Captchas To Keep Idiots Out Of Comment Threads

http://www.dangerousminds.net/comments/captchas_to_keep_idiots_out_of_comment_threads/

by Udo • 15 years ago

233 202 15 years ago

Luyt3• 15 years ago

Still a 33% failure rate. And oh, http://www.youtube.com/watch?v=N4vf8N6GpdM

Udo3• 15 years ago

In my experience simple, domain-specific CAPTCHAs work pretty well. They don't even need to include a frustrating image with hard-to-parse words. Just ask commenters a somewhat randomized question that any visitor to the site would be able to answer easily. It can be really, really simple - like for example asking them to enter the abbreviation for something:

> Whats the abbreviation commonly used for the Hypertext Transfer Protocol?

This weeds out bots (because they generally don't bother with site-specific stuff) and many trolls (because they are either to stupid or to lazy to go through this process).

joelthelion• 15 years ago

That would work well provided your site isn't a high traffic site.

Udo3• 15 years ago

Yes. The lower traffic your site is, the easier you can make the captcha. The truth is, most of us have pretty low traffic sites with no more than 100k hits per month, blogs for example. The key here is to have a simple but different captcha compared to the standard stuff. Pairing this strategy with a little bit of site-specific domain knowledge goes a long way towards discouraging trolls and idiots.

I think overall, bots would be negatively impacted by a more diverse captcha ecosystem even if the individual implementations are easier to crack. Spammers don't invest 10 or 20 minutes in cracking a domain-specific captcha system just so they can spam one single blog - they'd rather move on to thousands of sites that are protected by the same standard word reading puzzle. Of course, if your site rises into the upper strata of the web, you're also quickly entering a zone where the standard, hard-to-decipher image strategy isn't working so great anymore either.

rocktronica• 15 years ago

And that the question randomizes pretty well. Otherwise posts could be automated.

CWIZO• 15 years ago

Jeff from coding horror for a long time used a very very simple "CAPTCHA". You just had to enter 42 (if I recall correctly), and it was always just that, nothing random. It served him well, and he has a high traffic site (http://www.codinghorror.com/blog/2006/10/captcha-effectivene...)

dredmorbius4• 15 years ago

The other use-case is defeating chaptcha-solving farms often run out of locations with low labor costs. I'm not entirely sure a specific cultural / intellectual / domain knowledge question would be more effective than simple character recognition, but it might be an interesting test.

dbshapco• 15 years ago

This seems to weaken the captcha for bots, since a simple Google search for (for instance) 'do you know we are going' completes the sentence. No need to defeat the obsfucated OCR test.

If the goal is to keep out some idiots (since these tests will admit 33-50% of idiots who can at least complete the OCR test but answer the question randomly) but permit moderately sophisticated bots, mission accomplished.

Providing more context to the captcha solution in general strikes me as helping the bots, even if it confounds idiots.

xd5• 15 years ago

Love the idea, but wouldn't it be simpler and more effective to just have them fill in the missing word without the images.

lfdias• 15 years ago

Agreed. If you're going the grammar route, why distort the words at all?

dhimes3• 15 years ago

It would take much more proofreading to make sure the questions aren't ambiguous. Look over ______. What am I thinking? U = {these, there, this,....}

xd5• 15 years ago

I thought that as well. Each question would need a good degree of context.

fferen• 15 years ago

I was thinking of giving 5 or so sentences and having checkboxes for whether they're grammatically correct or not. I feel that even people that don't exactly know the correct usage can still get the OP's test right because one "looks" better than the other, whereas with mine there's only one choice presented.

pr0filer_• 15 years ago

So, i can be a d*ck as long as i'm smart?

Edit: I'm disturbed by the amount of grammar nazi's here that think the quality of someone's grammar reflects their personality...

If the goal is to keep the conversation civilized, why not do something along the lines of "Doing X is a more nice thing to do".

alecbenzer3• 15 years ago

I guess smart dicks are slightly better than stupid dicks?

xd5• 15 years ago

I don't think it's really about being smart. Having a basic grasp of language so you can adequately articulate yourself in a conversation is critical to discussion. If you can't, then you probably shouldn't be taking part in the discussion anyway.

Too many people will spill there emotions into a discussion with sub standard spelling and grammar which makes it very difficult for others to understand them, which leads to inevitable misunderstandings.

raganwald9• 15 years ago

The grammar issues presented on this page are picayune distinctions with very little potential for major misunderstandings. Getting them correct is more of a social signal than anything else, like eating salad with the correct fork in some places and not eating food with your left hand in others.

I have real trouble imagining that a good point made with substitution erros is somehow less critical to the discussion than a poor point that doesn’t confuse it’s/its.

cromulent5• 15 years ago

> there emotions

Another good Muphry's law example.

raganwald9• 15 years ago

Am I to understand from the title that we are equating “poor spelling/grammar in English” with “idiot”? Or is there some more subtle mechanism at work, such as equating “can’t be bothered to stare at hard-to-read-text just to post lulz, wat?” with “idiot”?

JonLim• 15 years ago

I hate to do this, but it's "grammar".

Not that I'm calling you an idiot, you have a very fair point.

raganwald9• 15 years ago

Thanks!

dhimes3• 15 years ago

Excellent point. Mathematics would be better. Simple questions or sequences.

ianterrell• 15 years ago

There's no excuse not to know common homonyms and misspellings. A smart person who has trouble with them knows they have trouble with them, knows what they are, and knows to double check himself. If you don't do at least that, you probably aren't that bright.

We'll forgive the less common words like your "grammer".

Iv2• 15 years ago

I'm not an English native speaker. Their/they're/their also stings my eyes when incorrectly used, but I must admit that I often do the lose/loose mistake. But that's fair, whoever uses a grammar that is loose, lose.

cromulent5• 15 years ago

>There's no excuse not to know common homonyms and misspellings.

Would you consider dyslexia an excuse, or English as a second language? Or even semi-illiteracy? I think it's definitely possible to be very clever and still not a good speller, and not bother to double-check on an Internet forum.

Einstein was notorious as a poor speller, which suggests he didn't double-check himself.

raganwald9• 15 years ago

This sounds like a social signal, not a cause-and-effect relationship. You are not establishing that some quality of being an idiot causes someone to have poor grammer, you are saying that the ought to know their grammar “just because.”

Your argument reminds me of the argument that someone ought to wear a suit and tie to work: There is no excuse for not knowing that a suit and tie is what people wear to work, and anyone who doesn’t do so is not fit to work here.

silverbax882• 15 years ago

We are equating that after years of education, with the subject of English (albeit American English) being a core class for EVERY YEAR of attendance, and with spell checkers on most forum communities, that if someone still goes out of their way to transpose "lose" with "loose", then, yes, they are probably an idiot.

clstrfckr2• 15 years ago

*then

markokocic2• 15 years ago

> they are probably an idiot.

Did you mean "He is probably an idiot." or "They are probably idiots."?

stcredzero2• 15 years ago

A part of this phenomenon, is that those of middle age or greater miss the days when there was good quality control over widely published text.

The Internet has democratized widely published writing. There is no quality control over comments, except through communities. At the same time, there are large populations who place are undereducated and place little value on education, while they do prize the ability to get a rise out of others.

An unfortunate consequence of this: reading more used to make me feel smarter and I'd sound more literate. Nowadays, through reading reading I tend to absorb things like "loose" for "lose" and other atrocious sounding quirks.

There is a clear market for quality text, and there is power in being the provider of it.

damoncali2• 15 years ago

Maybe they're smart, but suck at spelling. Not everyone is wired the same.

alecbenzer3• 15 years ago

this would also seem to discriminate against people who aren't native english speakers (or w/e other language).

cturner• 15 years ago

It might discriminate in their favour. On average, foreign students I studied with had much better basic grammar than native English speakers.

cryptoz• 15 years ago

In my personal experience, non-native English speaks are far more well-versed in grammar than native speakers.

dchest• 15 years ago

I think native English speakers are more likely to make such mistakes. Non-native speakers usually know the difference between those words, because they have the ability to translate the words back into their native language to see the difference.

pygy_• 15 years ago

Not necessarily especially if you've learned the language at school, analytically rather than organically.

These words/phrases are wildly different in other languages, and thus well distinct in the mind of non-native folks.

For example, in French:

were -> étais; etions, etiez, étaient (2nd person, singular; 1st, 2nd and 3rd person, plural)

we're -> nous sommes

where -> où

cabalamat• 15 years ago

These people often have better written grammar, because they've had to take the trouble to learn it as a foreign language.

darklajid6• 15 years ago

Reminds me of a common beginner's problem if you're starting with German:

Who is 'Wer' in German

Where is 'Wo' in German

Trivial, but I see lots of people (and sometimes myself) falling for this trap.

drv2• 15 years ago

It happens the other way around too (German to English), and sometimes you end up with non-words like "somewhen". It's an interesting exercise to determine a writer's native language by the oddities of their writing.

alecbenzer3• 15 years ago

I was talking about people who just don't really know English very well (but well enough to communicate on some level)

nraynaud3• 15 years ago

I've the feeling that when you're speaking in a foreign language, you are more careful than in your own language. Maybe it's language thing, but I make ton of typos in French (double consonants, en/an, or/hors), and I can correct native English speaking people on stuff like your/you're.

ZoFreX3• 15 years ago

Exactly. I now live by the credos that someone making these mistakes may actually just not know better, or be struggling with a foreign language. So far (on HN at least) this assumption has held in many cases. It also makes me far less angry when reading the internet, regardless of whether the assumption is true that's hardly a bad thing.

desaiguddu2• 15 years ago

i like this idea..! I am planning to open source my CAPTCHA research if any one interested please contact me ..! you can find my research on http://dndcaptcha.blogspot.com/2010/04/textareaid.html

You can find presentation of the same concept here: http://www.slideshare.net/desaiguddu/drag-and-drop-captcha-a...

If anyone of you are researching or preparing Survey presentation or educational research on CAPTCHA feel free to use the content from the presentation.

n1ck4n• 15 years ago

He's missing this one

Fill in the gap: a) it's b) its

"The hacker got access to the server and killed ... processes"

lolz

burgerbrain3• 15 years ago

The problem with this one is that at one point in the English language's history (17th/18th century I believe), the correct possessive form of "it" was "it's", which is arguably more correct^.

^You make nouns possessive by adding a saxon genitive to signify the distinct possessive-izing "s" sound. The word "it" is also made possessive with this same sound when spoken, but for the sake of a contraction (read: abomination ;)), we have decided to arbitrarily remove the saxon genitive and replace it with a simple "s". The excuse for this inconsistency is "well it is a pronoun, not a noun, so this is not an inconsistency" does not take into account that English is a language primarily spoken. Written English, where reasonably possible, should approximate the spoken constructs.

Demonstration:

  He stepped on the cat*'*s tail.
  He stepped on its tail.

Notice that although these two sentences are expressing the same idea and are of the same approximate form (they would be spoken similarly), the second has dropped the saxon genitive. This is clearly, if you look at it ignoring what you were taught in primary school thanks to Webster worship, an absurd change.

At the very least, people who use saxon genitives with the word "it" do not deserve the ridicule people like to heap on them. "it's"/"its" is not a "there"/"their"/"they're" scenario.

corin_6• 15 years ago

I disagree, I think it/it's is completely comparable. Sure, there's a long history to it, but it has evolved, and there are two reasons why it is comparable.

First, we weren't alive back when "it's" was used this way. Nobody currently alive was. There's no argument of "they're doing it the old way", because it's not even in living memory. As of the current standards, and those of the past 200 years, they are simply doing it the wrong way.

Second, it can't even be argued that they are confused because of the old way. Nobody accidentally writes her's, their's or your's. The reason people write it's is nothing to do with the possesive apostrophe, it's purely because of confusion with the "it is" contraction.

jokermatt999• 15 years ago

My way of remembering its/it's is to associate it with his/hers rather than the typical 's.

`The cat stepped on his foot.`

`The cat stepped on its foot.`

Still, I don't get annoyed at its/it's because it's such an easy mistake to make. I still do it occasionally without meaning to. I don't quite get how people consistently get their/there/they're or loose/lose wrong, but that's just me.

abend• 15 years ago

Might be effective at keep the idiots out, but not the bots. The challenge sentence should not be in straight-up HTML.

gmac• 15 years ago

This made me smile, but I don't think it would work. I suspect a lot of people could use the right form when asked outright like this (and as a last resort, they could Google it). The bigger problem is probably when people don't care whether they get it right or not when they're writing their comments.

[Pre-submit verification of 'there' homophones . . . . 100%].

51Cards• 15 years ago

Seems to be a lot of serious comments about this when I believe the content was created with tongue planted firmly in cheek. That said, a valid moron captcha would be a wonderful thing in some places. It comes down to the old saying:

"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning." - Rich Cook

sqrt17• 15 years ago

a cloze test (i.e. some text with a gap for a word) would do very poorly in filtering out bots, as they're much easier to do automatically than captchas.

hugh4life• 15 years ago

What would an anti-hipster Captcha look like?

http://www.dangerousminds.net/contributors

colanderman• 15 years ago

What does that link have to do with your comment or the story?

desaiguddu2• 15 years ago

thats why the Smart CAPTCHAs are needed !

BasDirks• 15 years ago

Idiocy is not the privilege of bad spellers.

DanielBMarkham2• 15 years ago

So people who are detail-oriented grammarians are necessarily good commenters?

Since when?

The assumption here is that you can tell a comment is going to be stupid by testing whether or not the person understands homonyms. The only thing you check with that is language ability. I'd argue you'd be better off searching for common internet memes and cliches instead of original thinking -- since that's a sign of a lazy or uninteresting mind. The author is kind enough to inadvertently supply us with one "...If Fox Nation implemented something like this, they’d have zero commenters..." But you could test stuff like this based on any preconceived viewpoint, such as Obama-socialist, Obamacare, etc. The use of shortcuts and blindly-repeated jokes and phrases is a good sign that you're not going to be getting much from the comment.

Not sure how you'd code it, though, but I'm certain you could come up with some semantic magic given enough input text. I'd imagine you'd use n-grams and some Bayesian logic. You'd have to have a pre-existing corpus of the person's writing, though.

corin_6• 15 years ago

They are homophones not homonyms - normally I wouldn't be so pedantic, but in a thread around the idea of people having prove they know meanings/spellings/etc, it seems suitible.

DanielBMarkham2• 15 years ago

I actually doubted that was the correct word -- but didn't take time to look it up. Thanks.

mirkules3• 15 years ago

Don't you think it's ironic then that you misspelled "suitable"?

moss• 15 years ago

It's suitable, not suitible - normally I wouldn't be so pedantic, but in a thread around the idea of people having to prove they know meanings/spellings/etc, it seems apropriate.

dredmorbius4• 15 years ago

Good grammar and basic wordsmithery make idiocy much easier to manage.

Say: George F. Will vs. Rush Limbaugh.

neutronicus3• 15 years ago

I see an unfair prejudice against the vocal and valuable Indian and Eastern-European HN contingent.

mey• 15 years ago

several problems that need to be addressed. Non-native speakers would be arbitrarily penalized. This is very easy for a computer to crack. (small choice selection, systems to do grammatical checking have been around for a long time.)

MatthewPhillips• 15 years ago

This makes me wonder how much AI development is done for nefarious purposes. Someone's going to crack this and then they're going to have to make a more intelligent captcha system. Some day we're going to have strong AI and spammers will be thanked for it.

apl• 15 years ago

Highly relevant --

http://xkcd.com/810/

darklajid6• 15 years ago

If you haven't read this, please check out Cory Doctorow's story on exactly this point.

First link I could come up with: http://singularityblog.singularitysymposium.com/pester-power...

raganwald9• 15 years ago

Here’s my test for whether you’re qualified to post to Hacker News.

Pair the following five comments with their corresponding definitions:

1. “Raganwald is a blowhard whose pathetic attempts to score karma reveal him as an insecure dweeb who can’t get over being kicked around in grammar school. He and his mumblings should be should be flushed down the toilet bowl like the turds that they are.”

2. “What, raganwald is talking about beauty in code? Have you seen some of his code? Ignore him.”

3. “Raganwald sounds a lot like a Ruby fanbody, and we all know how those people think."

4. “Anyone who has spent that much time on Java clearly has no taste in software and cannot be relied upon for sound reasoning. Ignore raganwald.”

5. "Of course raganwald would say that there’s something wrong with Waterfall, he’s a Certified Scrum Master, he’s just pimping his own credentials.”

And the definitions:

A. Fallacy: Ad Hominem Abuse

B. Fallacy: Circumstantial Ad Hominem

C. Fallacy: Tu Quoque (“You Also”)

D: Fallacy: Guilt by Association

E: Not a Fallacy: Insults

Example: 1-E, This is an insult, but not fallacious.

mfontani• 15 years ago

http://en.wikipedia.org/wiki/Tu_quoque

raganwald9• 15 years ago

https://secure.wikimedia.org/wikipedia/en/wiki/Ad_hominem#Ty...

corin_6• 15 years ago

I know people smarter than the vast majority of HN users (how can this really be judged, sure, but I'd place money on it, certainly far, far smarter than me) who wouldn't know those, and I know many, many people who are damn smart (just not fitting them into the "smarter than most people here", but definitely smart enough to belong here if they wanted to) who also wouldn't know them.

raganwald9• 15 years ago

I beg your pardon, I was trying to crack a funny. It seems it is impossible to parody an extremist belief without somebody mistaking it for the general article. I was attempting to parody the fundamentalist debating pedant point of view.

Poe’s Law strikes again: http://rationalwiki.org/wiki/Poes_Law

mikeash• 15 years ago

They could look it up....

I have a really dumb spam filter on my blog which simply asks that the user "Type the word 'humour', but with American spelling". I've had a couple of complaints that this request is unfair to non-native speakers. My response is that you don't have to know it, you just have to know how to look it up.

(And yes, I saw that this stuff wasn't serious, but I thought it was a worthwhile point to make in any case. CAPTCHAs don't need to be things that users can pass without outside help.)

olavk• 15 years ago

Never mind if people are smart or stupid. The important thing for a forum is if the participants are willing and able to provide thoughtful and interesting comments. Simple quizzes like the above (joke or not) might filter out a lot of lazy trolls up front. Everybody with a genuine interest in the forum should easily be able to find the answers using Google, and might even find it fun to learn a bit.

More importantly, after having made a bit of effort to be allowed in, the users might value the forum more.

burgerbrain3• 15 years ago

Really I think you could just get away with testing to see if the user knows the difference between an Ad Hominem and a standard insult.

There is little more annoying than someone who thinks they see logical fallacies where none exist and take the conversation from bad (real arguments intermixed with old fashioned insults) to worse (nothing but an argument from fallacy)

"I don't claim to know what an Ad Hominem is." would also be a passing response in my book...

pbhjpbhj2• 15 years ago

Showing a person's argument to be fallacious doesn't show there position is wrong. Many it seems fail to grasp this point.

SkyMarshal• 15 years ago

<3 it. Reminds me of all the good reads on the multitudes of logical fallacies out on the net:

http://duckduckgo.com/?q=list+of+logical+fallacies

intellection• 15 years ago

Anyone else think it is someone else posting, before seeing it is self-referenced?

Very measured writing about yourself.

Copacetic shield.

tomelders2• 15 years ago

is this a trick question?

narag• 15 years ago

I can't know but, after reading the comment, it ocurred to me that identifying "idiots" with bad spellers is fallacious.

hammock• 15 years ago

1E 2C 3D 4A 5B Do I pass?

kenjackson• 15 years ago

I thought you were trying to be really clever by responding in hex.

pbhjpbhj2• 15 years ago

WRT 1. you're saying that

>"He and his mumblings should be should be flushed down the toilet bowl //

So arguably you're making the case to ignore his [implied] argument based on, in the antecedent sentence, his character. Sounds ad hominem to me.

Just saying.

Jach2• 15 years ago

It's possibly an ad hominem, but not likely to me. http://plover.net/~bonds/adhominem.html

nraynaud3• 15 years ago

(Non native english speaker) isn't the verb "to go TO" and the whole sentence: "Do you know where we're going TO?"

sid0• 15 years ago

To my (non-native) ears, that sounds overly formal.

andrewaylett• 15 years ago

"Do you know where we're going?" would be what most UK speakers would say, and "Do you know to where we are going?" would be an archaic variant which implies a location is requested rather than an action. The verb used is "to go"; the answer might be "[we are going] to London" or "[we are going] shopping".

There's some potential for confusion in that the verb "to go" can be used for the future imperfect (I think, I was only taught my tenses in French and German lessons) as well as the present tense: note the difference between "We are going to London" or "We are going shopping" and "We are going to go to London" or "We are going to go shopping", or even "We are going to shop". The former is talking about an action we are currently undertaking (to go: implying motion), the latter about a future intent (to go: expressing an intent to do something in the future) to carry out the action in the former case. Using a different verb: "We are riding a bicycle", compared to "we are going to ride a bicycle"

nraynaud3• 15 years ago

Thank, good one with the shopping.

tedunangst• 15 years ago

No, TO is a preposition, but not part of the verb. No more than "by car" would be part of a "to go by car" verb.

kragen2• 15 years ago

"Where" can substitute for phrases of the form "at PLACE" and "to PLACE" as well as bare placenames, so, "Where are we going? We're going to the market," is perfectly standard.

"Where are you at?" is considered nonstandard, but it (and the form "Where you at?" are standard in AAVE.

Ygor• 15 years ago

I am very good at failing standard captchas on many sites.

I always blamed Captchas as being to hard and not user friendly.

This title got me wondering - maybe it was by design, and I am an idiot.

mechanical_fish• 15 years ago

If you don't want to hear from anyone who isn't a pedantic, highly literate writer in English, that's your prerogative.

I, however, wish my Web had more input from illiterate people. There are a lot of them -- the majority of the world, an overwhelming majority when you restrict yourself to the English language. They have lives, thoughts, and stories too. And if it weren't for recorded music, oral historians, and the occasional documentary they'd be completely invisible in our media.

Of course, the average comment thread on the web is a terrible way to interact with the literate and the illiterate alike. ;)

tokenadult2• 15 years ago

The better idiot screen on any discussion forum with a well defined subject scope would be a factual knowledge test. But the factual knowledge test wouldn't maintain the existing community if the existing community already has "hivemind" about factual issues central to the forum's subject, contrary to fact. I wonder what we would all consider the subject scope of HN? Are there any issues on which the HN consensus about the facts of the real world might be contrary to "objective" fact?

erikb• 15 years ago

Also keeps (some of) the non-mothertounge speakers out.

kilburn2• 15 years ago

I am a non-native speaker, and to me it looks like the second captcha is wrong.

According to what they've taught me, it should be "Do you know _____ are we going?" instead of the proposed "Do you know _____ we are going?", shouldn't it (notice the "we are/are we" swapping)?

drv2• 15 years ago

No, "Do you know where we are going?" is correct; however, without the "Do you know", "Where are we going?" would be correct.

kilburn2• 15 years ago

Thanks, it totally makes sense now. "Where we are going" was an affirmative (sub)sentence because the interrogative is already expressed in the "Do you know" part.

kragen2• 15 years ago

The sentence already has a VS inversion: "Do you". It is not correct standard English to add a second VS inversion, such as "are we", although it does sometimes occur in nonstandard native speech. The "we are" is in a subordinate clause, and the subordinate clause is not structured for the interrogative.

boredguy82• 15 years ago

As is often the case with these "I'm so smart and other people aren't" posts, the author makes what I bet is an unintended mistake. I assume they want each option to have only one solution. The last example ("It is poor form to ____ you temper in a discussion group"), either option (lose or loose) works. The verb form of 'loose' means 'to set free', and temper has come to mean an agitated state of mind. So you can say, "It is poor form to lose (be deprived of) your temper (calm state of mind) in a discussion group."

But you can also say, "It is poor form to loose (let free from restraint) your temper (agitated state of mind) in a discussion group."

ZoFreX3• 15 years ago

I don't think anyone would ever use loose in that manner without preceding it with "let"? Feel free to correct me, but to me that sounds very wrong.

yid• 15 years ago

It's correct without "let", but a little archaic. You're unlikely to see that sort of usage outside slightly pretentious plays, but it's correct nonetheless.

defen2• 15 years ago

Turning and turning in the widening gyre

The falcon cannot hear the falconer;

Things fall apart; the centre cannot hold;

Mere anarchy is loosed upon the world,

The blood-dimmed tide is loosed, and everywhere

The ceremony of innocence is drowned;

The best lack all conviction, while the worst

Are full of passionate intensity.

neutronicus3• 15 years ago

One would loose one's temper \on\ a discussion group, not \in\ it. I enjoy the perspectives of the ESL folks on HN, though, so I don't want to see any filtering by grammar nitpicking.

boredguy82• 15 years ago

I like the ESL comment: that's cute. But to your broader point I ask, "Why?" If I can comment in an online discussion group or I can post on an online discussion group, why is the relation of my temper any different than the relation of my post?

The OED is useful here: it literally defines in as equal to on in the second definition of the prepositional form of 'in'. There's an extended history of the relation between 'in' and 'on', which includes the Latinate source of the Saxon term, its evolution in Old and Middle English, and its use today. Suffice it to say, the distinction is not a trivial one, and 'in' and 'on' are often interchangeable.

o1iver• 15 years ago

Not that that is a place where idiots are likely to venture, but I remember this "captcha" for the Arch Linux forums:

  What is the output of "date -u +%W$(uname)|sha256sum|sed 's/\W//g'"?

(https://bbs.archlinux.org/register.php)

kevinpet• 15 years ago

That's a nice one. A bit like "how many were going to St. Ives" or "what color are the bus driver's eyes."

EDIT: whoops, I interpreted it as 's/\S//g', which I assumed would discard everything.

emehrkay• 15 years ago

The page loaded a 500 and I thought that it was some sort of gag or that my understanding of HTTP response codes was lacking.

trotsky• 15 years ago

If Fox Nation implemented something like this, they’d have zero commenters (unless, of course, they made the wrong answers “right”).

I wholeheartedly disagree. I am pretty much a total idiot (based on many years of evidence) and yet I can easily use the words listed correctly. The fact that I espouse bizarre and unworkable methods of government (see my username) certainly doesn't prevent me from spelling correctly. In fact, unworkable and impractical ideas may be a significant indicator of post-secondary education.

I'm surprised to see such a thinly veiled insult directed at one set of political beliefs sit at #1 on HN.

Udo3• 15 years ago

  I'm surprised to see such a thinly veiled insult directed at one
  set of political beliefs sit at #1 on HN.*

Things tend to sit at #1 on HN because they're interesting and engaging. I didn't post the link because I wholeheartedly agree with it, and I suspect that's not the reason why people upvoted it either.

That being said, I think the not-so-thinly veiled insult directed at the average Tea Party clientele is backed up by casual observation. And that's saying something, coming from a foreigner who has no stake in the US party system (=me).

  I am pretty much a total idiot (based on many years of evidence) 
  and yet I can easily use the words listed correctly.

You're obviously not an idiot. You can articulate yourself, bring points across in an eloquent manner, thus enriching the conversation. Compare that to someone who only communicates in lulz or derrrp-speak. The point of having an online discussion is not to surround yourself with people who agree with you 100% of the time, it's about having a stimulating discussion in the first place.

dabent• 15 years ago

This is an interesting concept, but won't keep out comments from the reasonably educated, but devious. Those often are the worst kind. It's usually possible to simply down vote puerile comments into oblivion, but one intelligent troll can waste considerable bandwidth.

steins• 15 years ago

those would be pretty easy to solve with a bit of code

__rkaup__• 15 years ago

I don't get it. How come this article has not been flagged to death? How come this article has over 100 comments? Am I missing something?

Uchikoma• 15 years ago

(non native speaker)

In 20% of my HN comments I need correct "there"/"their", "buy"/"by" [...] which I write down correctly when re-reading my comments, but subconsciously write incorrect the first time when I do not think about the words but the thoughts.

roser1372• 15 years ago

interesting...

roser1372• 15 years ago

I think, This seems to weaken the captcha for bots, since a simple Google search for (for instance) 'do you know we are going' completes the sentence. No need to defeat the obsfucated OCR test.

tomelders2• 15 years ago

Hands up if you were especially careful when commenting on this post?

sedachv• 15 years ago

This IMO is something that Disqus or Quora or similar systems should jump on - automated techniques for knowledge base quality management. I can see taking some of the existing essay grading software, combined with something like these captchas, being used to produce customer support knowledge bases from input from technicians and discussion forums.

joeyespo• 15 years ago

The problem is there's a good chance of measuring the wrong thing.

If you need to pass a test to post, sure you'll keep out a lot of bad comments. But you'll also limit the pool of acceptance. There could be plenty of people who fail the test having something positive to add. Standardized tests prove this. There are those who are lousy test takers but end up being very successful. Also, certain topics might be easy to answer with someone with sufficient experience in the subject without requiring even average intelligence.

So this limitation might increase the quality of posts, but also adds a kind of tunnel vision to the site.

tokenadult2• 15 years ago

Does noticing whether a link is the original link

http://www.buzzfeed.com/chrismenning/captchas-for-keeping-id...

or blogspam count as a useful screen?

ambler0• 15 years ago

Unfortunately, grammar captchas like this would keep out a lot of second language learners as well.

Revisor• 15 years ago

Via via via via...

The original source is here

http://www.defectiveyeti.com/iacaptchas/

chc• 15 years ago

The best Captcha IMO is just to write up a set of questions about the article. Actually reading the whole article and absorbing the details is a reasonably good predictor of comment quality.

Jupe• 15 years ago

At first take this seems like it would be easy to circumvent, (given a dictionary of homonyms) but...

google "look over there." == 1.5M hits

google "look over their." == 16.2M hits

google "look over they're." == 1.0M hits

Google doesn't support periods, apparently. But, other examples from the article are a little more easy to game:

google "do you know were we are going?" = 8 hits

google "do you know we're we are going?" = 0 hits

google "do you know where we are going?" = 6.1M hits

jhuckestein• 15 years ago

Do yourself a favor and read the comment thread on that post.

Jach2• 15 years ago

I know ESR isn't really popular here, but:

"Being a native English-speaker does not guarantee that you have language skills good enough to function as a hacker. If your writing is semi-literate, ungrammatical, and riddled with misspellings, many hackers (including myself) will tend to ignore you. While sloppy writing does not invariably mean sloppy thinking, we've generally found the correlation to be strong and we have no use for sloppy thinkers. If you can't yet write competently, learn to."

I still try to read through poor writing, but the comment has to contain something fairly insightful for me to get over a "then/than" mistake. That said I think a Naive Bayes classifier for "sloppy" and "not sloppy" would work better than this captcha system. Not to mention it's super-easy to get around by a bot, and getting rid of bots is the point of a captcha in the first place.

Luyt3• 15 years ago

"If you can't yet write competently, learn to."

Excellent advice.

Ayn Rand wrote: "If you don't know, the thing to do is not to get scared, but to learn."

spc476• 15 years ago

I liked the method the USENET group alt.hackers uses---it's moderated, but there's no moderator. Good luck in posting.

jgeerts• 15 years ago

The funny thing is that the smart captchas are easier for a computer to guess (if it wasn't a joke).

Too bad that the current implementation of captchas ever saw the light, they are pushing a security issue to the end user. Seems like a good opportunity :).

The acronym is "Completely Automated Public Turing test to tell Computers and Humans Apart" but it isn't really that automated, so do captchas actually exist?

Captchas To Keep Idiots Out Of Comment Threads

Related Stories