META: Tiered Discussions

post by Pavitra · 2010-10-30T22:37:46.767Z · LW · GW · Legacy · 20 comments

Contents

20 comments

(Edit: It seems lots of people thought this was a terrible idea. I'm keeping the post as it was, though, mostly because I still think it's an interesting experiment and it ought to have been tried at least once somewhere on this site. Also, blah blah something about preserving the historical record so that earlier comments still make sense, whatever.)

You aren't allowed to know what this post says unless you can figure out what LW post this sentence is a clever reference to. The URL of that post is the CAST5 symmetric key for this one. Please help downvote spoilers into oblivion.

-----BEGIN PGP MESSAGE-----
Version: GnuPG v1.4.10 (GNU/Linux)
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=sPUV
-----END PGP MESSAGE-----

20 comments

Comments sorted by top scores.

comment by AngryParsley · 2010-10-30T23:31:49.864Z · LW(p) · GW(p)

I spent more time figuring out gpg than guessing the password. Now to save everyone else the time and effort.

SPOILER ALERT

The password is http://lesswrong.com/lw/p0/to_spread_science_keep_it_secret/ and the decrypted message is:

As has been discussed elsewhere, Less Wrong is currently fairly unapproachable to newbies. In particular, the learning curve is ridiculously steep. People who have read a few posts are told, in effect, to read the entire archives [TVTropes] before commenting.

Suppose there were a separate discussion area dedicated specifically to those who are just starting out, or perhaps who have taken a few of the first steps down the road. When posting there, it would be considered poor etiquette to expect others to have read later material, like expecting fourth-graders to know trigonometry, or high school graduates to be able to read Feynman diagrams. Ideally there would be a series of tiers of discussion, so that people can learn the material interactively, like in a classroom, rather than doing the equivalent of shutting themselves in a closet with a textbook.

In order to motivate people, of course, we should lock up the higher-level discussion areas behind tests of merit. We should also think of some sort of standard miracles people can do to publicly demonstrate their Level Six superpowers.

Secret discussion areas don't scale. All you need is one defector (me) and everyone knows the secrets. A newbie area or a "let's read the sequences" thread would be handy. A delayed RSS feed of the sequences would also be nice.

Replies from: Pavitra
comment by Pavitra · 2010-10-31T00:26:04.254Z · LW(p) · GW(p)

Hopefully community norms can be upheld with a thorough karma beating; downvoted. (This requires, of course, that the community in fact disapproves of spoilers.)

Srsly tho, you could at least rot13 that.

(This comment is now at -3. Point taken.)

Replies from: NihilCredo, Alicorn
comment by NihilCredo · 2010-10-31T01:54:55.127Z · LW(p) · GW(p)

It is a fact of life on the Internet that there will always be spoilers available to any puzzle or story that someone has finished.

Trying to stop them from appearing is foolish; all you can do is put forth your best effort to encourage readers to refrain from scrolling down (and for this particular post your encouragement was quite weak, to put it mildly).

Replies from: Pavitra
comment by Pavitra · 2010-10-31T06:41:32.225Z · LW(p) · GW(p)

In a "real" implementation, what's on the other side of the door would be not just a static text that can be reposted, but an ongoing living discussion. Then the locks can just be changed.

And yes, I did it badly. I might try to fix the implementation at some point. Do you think that should be an edit to this post, or an entirely new post, or the idea outright abandoned?

comment by Alicorn · 2010-10-31T00:51:20.769Z · LW(p) · GW(p)

It's hardly a spoiler. You are not telling an exciting story with a twist ending. You're just suggesting to lock up parts of the community and test people before they get in. At worst, it's a puzzle solution, and not to a particularly interesting puzzle.

comment by AngryParsley · 2010-10-30T23:11:27.435Z · LW(p) · GW(p)

Not all of us know how to use gpg. I think the right way to decrypt that message is to run...

gpg --output decrypted.txt --decrypt input.txt

...and then enter the password. Is this correct?

Edit: Yep, that's it.

comment by Pavitra · 2010-10-31T14:59:40.858Z · LW(p) · GW(p)

I'm preparing to rewrite the top post completely. To that end, I'm archiving the original version here.


You aren't allowed to know what this post says unless you can figure out what LW post this sentence is a clever reference to. The URL of that post is the CAST5 symmetric key for this one. Please help downvote spoilers into oblivion.

-----BEGIN PGP MESSAGE-----
Version: GnuPG v1.4.10 (GNU/Linux)
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=sPUV
-----END PGP MESSAGE-----
comment by Morendil · 2010-10-30T23:23:44.906Z · LW(p) · GW(p)

A data point on the effectiveness of this challenge: it took me about four minutes and two tries to get the passphrase. (My first try was too clever and used the URL to a post which has been censored.) I've been on LW for a little over a year.

It's hard to tell what you're expecting people to say in this discussion, given your spoiler downvote request. Any sensible response to the encrypted content is going to be more or less of a spoiler.

In the spirit of having fun, though, you can get at the (probably disappointing) remainder of this reply by using the full name of the writer who is the target of the allusion in the seventh paragraph of the post the URL of which is the above's passphrase.

-----BEGIN PGP MESSAGE----- Version: GnuPG v1.4.8 (Darwin)

jA0EAwMCC6+CgqSHk/5gycCAbdqenzgWRTAMf6txIWGx9XjBlZ2KTFJKuSy5HhZN HdryjGegrR0lbYJzd7V3pgQ1RFRrsgzEH7JE+r5g3q+nufHIr33BkEj4VoDui56d 2SmdfIthdjYNvk1IM4H6N8zegsBSlptB2BYWv7xBE3MzekqDRO5QBkP1mDar0ca2 zx9bfU5L4XtMcgNc3rVGpyRWARDZkzluozVQwwLuTmeNrjpVubtbUKNlaoy+LgiR BHuRC5We6NBZR1mc8Q/TIhDltW2y4G81pzChff0yJc6YkE7uasMgmmJC9qtu7VSW wcPkVgLOoQrXQQmHKMb3eOrr9iAyPQlYaRgPLcOazloQj7CeCqAFm6zQrNoZrUUH aX1Sm4nS6gQK7fa5hS8Lw/4oT/CQJrYj33dWtseial8phYW5U7rGe8i8qyK5G136 5l8= =jgi1

-----END PGP MESSAGE-----

Replies from: Pavitra
comment by Pavitra · 2010-10-31T06:57:36.231Z · LW(p) · GW(p)

Since apparently GPG is too much of an annoyance, and serious security isn't really going to work, I'll try a Vigenère cipher for this comment. I'll also try to make the password more closely dependent on the real dependencies of the discussion, as an extension of the exercise.

Here is the clue: After reading the article that is the original post's key, you will have heard of two things that are needed for an idea to be taken seriously. The second of the two is Masks; the use of codes fulfills this function. The first of the two is not science-as-attire, but rather Science's attire: the rightful dress of those who genuinely possess the Hidden Truth. This first is the Vigenère key.

Dm qsaeh jw lyou tsihjgagoumge wo xzv rjwulgtmge wt edisbhq sosvatoeiv sm b rgehsmnzom mftcozwewfruv: hiel ft sisuwok lys Tiilsogwj. Xvwl rg xmly gdmweqf, xzv sgjgih siilwsiv kc pzwiqpqw kvf kweijrw uwgjatimxq too fw dcumnrhfh tp abrmwodxmiwok s tcowhzqvsmj, hbrlrzjdaeu, mikjss hawtjgmchz.

Sx tcvvkv kf aglzer'l noox lf vbzw jsqejrhf tsjgxsjug brv iwehdvg pr wrqi tgjh -- nyuy zfwk vodl ufanifk -- fbxzvf, xi'v noox s jwokdv zpgc fb feuy hjij'j tpvmd.

comment by Alicorn · 2010-10-30T23:51:03.045Z · LW(p) · GW(p)

This is the first post I've laid eyes on since gaining mod powers that I considered banning. I just downvoted, instead, but wow, do not want.

Replies from: Pavitra
comment by Pavitra · 2010-10-31T00:24:46.916Z · LW(p) · GW(p)

The GPG thing is a not-very-good workaround to the fact that what I'm trying to do isn't supported by the software.

Replies from: Alicorn
comment by Alicorn · 2010-10-31T00:44:56.306Z · LW(p) · GW(p)

I fail to see that you are trying to do anything which is valuable, or which would be valuable with software support.

Replies from: Pavitra
comment by Pavitra · 2010-10-31T02:59:08.738Z · LW(p) · GW(p)

Would you be willing to specifically address the arguments I made in the post regarding why I think it would be valuable?

Replies from: Alicorn
comment by Alicorn · 2010-10-31T04:01:27.070Z · LW(p) · GW(p)

Did you make arguments? I only read the helpful translation, so it's possible that it's missing an argument part. The helpful translation seems to contain only a problem statement, followed without support by a proposal phrased as a statement of prudence.

Replies from: Pavitra
comment by Pavitra · 2010-10-31T06:21:04.804Z · LW(p) · GW(p)

The translation is complete.

I must have expected an inappropriately short inferential distance, because I thought I was presenting a well-reasoned argument, but I can't now find any specific section of text that constitutes reasoning.

Could you specify which of the following claims you disagree with? (Spoilers rot13'd, since the text of the post is now available without significant effort to those who are interested, and I still think the barrier should at least be available to those who want it.)

Replies from: Alicorn, Relsqui
comment by Alicorn · 2010-10-31T13:34:53.451Z · LW(p) · GW(p)

Item Two: False assumption. Newbies aren't in fact obliged to read the entire sequences before commenting. They can comment about things like their personal experiences, their areas of instrumental expertise, and matters of opinion. They can request citations for statistics, correct many types of factual errors, issue praise, introduce themselves, make jokes, and talk about locally popular fiction and fanfiction. Any of these types of comments can be valuable, can garner upvotes, and does not need to be suppressed.

The type of newbie comment that is problematic is uncommon compared to the unproblematic newbie comment types. When problematic newbies say annoying things that really would be solved by their reading the sequences, they come in one of two types: one, they only needed a link to the correct sequence, and will promptly stop making problematic comments when so linked. These are fairly uncommon; LW has a lot of obvious conceptual house rules, and most newbies seem to stay out of the deep stuff on their own. The second type is a borderline troll or worse, and while such trolls might be kept out of part of the community by the puzzle you describe if you could control solution leak, downvoting and (in a pinch) banning also do the trick without filtering out commenters who, like me, have no interest in solving silly puzzles you make up.

Item three suffers a similar assumption.

Four: Assumes that people make the same inferential steps from concept to concept. I think this is false. Also assumes that people interested enough in LW to try to comment are the "general population"; I think this is also false.

Five: Of course the idea is neat. But that doesn't mean it is a prudent idea, or that it should be imposed on everyone. As a quick-and-dirty test, I would have abandoned the site on Stumblingupon it if this sort of system was in place (and solutions weren't leaked). Assuming for the sake of argument that I'm a valuable commenter, we should tread carefully when considering the introduction of filters that would have driven me away.

Six and seven suffer similar problems.

Replies from: Pavitra
comment by Pavitra · 2010-10-31T14:55:42.431Z · LW(p) · GW(p)

Two: I had offered these links in the anticipation that they would be taken as sufficient evidence that newbies are in fact obliged to read the entire sequences before commenting, and it was from there that I derived my beliefs that there was a problem that needed solving. Since you're a mod, it seems reasonable that you would have strong first-hand knowledge of the balance of good and bad comments; but on the other hand those links contain even more tightly first-hand assertions from newbies that something suboptimal is going on. I'm not sure what to believe at this point, which means it's probably not a good thing that my brain is still trying to propose solutions.

Three: I don't have real evidence for this one; I had imagined that newbies are turned off by an apparent social pressure to have read the sequences, so if Two is out and you say it doesn't work like that then this is likely wrong.

Four: I'd assumed, roughly, one inferential step per post. It's already discretized, posts have globally-unique names (at worst, we can use the URL), and the dependency graph is explicit. And even if there are in fact several inferential steps for a given person in a given post, it's a reasonable amount of reading to ask someone to absorb one post.

LW-baseline being higher than humanity-baseline just means we don't need as many tiers.

Five through Seven: I said before I shouldn't still be proposing solutions, but...

It bugs me that the dependency pointers are only listed in one direction. I want to be able to start at the beginning of the Sequences and chase pointers to the end. And, wherever I am in the Sequences, it would be nice to have a discussion area for that stage, where I can quickly find questions from other people who might have had the same questions I have. (The comments don't really work for this -- commenting on a blog post from 2007 or something always feels like thread necromancy.)

And, as long as we're setting up a big separate additional optional system for reading through, we might as well structure it as a conspiracy rather than a university, assuming the Do Not Want problem discussed under Five can be solved. (Conspiracies are fun, right?)

The first tier, for people who are reading the dependencyless posts, would of course be publicly accessible. And people who just fundamentally DNW can just continue using the regular LW that currently exists.

And I'm rationalizing aren't I. I have two separate desires: I want a neat fun conspiracy with fearsome rituals, and I want to chase pointers from bottom to top with individual discussion areas. I have no particular reason to believe that either of these two things would benefit from being combined with the other.

Replies from: Relsqui
comment by Relsqui · 2010-11-02T07:13:46.329Z · LW(p) · GW(p)

I had offered these links in the anticipation that they would be taken as sufficient evidence that newbies are in fact obliged to read the entire sequences before commenting

I think you're using "obliged" to mean encouraged/pushed, and Alicorn is using it to mean required-before-being useful. The former is accurate, but to a variable degree; the latter, as she demonstrated, is not.

I had imagined that newbies are turned off by an apparent social pressure to have read the sequences

Yes. and being restricted to a newbie-only (i.e. devoid of useful content!) area until we could prove we had done so would not be less of a turnoff. :P

Replies from: Pavitra
comment by Pavitra · 2010-11-03T00:55:53.815Z · LW(p) · GW(p)

I've since come to think of the tiered discussions as being in addition to what currently exists, rather than instead of it.

But, uh,

... you used a ciphered message to propose that the site should be more accessible?

yeah, that's a good point.

comment by Relsqui · 2010-11-02T07:08:20.908Z · LW(p) · GW(p)

... you used a ciphered message to propose that the site should be more accessible?