, but I can think of many examples of software that are neither crap nor Smalltalk. (How many of the SoftwareMasterpieces were writtein in ST?)
- Let's analyse this logically. If 90% of everything is crap, then 10% is not crap. That seems reasonable. Obviously that which is not crap is written in Smalltalk. So the only possible conclusion must be that you are wrong about Smalltalk's market share. I suggest you revisit that assumption :) GlenelgSmith
- However, it does seem that >90% of enterprise code these days is written in either Java, WhateverDotNet?, or pre-DotNet VisualBasic--which does jive with SturgeonsLaw. (The delta between the two figures is doubtless due to GrandMasterProgrammers writing good quality code despite being instructed to do the SW equivalent of building houses with the following in their toolbox: a butter knife, a xylophone mallet, a corkscrew, a pair of tweezers, a box of thumbtacks, and a nail file)
Which means, since I STILL haven't learned how to program in Smalltalk, that 100% of everything I've ever coded has been crap. Hmmm... -- RobMyers
Actually as a theory its interesting since you have to wonder what structure everything is in order for 90% of every part of it to be crud. Is total homogenity of crudness the only solution, are there interesting fractal structures which are 90% crud however you look at them or is there some rule about the residual of any meaningful operation (e.g. resturcturing a company and getting rid of loads of people) readjusting so that the new composition is 90% crud? Can't solve it after a glass of wine... -- AndrewCates
As much as I admire the sagacity of Sturgeon's commentary, I believe he was overly optimistic. In most cases, his "90% rule" applies recursively to the remaining 10%, thereby producing a most depressing continued fraction. -- Mike O'Dell
Mike, if you're right (and I'm not even going to try to disprove it) you must realize that this implies that EVERYTHING is crap. You would end up with 99.999...%, which is equal to 100%. -- RobMyers
Making Sense by Mining and Refining
In order to make things UsefulUsableUsed
, may I recommended the process of mining and refining. Selection and Filtering of information and processes in a manner similar to making metal from its ore. -- DonaldNoyes