| |||||||||||||
Exceptions and rulesExceptions and rulesPosted Oct 29, 2012 14:52 UTC (Mon) by rfunk (subscriber, #4054)Parent article: Thoughts on the ext4 panic
Argh. This is an excellent article, marred only by the serious (mis)use of the nonsensical expression "exception that proves the rule". The net is full of explanation about the original meaning of this phrase, and why almost every modern use of it is wrong.
(Log in to post comments)
Exceptions and rules Posted Oct 29, 2012 14:57 UTC (Mon) by corbet (editor, #1) [Link] We have an internal reviewer who will be gratified by your comment.. sorry to have marred your experience. I'll make a rule not to use that phrase again, with only rare exceptions.
Exceptions and rules Posted Oct 29, 2012 16:53 UTC (Mon) by Trelane (subscriber, #56877) [Link]
> I'll make a rule not to use that phrase again, with only rare exceptions.
I cannot take exception to that exceptional rule.
Exceptions and rules Posted Oct 30, 2012 11:40 UTC (Tue) by zack (subscriber, #7062) [Link]
> I'll make a rule not to use that phrase again, with only rare exceptions.
Chapeau.
Exceptions and rules Posted Oct 29, 2012 15:33 UTC (Mon) by nix (subscriber, #2304) [Link]
It wasn't a misuse. This exception did indeed prove (i.e. test) the rule, and find it to be unchanged. ext4 is stable as hell.
Exceptions and rules Posted Oct 29, 2012 15:40 UTC (Mon) by rfunk (subscriber, #4054) [Link]
If there is an actual exception, then the rule fails the test. If the rule passes the test, there's no exception.
Exceptions and rules Posted Oct 29, 2012 16:29 UTC (Mon) by khim (subscriber, #9252) [Link] If there is an actual exception, then the rule fails the test. If the rule passes the test, there's no exception. Yup. 1001th wrong interpretation of the phrase. If you'll stop complaining for the minute and actually read the meaning of phrase in question you'll find out that original principle is exceptio probat regulam in casibus non exceptis ("the exception confirms the rule in cases not excepted") and that this article used phrase correctly The very fact that this tiny quirk which needed non-default (and non-recommended) mount options warranted such a huge ruckus means that the rule is ext4 is rock-solid in accordance to the exceptio probat regulam in casibus non exceptis principle.
Exceptions and rules Posted Oct 29, 2012 16:38 UTC (Mon) by rfunk (subscriber, #4054) [Link]
I was responding to nix's interpretation above. You're correct about the original meaning, of course. But I disagree that the usage here fits the original meaning. Even if I accept that it does, however, just the fact the the phrase is widely misunderstood and misused should consign it to the dustbin; if most people don't understand it correctly, communication isn't happening.
Exceptions and rules Posted Oct 29, 2012 16:49 UTC (Mon) by khim (subscriber, #9252) [Link] I was responding to nix's interpretation above. You're correct about the original meaning, of course. But I disagree that the usage here fits the original meaning. It does. We had a rule: ext4 is rock-solid and is robust no matter what you throw on it (unless you are doing something totally insane like using broken RAID controller). Now we have an exception: if you use ext4 with non-default mount option journal_checksum then you may expect data corruption in some rare cases. The fact that such a small exception raised such a ruckus can be used as a support for the original rule (after all if the file system corrupts data all the time then yet-another-bug in it will hardly be a newsworthy material). Even if I accept that it does, however, just the fact the the phrase is widely misunderstood and misused should consign it to the dustbin; if most people don't understand it correctly, communication isn't happening. Most people on this planet don't understand English. Should we stop using it and close LWN?
Exceptions and rules Posted Oct 29, 2012 18:40 UTC (Mon) by randomguy3 (subscriber, #71063) [Link] See Wikipedia's article on the phrase, which enumerates 5 usages (from most correct to least correct) from Fowler's Modern English Usage. I think that the article's usage would fit in number 3 (loose rhetorical): the use of the phrase is to draw attention to the fact that something is uncommon. The existence of the bug does not imply that bugs are uncommon. Nor is the ruckus itself an exception from anything.Most people on this planet don't understand English. Should we stop using it and close LWN?"Most people on the planet" is irrelevant. "Most people who read LWN" is much more relevant, and "tech-savvy English speakers" is a good approximation of that.
Exceptions and rules Posted Oct 29, 2012 16:29 UTC (Mon) by epa (subscriber, #39769) [Link] I think the idea is that the apparent exception to the rule, on closer examination, turns out not to be an exception after all. So the rule still holds. It is silly, though; as you say, if the rule holds then there was no exception after all.
Exceptions and rules Posted Oct 29, 2012 15:37 UTC (Mon) by tialaramex (subscriber, #21167) [Link]
"almost every modern use of it is wrong"
Or, to look at it another way, perhaps it's the usage mavens who are wrong. We don't all have to be Humpty Dumpty, but if the vast majority of people choose to use a word, a phrase or even an entire language in a way that overturns precedent you eventually have to bend to their will or be broken by it. This isn't an example like misnegation where you can argue that perhaps someone didn't mean what they wrote, the author intended exactly these words, they just didn't intend it to mean quite what it once did.
Even eggcorns, which begin with a mistaken re-analysis of a spoken word or phrase, sometimes enter the mainstream. Once upon a time "upmost" was clearly sometimes a mistake for "utmost" today it's unclear, in another century there may be usage mavens insisting "upmost" should be preferred as more logical and "down the pipe" might likewise displace "down the pike".
Exceptions and rules (digression) Posted Oct 29, 2012 15:55 UTC (Mon) by Richard_J_Neill (subscriber, #23093) [Link]
The modern use of this is confusing. After all, "The proof of the pudding is in the eating" means "The right way to test a pudding is by eating it".
Also, "80% proof whisky" means "tested, contains 40% ethanol".
Exceptions and rules (digression) Posted Oct 29, 2012 19:54 UTC (Mon) by davidstrauss (subscriber, #85867) [Link]
It's "80 proof," not "80% proof." "Proof" is like "percent," except with a denominator of 200 -- and only really used for liquor.
Exceptions and rules (digression) Posted Oct 29, 2012 23:10 UTC (Mon) by barryascott (subscriber, #80640) [Link]
I like the historic origin of testing alcohol content by seeing if gunpower will ignite.
http://en.wikipedia.org/wiki/Alcohol_proof
Exceptions and rules Posted Oct 29, 2012 15:56 UTC (Mon) by rfunk (subscriber, #4054) [Link]
But those exact words don't mean anything sensible in most cases. An exception cannot logically show that a rule is valid in the usual modern sense of "prove", and this is what most people mean. The modern idea seems to be that because this phrase exists, it must be valid to use an exception to prove that a rule is valid, which is of course logical nonsense. Logical nonsense is much different from simply giving new definitions to words or coining new words.
Exceptions and rules Posted Oct 29, 2012 16:34 UTC (Mon) by tialaramex (subscriber, #21167) [Link]
Yes, I think we're all familiar with the "But it's not logical" argument from the Lynne Trusses of this world. The problem with that argument generally, and here specifically is that language isn't about sharing propositions of formal logic. Interpreting other people's utterances as logical propositions makes for a briefly amusing diversion in a light drama but it's a terrible way to carry on a real conversation.
"It's like the Somme out there"
The legal origin of this idea of an "exception that proves the rule" is fascinating, but the phrase has taken on a life of its own. You will notice that people are also content to say "third time is a charm" (there is no logical reason to believe that third attempts are special in any way) and "If a job's worth doing it's worth doing well" (likewise, a shoddy job may be the only economic or practical option) and many other phrases which can't be defended logically. They're not doing it wrong, these phrases aren't intended to be truthful statements about the world, any more than anybody thought exceptions /actually/ prove a rule.
Exceptions and rules Posted Oct 29, 2012 16:42 UTC (Mon) by rfunk (subscriber, #4054) [Link]
The other examples you cite at least have internal logic. People may or may not believe them (and many people do in fact believe them as truths about the world), but they make internal sense without needing to believe them as true.
Exceptions and rules Posted Oct 29, 2012 16:41 UTC (Mon) by khim (subscriber, #9252) [Link] "Exception that proves the rule" is part of the phrase. The full legal principle is exceptio probat regulam in casibus non exceptis ("the exception confirms the rule in cases not excepted") as you, I hope, know. Classic example will be Special leave is given for men to be out of barracks tonight till 11.00 p.m.; "The exception proves the rule" means that this special leave implies a rule requiring men, except when an exception is made, to be in earlier. The value of this in interpreting statutes is plain. Modern example will be Now we know that exceptional case of ext4 with journal_checksum is not stable. Application of exceptio probat regulam in casibus non exceptis principle will mean that: ext4 is stable unless non-standard option journal_checksum is used - and this is indeed the case. IOW: this article is rare case where phrase "exception that proves the rule" is used correctly. P.S. When you start badmouthing people and explain that the net is full of explanation about the original meaning of this phrase, and why almost every modern use of it is wrong it's good idea to refresh your own knowledge and see if you understand what the phrase means and when it's appropriately used.
Exceptions and rules Posted Nov 8, 2012 12:01 UTC (Thu) by Wol (guest, #4433) [Link]
This comment actually gets to the root of the problem. The modern understanding of the word "prove" has changed.
Hence, the saying no longer means what it did, because the meaning of the words have changed underneath it.
Cheers,
Exceptions and rules Posted Oct 30, 2012 20:01 UTC (Tue) by cmccabe (guest, #60281) [Link]
Usually comments on LWN are very informative. This thread however, is the exception that... never mind.
Exceptions and rules Posted Oct 31, 2012 0:43 UTC (Wed) by nix (subscriber, #2304) [Link]
Agreed. I guess this just shows that the ext4 problem really *was* a tempest in a thimble (too small to be a teapot): thoroughly dull once the panic died away and its complete lack of terrifying implications sank in.
|
|||||||||||||