## LtU: Style and Focus

LtU is primarily a weblog dedicated to posting links to interesting projects, research papers, and programming languages related news, by contributing editors who can post to the home page - items intended to be discussed in the comments section.

Editors (as opposed to guest bloggers who are asked to join us because of their contributions to the field) are encouraged not to post about their own work, nor to post rants and opinions to the home page.

This allowed LtU to maintain a very high standard of quality for the news items posted to the home page, and resulted in deep and informed discussions.

As LtU grew, the attached forum became more and more active. This level of activity, while exhilarating, poses some predictable problems. Too many of the problems that plague other forums related to programming languages, such as religious wars about typing, Microsoft bashing, and possible trolls are becoming a real issue for LtU. Newcomers, attemtping to learn their way into the field, will always be welcome on LtU. Trolls abuse this by asking "naive" questions, and arguing positions that only serve to confuse rather than inform.

A few LtU regulars voiced their concern. I agree: I have been following the discussion group much less closely myself, recently, because of these issues.

LtU wasn't intended to be the only web forum on programming languages, nor do I intend to manage a multitude of forums for all tastes and styles. LtU is intended for the discussion of items thought to be news worthy by the contributing editors. Short topical messages to the forum were ok, as were off topic messages, when LtU was small and not well known. It seems that we as a community need to decide how to keep the original style and commitment to civilized, informed and professional discussion.

Some of the suggestions I received regarding possible modifications to the site and its policies:

1. Closing commenting entirely. I think this is too harsh, since the discussions often contain great ideas and suggestions.

2. Eliminating the option to create new threads in the forum, thus only allowing contributing editors to set the agenda, so to speak.

3. Establishing some sort of comment moderation. I think this is too problematic and requires to much work from the volunteers running the site (including myself).

4. Banning users who use LtU to advertise, post trolls etc. I don't like banning, but I also don't want trolls to alienate long time members. We aren't sure whether there are (non-troll) readers who feel strongly that we should not ban people. How do we decide on policy, as regards banning?

5. Eliminating user signatures (which are appended to all their posts), since they seem to mostly add noise.

My hope, originally, was that the norms of behaviour here will be enough to ensure newcomers understand and maintain the LtU style and unwritten code of behaviour. Perhaps waiting for this to happen is still the best option. We may decide to not let trolls and such dictate our policies, deciding that we can live with them even though they are irritating.

I'd like to hear your opinion. Especially if you are a long time reader and contributor.

Some rules of behaviour are obvious, yet seem to require reiterating and strengthening: personal slurs and attacks and advertisements will not be tolerated. They will be deleted without warning, and repeat offenders will be banned from the site.

## Comment viewing options

### comment moderation

LtU uses Drupal, right? I'm used to http://br-linux.org, which also uses Drupal with a very effective moderation system. Let LtU readers themselves moderate bashers and other annoying posters.

Banning users is likely to just make them come back to haunt us under other nicks...

### A Group Is Its Own Worst Enem

A Group Is Its Own Worst Enemy

"The likelihood that any unmoderated group will eventually get into a flame-war about whether or not to have a moderator approaches one as time increases."

### Contributing Editors

I think it would be a bad idea to allow only contributing editors to set the agenda, especialy with your (Ehud's) idea that contributing editors should be prolific.

It leaves only two options:
1. Don't contribute stories.
2. Contribute all the time.

I think there's a middle ground between these that is very valuable.

### Huh?

I fail to see any reason for the "only two options" you mention. Mayn editors contribute only occasionaly.

### Then I'm wrong

I either misunderstand what is expected of contributing editors, or I misunderstand what a contributing editor is.

Is a contributing editor one who can post to the front page?

I probably just misunderstood what is expected; I don't know first hand, as I never pay attention to who posted a story.

### Contributing Editors

Is a contributing editor one who can post to the front page?

That's essentially it. They are usually regulars and long time readers as well, naturally.

### If it can't be expressed in figures, it is not science

"If it can't be expressed in figures, it is not science; it is opinion. --Lazarus Long"

Talk is cheap. Claims that include links to executable code are worth investigating.
That and E-Prime usually make computer science discussions easier to deal with.

The rules for #haskell are "Be nice. (or else)" When people fall outside of nice the first time, I privately explain carefully to them what the community expects and how the person's actions fell outside of those expectations. The second time I say the same thing publically. The third time, I remove the person from the channel.

I don't have time to waste on difficult people, there's so much cool stuff to learn!

Also, I've removed my signature. I'll keep that sort of information on my user information page in the future.

### More suggestions

If you mainly worry about the website getting filled up with troll comments, maybe running a usenet server at LtU could help. Usenet is the ideal medium for discussions with trolls ;)

If you mainly worry about the quality of discussions, maybe some aristocratic system would be good: users have ratings, and users can moderate users (not comments) up and down. So after a while I think that the regulars that post "good" comments would have a good rating, while trolls don't. By marking every comment with a color (shades of green or yellow, maybe?) or with a numbers of little stars next to it users could quickly skim comments for the good ones (or set a level they want to browse at, like on Slashdot).

Well, I certainly LtU stays as good at it has been for the last years...

### mod_virgule?

There is even code available from the community site http://advogato.org/ which implements this kind of trust metric.

### Edit Queue

I know this site uses Drupal, but perhaps a queue system the one used on Kuro5hin.org could be implemented for the forums.

Kuro5hin uses the Scoop engine. The basic idea is that all registered members can vote on new topics before they are allowed to bubble up to the main forum.

They initially exist in an editors queue where people can comment on style, content and grammar, etc.

When enough votes are cast in the editors queue to allow the publishing of a topic the post finally achieves public viewing.

Kuro5hin allows for this queue to work for even front page stories, there's no reason that you couldn't eliminate that feature.

Since all registered users are allowed input to this queue, it is only discriminatory to anonymous users.

I think it would provide a good balance. The Editors drive the front page and the day to day content of LtU. The users manage the forums, preventing the Editors from burning out moderating every little topic that arises.

### Re: Edit Queue

Drupal has support for a moderated edit queue. I agree that it could make a lot of sense to use that approach for forum topics.

### Clarification

I have tried to be nice and productive, but I fear that I have also posted questions and ideas to the forum that aren't in the spirit of what this site's community would like to see... could you post some guidelines for what is appropriate for LtU forums, and what I should try to take elsewhere? This is a great place with thoughful and intelligent readers and contributors... I don't want to abuse it.

### With an attitude like that...

...I'd say that you fall outside of the range of a recidivistic offender. :-)

### Agreed

I also agree that "some guidelines for what is appropriate for LtU forums" are needed, beyond what's in the FAQ. We can't really fault newer members for not instantly divining the site's gestalt.

### Guidelines/FAQ

I also agree that "some guidelines for what is appropriate for LtU forums" are needed, beyond what's in the FAQ. We can't really fault newer members for not instantly divining the site's gestalt.

Does anyone really think the problem is with people who don't understand the site's "gestalt"? It seems to me this happen rarely, and never causes more than a slight bu,p (i.e., one or two OT messages). The problem I see is with people who try to be disruptive on purpose.

As for guidelines/FAQ, this sounds like a good idea, but I am at a loss as to what to include in one. To give a specific example (I apologize if this sounds too harsh): Do you think this exchange is explained by the norms being unknown or misunderstood? If the comment is genuine, it should be discussed (the issue involved isn't advanced, but an explanation is appropriate), if not then the problem won't be solved by guidelines.

Guidelines can tell normative users when to ignore threads and not answer, but I don't think this is what you had in mind.

Anyway, if someone wants to come up with a FAQ or has specific suggestions about what to include in one, I'd appreciate the help.

### Re: Guidelines/FAQ

I didn't intend to suggest that the more serious problems would be solved by more detailed guidelines. I was more responding to letdinosaursdie's concern that he was abusing the site, and saying that in the absence of better guidelines, "innocent" violations of the unwritten rules are to be expected.

Regarding the response you linked to, I agree that the usual polite approaches aren't going to help.

Guidelines can tell normative users when to ignore threads and not answer, but I don't think this is what you had in mind.

I think every little bit might help.

Anyway, if someone wants to come up with a FAQ or has specific suggestions about what to include in one, I'd appreciate the help.

I'll give it some thought.

### Yes, a prominent policy statement, please

A newcomer has no idea that LtU is intended "only for the expert stuff." On my first forays, it looked like any other FP forum. Complete with a wide range of expertise, and the usual oscillation points that aren't terribly interesting. In other forums this is not considered Bad / Evil [TM], just the product of open public debate. It is clear that the forum admin doesn't want that kind of debate, that he wants to exercise some kind of firm editorial policy. Also clear, is that others share the sentiment. That's neither good nor bad; it's "what is," and I can't fault a community for attempting to design itself. I just hope the community takes responsibility for designing itself and does not seek scapegoats.

I would suggest doing the hard work of manual, all-posts-must-be-approved moderation. You don't seem to want huge traffic flow anyways, and "hard moderation" is an excellent squelch tool. It allows busy professionals to fit forum participation into their otherwise hectic lives. I say this as one who owns several mailing lists and has used various moderation policies at various times.

You may find that with hard moderation, the number of posts diminish enough that "seniority" and "5 posts a year" systems become unnecessary. But, of course consider those more radical proposals if you want to (re?)engender a cult of expertise.

### LtU is not a debate forum

It is clear that the forum admin doesn't want that kind of debate

The intent of the forum owners (both the admin and those of us who feel like defacto owners) is not to create a debating forum for advocacy, posturing, attacks, vendettas or advertising. As one who engaged in competitive forensics for some 10 years, I've had bit of experience with debates.

There is a difference between the concept of a debate versus the concept of exchanging ideas. Exchanging ideas may result in arguments (some of which can be passionate), but that is still very much different than the purpose of a debate. An exchange of ideas involves two or more people that are sharing their thoughts to either communicate their ideas, or to discern what another has to say. A debate involves at least three parties: the two advocates and a target audience (judge, jury, popular opinion). The two advocates talk past each other - and indeed have hostile intent and are more than willing to subvert their opponent. (Crush the weak is the motto of most debaters).

### Like any other FP forum

Don't know where you've been to, but when I first saw this site, it scared me away with all the theoretical stuff. When I stumbled over another link to here maybe 1-2 years later, I had the courage to stay :)

Only lately do I see tons of articles involving C dialects, Python and Ruby...

### "FP forum"

I think it's worth mentioning that LtU is not specifically intended to be an "FP forum". However, it is intended to have a strong connection to and grounding in programming language research. Of course, a majority of current PL research leans strongly in the functional direction, for fairly obvious reasons related to the connection to various mathematical subjects. (If someone wants to dispute that, start a new forum thread ;)

The FAQ says:
"This weblog is dedicated to the study of general properties of programming languages ... Mostly this site deals with issues directly related to programming languages, and programming language research."

Although the natural overlap between FP and PL research does make for a fuzzy boundary between the two, there's still an intended difference in focus. The FAQ page has a bit more on the subject.

### overgeneralization

Let's not overgeneralize from what is (as I take it) a problem with one notorious troll. Would a specialized solution work best?

It's surely not sufficient, but I'd guess it would be helpful to eliminate threads, which always seem to be an invitation for discussions to wander off topic. If new topics arise that would best be discussed in a new space, one can always start a forum topic.

Having a detailed FAQ that links to expectations of user conduct and itemises issues where we have agreed to differ is likely helpful: perhaps it would best be a wiki? We could discourage wiki-sprawl by restricting the ability to create new pages, or some such.

Having recently finished a study on trust & reputation systems, this topic is very timely.

I've been a reader for some months now, but this is my first post. Go easy on me :-)

It has been shown [1,2] that the optimal strategy for enforcing best behavior among a community with cheap pseudonyms is the "Pay Your Dues" strategy. Essentially, this means newcomers must 'pay' something: time, money, 'work', in the context of the community before becoming full members.

Trolls can't be trolls if they have to pay to create each new pseudonym - assuming existing troll pseudonyms are blocked.

Several techniques have stemmed from this strategy [1,2,3,+]:
- Pay  for membership.
- First X messages are moderated.
- New members must wait X days/weeks before posting.
- Messages from members newer than you are subdued/hidden (as are any replies to them).
- New members are restricted to X replies per day/week/etc.

Some other interesting findings in my research [2,3]:
- Binary (good/bad) ratings are optimal.
- When feedback is missing, "good" is the optimal default choice.
- When cheap pseudonyms are available, all unknown (ie: new) members should be treated as a single, anonymous entity. All ratings of an unknown member (until they have paid their dues) also rate the anonymous identity. This causes trolls to reduce their own effectiveness over time.

Back to the options presented, I think option 5 is a good one to get started. Option 4 to some degree (for the worst offenders). Then some form of "pay your dues" for new members before their replies and forum posts become publicly visible.

### Answering the question in the

Answering the question in the post, I generally agree with Bryan. I would start with option 5, and option 4 might be necessary.

As a "pay your dues", I don't know about actually paying money, but I agree that a moderating system where the first X messages go through a moderator should avoid abuse, while limiting the load on the moderators.

### keeping out trolls

We just need a CAPTCHA-like system, where you have to solve an exercise from SICP or CTM before you can post. :-)

As for the various other suggestions, I support moderating new forum threads by non-editors (but not moderating all comments), banning trolls, and killing off signatures.

### Editors make the site

To the degree that the tone of a site that takes submissions from the public does and must reflect editorial policy, I don't think it makes sense to devolve power away from the editors. To this end, I would allow the editors to delete any reply or freeze any forum thread. Sure, there can be abuses: editors who are quick on the trigger should get a lecture and told to take a vacation, and those abuse power should no longer be editors, and there may be the occasional gripe about it. On the other extreme, there's little value in letting the inmates run the asylum just to guard against the bogeyman of an "editorial cabal". I value the content of this forum more than some abstract egalitarian ideal, and I'm perfectly fine with this forum belonging to a select group of editors, and not to all of us.

### Overkill

If the feeling is that the problem is restricted to a couple of users, it seems to me we should handle that directly and not change the entire way LtU works. No?

### Sounds ok to me

I think a few posts by the editors/threadstarters to keep it on topic and the tone down should suffice without any serious new infrastructure.

Nowadays it's pretty easy to dive into the comments without even reading the article being discussed. It's up to all of us to keep it on topic. On the other hand, off-topic comments also have value and it's again up to us to recognize our off-topic comment and turn it into a separate thread if necessary.

We must recognize that the papers/works that are being talked about are not easy reading. Many threads are advancing much faster than I thought they would. I'd expect a lot more questions of the sort "What does the symbol on Page X mean?", "Is it me or is his example on page Y really broken?", etc. I think it'd be nice if we all could slow down a bit and maybe even look at older threads.

### Sure, why not

We must recognize that the papers/works that are being talked about are not easy reading. Many threads are advancing much faster than I thought they would. I'd expect a lot more questions of the sort "What does the symbol on Page X mean?", "Is it me or is his example on page Y really broken?", etc. I think it'd be nice if we all could slow down a bit and maybe even look at older threads.

True. I tried to start such threads a couple of times. These aren't the things that bother me. Asking questions of this sort about the papers mentioned is a very good idea.

### Not sure how much of a problem there is

I find LtU refreshing flame- and troll-free. Of course, I hang out on Slashdot and DailyKos on a regular basis, so I may have different standards.

I would like to comment that there is a danger in going the other direction (too much moderation) as well. I've seen online communities get to the point where any topic more controversial than "my cat Fluffy" stories got shut down out of a fear of flame wars. I'm not saying that LtU is anywhere near there (because it isn't)- I'm commenting that there is such a thing as too much moderation, just as there is such a thing as too little. There are people, and I've met them, who feel saying to them "No, you're wrong, and here's why" is a personal attack. Letting these people set the standards for moderation is as bad as letting the trolls set the standards.

Another comment I will make is that while compile time vr.s run time type checking, or ocaml vr.s haskell vr.s python vr.s lisp arguments are boring and repetitive to the old timers, they aren't to the newbies. I myself have learned a hell of a lot in the go-arounds I've seen on this site. Perhaps instead of banning them, they should be "ghetto-ized"- create semi-permanent "advocacy" groups (like the comp.lang.*.advocacy goups of usenet of yore) where these arguments can go on to the participants heart's delight, and those who don't want to read them don't have to.

Just my two cents.

### boring universe?

What I want to see is interesting code and ideas in languages that other LtU readers know and I don't. I'm lurking for snippets of Smalltalk, Factor, J, etc. I don't really understand if that sort of things is welcome on LtU or if it's not "meta" enough.

What do the rest of you want to see here that we don't have lately?

### Sure

I like the cool examples too. But the reason this is ok isn't because "Ehud says so". If Luke, Chris, Anton or any other genuine member of the community (I won't list the > 1000 members here :) posts something directly related to PLs, fostering informed and meaningful discussion - it's welcome on LtU. That's how it has always been.

What do the rest of you want to see here that we don't have lately?

It's not what's lacking. It's what we have too much of: silly trolls, flames etc.

All the messgaes here about being accomodating to all levels of discussion and all aspects of the the field are missing the point: we've always been open, and want to remain so. I've written about the vision many time... The question before us is how to save this in the face of some posters who undermine this and want to destroy what we have here.

### "PLs"?

something directly related to PLs,

I don't understand what this means. I'm interested in lots of individual programming languages ("PLs" plural) but not in programming language theory ("PLs" meta). There is much more meta and less plural on LtU this past year or two.

For example, is a small joke in Emacs Lisp directly related to PLs?

(defun activate-peril-sensitive-sunglasses ()
"Instant protection from the sight of ugly code.
\(With apologies to Zaphod Beeblebrox.)"
(interactive)
(mapc (lambda (face)
(set-face-foreground face "black")
(set-face-background face "black"))
(face-list)))


### hahah

Good jokes are always welcome, i guess. Even in bad languages... ;)

### Post limiting and seniority

How about giving new members a maximum of 5 comments for the first year, and not until they have read at least twenty posts so they get a better feel for the tone of the forums?

In addition, there are newbie questions which may be quite appropriate. We can't be everything to everybody, so perhaps moderation of new topics is warranted. Any suitable alternative outlets?

is so low that it will drive people away before they ever give LtU a chance.

### It is not the quanity, it is the quality

I don't worry about users who post many quality comments. I worry about the users who post only junk that wastes my time.

I would agree that the first 5 posts of all users should be flagged, and an editor should review them ASAP, but once a user proves mostly good there is no need to continue to review their comments. Within a few posts you can tell if someone is trying to contribute something useful, vs trying to disrupt things.

A user who proves to be mostly good can be allowed some discretions where they cross the line. Everyone makes mistakes, so long as the mistake isn't too aweful we can forgive them for going over the line once in a while.

This though assumes that the moderators are unbiased. Generally not a problem on a place like LtU, but even here there could be problems. We don't want some Ruby fan would try to get rid of python fans, or vise-versa. (Or is it FP fans trying to get rid of OO fans?)

### Thoughts

On one hand, there are a couple of users I've noticed recently (I can think of two of them) who have engaged in what I might consider to be trollish behavior.

OTOH, often times a "troll" is a label put upon anyone who persists in advocating unpopular opinions (even if offered in "good faith"), rather than someone who seeks only to disrupt the site by seeding controversy. It's often hard to tell the difference. A person who were to post messages like "Haskell sucks" (but said more elegantly) would likely be labelled a troll; but someone who posts "C++ sucks" often gets a pat on the back.

Maybe the solution to the problem isn't to identify (and punish) trolls or problem users, but instead to identify problem content (or better yet, promote quality content). The following things, I think, ought to be discouraged in that spirit (feel free to agree or disagree, obviously).

* Personal attacks of any sort. Complaints about a user's behavior should perhaps go in an admin channel, not in the main discussion forums. Telling a person he's an idiot should go to /dev/null. In particular, accusations of trolling and such are fodder for the admin channel.

* Posts that just say "I agree", "I disagree", etc. If we go to a rating system, these become redundant anyway.

* Drive-by technology flames--including ones which are likely to be agreed upon by the participants. No matter how strongly you may think C++ is crap (or Microsoft or Java or whatever), simple declarations of such contain little useful content, and can easily venture into unprofessionalism. Reasoned critiques of legitimate LtU subjects should remain welcome--I'm only referring to the one-liners.

* Material which is scientifically discredited or just plain wrong. If this is the result of obvious error or newbie mistake, just correct it. Intentional and repeated posting of such may be considered abuse. This category should be limited to things which are demonstrated fact or dogma, not opinion--posts like "Java is the best language out there!" do not fall under this category (regardless of how wrong-headed you may think the claim is).

In addition, the following behaviors/posts always raise red flags. I wouldn't go so far as to ban these, but users should keep in mind that these make for weak arguments:

* Observations which are little more than personal opinion or experience--many comments on "intuitivity", "expressiveness" and such are of this sort. Likewise, proclaimations of superiority of some technology or technique, especially outside of a specific context.

* Anti-scientific and anti-academic remarks; disrespect of experts on mere "that's his opinion" grounds. (Substantial disagreement with scientific authority is another matter). Haven't seen to many of these on LtU, but topmind is notorious for this behavior on c2, where he has consumed much space on Ward's hard drive advancing his own version of "type theory".

* A strong desire to "protect the newbies" from material you disagree with, by placing objections on the record. People who read this forum are generally knowledgeable on the topic, and more than capable of making up their own minds.

In addition, I'd propose the following guideline:

* If you want to express a something which is largely a matter of opinion--such posts should always be professional, and should contain a new (and useful) reason why. (In addition, the reason why shouldn't also be mere opinion).

Quite a bunch of the noise is created when one person posts an opinion which is reasonable to him but anathema to someone else; who then posts a contrary opinion. Even when the argument remains friendly, such debates often degenerate into "is not"/"is too" affairs, which quickly lose any focus and value they may have had.

Finally--the purpose of this entire forum should be re-emphasized. If I understand it correctly, the main purpose is the sharing of knowledge and ideas, not to be a debating society. Some debate is often pedagogically useful, and debates which result in the sharing of knowledge and ideas ought to be encouraged. But debates which don't result in the sharing of knowledge (and often which are not intended to) ought to be discourage on those grounds.

### I'm not so sure...

If you want to express a something which is largely a matter of opinion--such posts should always be professional, and should contain a new (and useful) reason why. (In addition, the reason why shouldn't also be mere opinion).

While I agree that all such posts should always be professional, and should ideally contain new information, I think it's overkill to rule out "mere opinion".

Programming languages are many things to many people. For a programming language theorist, a language might be a testbed for a new theory, or a platform for investigating new idioms.

For many (most?) programmers, however, they are the primary user interface between their minds and the machines that they must get work done on. (I see some knees starting to jerk a little. No, I am not starting a discussion comparing "real programmers" with "academics"! Bear with me...)

Any time you are talking about user interfaces, you're getting into the realm where "opinion" becomes important. When I see a new paper on a type extension for Haskell, for example, my first thought is always along the lines of: "What could I use this for?", or "Does it make any of my existing programs, where I had to use a clunky solution, more elegant?"

Discussions about "expressiveness" or "elegance" or suchlike are by nature opinion-laden and subjective, but they're real (if vacuous) things which programmers actually care about. Scheme programmers like blurring the distinction between code and data; it makes for a large range of "elegant" solutions to some problems. Haskell programmers like strong and expressive type systems; they cut your debugging time dramatically. C++ programmers like getting close to the machine when it matters, while still having some pretty powerful high-level abstraction mechanisms. These are concerns which are based almost entirely on opinion, but the concerns are real.

Of course, the last thing we need is old discussions rehashed with no additional material. But we also don't want to require newbies to read the entire archives before posting. (Reading current discussion before delurking is more usual and more reasonable.)

So I'm torn.

### Signatures

Upon arriving, I discovered the signature capability. I also noticed that nobody else uses them. Figuring well, if it's available then it's allowed, I created one. I hope I am not ill regarded for doing so, as my actions were innocent.

My 1st one consumed 4 lines of text - fine by Usenet standards, but it looked awkward here. So I shortened it to 2 lines. I would have preferred 1 line, but I wanted people to be able to see my name at a glance, and I had too much text for 1 line. I suppose the underlying problem is the posting attribution identifies me as bvanevery rather than Brandon J. Van Every. I'll see if I can fix that.

My point is, I tried to reduce the noise, while still providing the signal I wanted to send. Many forums think it's ok to provide short bits of personal info in .sigs. Apparently that's not the culture here. If you decide to formally enshrine this culture, and make it a rule, then it would be best to remove the signature capability.

I include a sample signature so that we see what we're discussing. Otherwise I'll refrain from using signatures until people decide whether they should be allowed or not.

Cheers, Brandon J. Van Every
(cruise (director (of SeaFunc) '(Seattle Functional Programmers)))

### Advertizements, Announcements, Job offers, Resumes

I would note that none of these are the same thing. For instance, announcing a FP conference, or another FP forum, is much more topical than advertizing a bootleg copy of Visual C++ on sale for \$99.95. Lists, as they grow, always have to make decisions on these things - if for no other reason than to stop people from screaming "you're evil! you're evil!" just because someone tried to hire someone or some such. Oftentimes in forums, people regard as evil, bad, nasty, unsufferable, horrendous, etc., anything that is not expressly condoned. For instance, once upon a time I had to take a firm hand about job posting policies, because Usenet folks tend to regard headhunters as Evil [TM] and usually have nasty, snippy words for them. I didn't want that kind of poisonous anti-job culture in my forum, so I put a stop to it.

I hope you create a clear list policy for what's accepted in the forum, rather than relying on "unwritten rules." Also I hope you temper the severity of your moderator responses in accord with the offender's relevance to FP. "Will not be tolerated, deleted without warning" sounds a bit scary, and may lead you to unnecessary animosity.

### Not Much True Trolling AFAICT

I don't see much true trolling here, but must admit I don't read all threads. But I've seen people called "troll" here.

For example tkatchev has been called a troll, but I don't think he is a troll at all. I think tkatchev has a different (but interesting and usually correct) way of thinking about things. [I also think he's brighter than I am.]

Tkatchev catches me off guard. He'll sometimes make a statement but not explain it in detail. I become puzzled and sometimes get it but usually don't. Several posts later he explains and I realize he was correct all along and I have learned something. It's a form of Socratic dialogue.

### Hmm...

I agree that tkatchev doesn't seem to troll for the sake of attention-grabbing. However, he would probably find a far more receptive audience if he would tame the rhetoric and try to be more diplomatic in his wording. Being too terse isn't always a sign of superior thinking. Sometimes hiding the details makes it harder for others to disagree.

### Wiki/Dictionary for newcomers.

I've greatly enjoyed LtU and would willingly do my part to help it continue to succeed. Something that I think COULD help, and has been discussed before is a good collection of introductory material in a dictionary or encyclopedia like form that newcomers are greatly encouraged to read.

Not that that would solve all these problems, but certainly, I think it could help with some. A wikipedia for programming languages would be a valuable resource in its own right. Links to the right intro papers and books would help(especially since older members could point the newbies at specific entries local to the site).

I'll help in any way that I can Ehud, just send me marching orders :).

### MHO

1. Closing commenting entirely. Agree. This is too harsh

2. Eliminating the option to create new threads in the forum I'm not a fan of this either

3. Establishing some sort of comment moderation. I'd prefer this, but don't know how much work it would be for the administrators. I like the idea of readers flagging messages 'unsuitable' and having them hidden when some threshold of readers have tagged them like Craigslist works (e.g. posts X percent of readers tag as unsuitable with at least 15 reads are not shown).

4. Banning users who use LtU to advertise, post trolls etc. I haven't seen much advertising on LtU, but the trolls have been much much worse in the last 2 months and the quality of good posts has declined. After a few warnings about trolling, I would not feel guilt for banning an abuser. Another problem is cutting down on tiny 3 sentence posts without content.

5. Eliminating user signatures (which are appended to all their posts), since they seem to mostly add noise.This seems like a good idea. I also don't like aliases, people seem to put more thought into posts when their name is associated with it.

I like Bryan Turners suggested limiting posting the number of posts new members can make per day. Its pretty hard to make more that 5 LtU quality posts a day.

### LtU: Style and Focus

I have contributed little to LtU, but I've read and learned much by lurking over the last few years. Thank you.

I'm sorry that the lack of respect and courtesy that characterises our modern age is causing problems here. My advice is this:

Those who make a real contribution - the editors and so on - should react to these problems as they see fit. They have earned the right to do so by their efforts. The rest of us should gratefully and quietly support them. I do.

A possible improvement would be the ability to move (sub)threads to a different topic where it would form it's own thread, leaving only a link with the new topic name in the originating thread. Would make sure discussion remained topical while allowing the discussion to continue, fighting the tendency for it to form new subthreads to replace deleted ones.

The "static vs dynamic typing" topic would probably grow endlessly, and so would the new "personal insults" topic. Getting your discussion moved to either could possibly function as a remainder that you've strayed off-topic.

### My Thoughts

Hello. I've been reading along with you for over 2 years now, and I feel almost like I know some of you, so when I saw this topic, I had to post.

1. Closing commenting entirely. I think we'd lose the main value of the site this way, because often a comment from another user will provide enough mind-jiggling context to allow someone to see something they didn't before. Agree.

2. Eliminating the option to create new threads in the forum, thus only allowing contributing editors to set the agenda, so to speak. Well, maybe have normal users submit it to contributing editors for it to be submitted? No, I think this would stagnate us somewhat, unless there were a lot of contributing editors. Plus, it increases the burden on them. I like the idea of giving the more trusted members (editors) the power to intervene if they see fit, but forcing them to be the only driving force would take some of the steam out of the discussions, I think.

3. Establishing some sort of comment moderation. I don't think that this is worthwhile if each user votes a comment up/down with the same weight; the more trusted members should have more weight. If we do moderate, I think that the person who started the thread should be trusted to moderate it. That would distribute workload (more submitters than editors), but also keep responsibility for the thread with whoever started it. A slightly higher burden, but if a troll abused this power, it would be aplified immediately for all to see. If we did something like that, there should be guidelines: "brain-damaged language" -- borderline; "you're a dummy" -- right out. Hopefully a gentle "maybe you could word that less abrasively" from one of the old timers will be enough to quell flames (assuming editors have meta-moderate).

4. Banning users who use LtU to advertise, post trolls etc. I think SPAM should automatically get you booted, but rmalafaia has a good point about the nature of difficult people. I like the idea of "Pay your dues", but balancing the amount of effort required, and supression caused is key. As a new member, I'd support many of the pay-your-dues techniques, with the right parameters (waiting n weeks before/between posts, etc). I still think you'll eventually need to give the editors some power over editing threads and/or chastising users; cutting down religious wars, (company-name) bashing,and trolling can only be done with human intervention. Somebody has to call them on it.

5. Eliminating user signatures While I don't think this is essential the the (more important) trolling/flaming issue, I agree that this is mostly lost space (perhaps limit to one line?)

This is a tough topic in general, but I'm really not worried that much, because:

1. The number of people interested in this area is not likely to be a significant fraction of the population anytime soon.

2. Your content quality is much better than other technical discussion sites. A few shakey threads are okay, as long as they aren't growing in proportion to the site's volume.

Sorry that this is kinda long and incoherent, I just had a sudden urge to write, because I hope LtU keeps it's high quality.

P.S.
As the site is growing, perhaps it would be helpful to add a simple way to slap category/department tags from a pre-defined group (e.g. in the form of a bunch of checkboxes in the New Topic page). That way people can more easily find/avoid topics. That way you can avoid multiple inheritance :-) (I know the site volume isn't quite this big yet, though)

### Split up the site in scientific and public parts.

Let the scientific part work on invitation only, so scientists can present their papers there and discuss programming languages on a higher level. Newcomers in this area should have to present a paper or provide some piece of work that will enable them entrance. Participation to the scientific part will be limited either by time period of by topic.

Let the rest of us who are interested in programming languages but are not 'scientists' or academics battle it out in another part of the site.

It would be too bad for social reasons to ban people who say "Haskell sucks" or "C++ sucks". People say these things because they had bad experiences with language X or Y, or they had very good experiences with those languages and they blindly follow them. From a sociological point of view, these flamewars say more about programming languages than any calculus paper. Let the community expose them. If I say that Haskell is the best there is, then I would have to back up my claims. If I don't, then other posters could come up and say to me that I suck. It's the right thing to do, even if it is quite a darwinian process, because the good will be sorted out from the bad quite easily.

Please keep in mind that programming languages are for everybody, and not for the academic elite only. Many people will not understand many concepts, and good ideas can not be easily spread to the programming public if those that hold the knowledge have an elitist attitude.

Split up the site in scientific and public parts.

I would hate to see this happen. I may not be capable of understanding some of the deeply theoretical discussions here, but seeing tham helps me to understand (a) what the theoreticians thinks is important (and why), and (b) gives me some guideposts as to what areas I should be investigating myself. IMHO part of what makes LtU great is the free interchange between academics and "practical programmers" (not that it's really all that easy to draw a precise line there - many LtUers seem to blend both categories). How else will academic ideas make it into the mainstream, if not through the sharing of knowledge with non-academics? And how are academics going to get any understanding of how their theories impact the "real world" without discussing things with non-academics? Let's not create yet another ivory tower within LtU.

It would be too bad for social reasons to ban people who say "Haskell sucks" or "C++ sucks". People say these things because they had bad experiences with language X or Y, or they had very good experiences with those languages and they blindly follow them.
It would be preferable if we could just avoid the language pissing matches altogether. There are plenty of other sites that provide a forum for that kind of thing. The reality is that no one language is good for everything, and blanket assertions about the goodness or badness of a given language are completely useless unless there is also a discussion of the context in which language is used, and identification of the language features that may or may not be useful in such a context (i.e. how they aid abstraction, or reasoning, or reuse, or debugging). At that point, we might (might) be able to have a rational discussion. Which is, I thought, the point of LtU.
Please keep in mind that programming languages are for everybody, and not for the academic elite only. Many people will not understand many concepts, and good ideas can not be easily spread to the programming public if those that hold the knowledge have an elitist attitude.
Sure, many people won't understand some of the concepts presented here (I certainly didn't understand much to do with the recent discussion of *F* operators). But at least they will be aware that such concepts exist. That, to me, is the value of LtU. From my perspective, the other key part of LtU is that it focuses on trying to build understanding of these difficult concepts. I have in general found the LtU community to be supremely helpful when it comes to providing layman's explanations of difficult concepts, or pointers to more introductory material elsewhere on the web.

I've learned a lot here. I hope that I will be able to continue doing so.

### Me 2

I agree with most of the sentiments you expressed.

### I would hate to see this happ

I would hate to see this happen. I may not be capable of understanding some of the deeply theoretical discussions here, but seeing tham helps me to understand (a) what the theoreticians thinks is important (and why), and (b) gives me some guideposts as to what areas I should be investigating myself. IMHO part of what makes LtU great is the free interchange between academics and "practical programmers" (not that it's really all that easy to draw a precise line there - many LtUers seem to blend both categories). How else will academic ideas make it into the mainstream, if not through the sharing of knowledge with non-academics? And how are academics going to get any understanding of how their theories impact the "real world" without discussing things with non-academics? Let's not create yet another ivory tower within LtU.

But that exactly is what I am saying too: the academics must blend in with the croud. But that can't happen during a flamewar. The education can always take place in the public forum, whereas the academic discussions in the science forum.

It would be preferable if we could just avoid the language pissing matches altogether.

I want that too, but it ain't gonna happen any time soon.

I have in general found the LtU community to be supremely helpful when it comes to providing layman's explanations of difficult concepts, or pointers to more introductory material elsewhere on the web.

Me too...that's why I want LtU improved. It will not be improved in the current form.

### There can be an overlap betwe

There can be an overlap between education and academic discussion - often a layperson can find themselves pointing at something interesting that's been buried somewhere in the paper even if they can't necessarily describe it formally, for example. Or there can be a link to something that's clear to a layperson or an experienced hacker with the right background but non-obvious to an academic without it. Running two mutually-referencing threads in parallel isn't going to help much.

### Let's be clear

It would be too bad for social reasons to ban people who say "Haskell sucks" or "C++ sucks".

Let's be clear: comments of the form "X sucks" are a good example of the sort of comment which is not very welcome on LtU. At the very least, such a comment creates a high bar for the poster of the comment to justify it in a way that goes beyond their own unstated subjective experience.

We will probably be publishing some more specific guidelines about comments like this. That doesn't mean we'll ban someone for saying something like that once, but we want to strongly discourage such discussion.

Newer readers may have become accustomed to seeing some discussions along these lines here, but that's precisely why we're considering ways to deal with the issue. The alleged troll-like behavior seen recently is just a symptom of what has been a bit of a slide towards more subjective and technically ungrounded discussions. As has been pointed out, if you want to have "your-tool-sucks-no-your-tool-sucks" discussions, there are plenty of other forums for them.

### comments of the form "X suck

comments of the form "X sucks" are a good example of the sort of comment which is not very welcome on LtU.

they are also a good opportunity to ask the person that said 'X sucks' why he said that.

Here is another proposal:

Delete all comments that say 'X sucks', and offer a button to post a specific 'X sucks' item that should contain the arguments of why 'X sucks'. If the arguments are not good enough, then the post should be deleted with a note posted in place that explains the reason why the post is inadequate.

### Why bother?

Delete all comments that say 'X sucks', and offer a button to post a specific 'X sucks' item that should contain the arguments of why 'X sucks'. If the arguments are not good enough, then the post should be deleted with a note posted in place that explains the reason why the post is inadequate.

First of all: as I said before, blanket 'X sucks' statements are meaningless without context. Perhaps if they were statements of the form 'X sucks for doing Y' then we might be getting somewhere. But even then, you end up with a statement that is (a) deliberately confrontational, (b) still missing context (e.g. sucks compared to what?), and (c) based on the 'X sucks' posts I've seen so far, most likely based on subjective, anecdotal evidence.

Second: Who decides whether or not the arguments are "good enough"? The editors? They've already said they don't wnat these kind of posts at all. I imagine that it is unlikely that they will find any argument "good enough".

Third: again, as I said before, there are plenty of other forums for language pissing matches and 'X sucks' arguments. Let's not turn LtU into c2.

### "Sucks" is not appropriate.

Axilmar writes: 'It would be too bad for social reasons to ban people who say "Haskell sucks" or "C++ sucks".'

I disagree. Pointing out specific deficiencies in a language, and showing why doing something a different way would be better, is one thing. But blanket "X sucks" contributions are not very useful.

### No change

Although I don't make comments here, I visit LtU everyday and enjoy reading the posts. Sometimes I read a post many times. Thanks to all of you! Qrczak's comments on another site is actually how I found LtU (some there probably considered his comments trollish but I found them rather informative). I would prefer no change and just ignore trolls.

### Lambda the Ultimate Killfile

I wrote a Greasemonkey script that can provide some relief for Mozilla/Firefox users if anyone's interested: Lambda the Ultimate Killfile.

### Excuse me, but I think that b

Excuse me, but I think that blocking out comments from users you'd rather not read is an 'orwellian' solution that does not fit the democratic nature of our societies (and the word 'relief' sounds rather heavy - it's just a comment on the web, nothing that our life depends on).

I think that it would be better if those 'X sucks' comments are marked by other posters as 'needing explanation why'...then the poster that made the comment should have to justify his/her opinion, otherwise he/she will be banned after a few 'X sucks' rounds.

Killfiling is age-old tradition. A right to free speech is not the same as a right to be listened to, and it's not as if there's a central authority generating the killfile.

### Lambda the Ultimate Killfile

You can take more ideas/code from a similar Greasemonkey script for another forum.

I don't think there's anything violent or closedminded about using such tools; you may have a perfectly intelligent person who says things which happen not to interest you. We filter info out all the time; we don't visit every website.

### a newbie's perspective

I've been lurking here for a little while and recently I've posted a little. I have to say, if you're not satisfied with the content or quality of posts, why not as a first step make explicit somewhere on the site what is expected? I was kind of baffled to read this post; everything seemed just fine to me. Now I'm not quite sure what to make of the site or what's expected here. Do user signatures really bother you that much?

### Good Essay on the Subject

I didn't see it mentioned anywhere above, but Clay Shirkey wrote a great essay on the subject awhile ago, "A Group Is Its Own Worst Enemy". His basic thesis is that this sort of thing has been happening for a long time over and over again -- because people are people. He admits it isn't a solved problem but does offer some tips. At the very least, its a fantastic read.

### Troll??

Troll? What is a â€œtrollâ€? The first time I heard this term was only recently and I have grown more and more curious about what it might mean. A troll is some sort of web log poster who for some reason may be metaphorically related to the dictionary definition of the word â€œtrollâ€. Now the dictionary definition has nothing to do with web logs. Basically there are two definitions. The one I am most familiar with is fishing. To troll is to slowly drag a fishing line behind a boat hopping for a bite. The other definition is a mythological creature, usually very short, that pops out of nowhere to pose some unexpected problem for the innocent person who is simply going about their business.

Both definitions seem to work. Trolling could mean â€œfishing for informationâ€. This usually means posing some problem or difficulty that the some other reader is willing to respond to. Simply saying something that everyone agrees to doesnâ€™t get you any new information. The internet is uniquely suited for this and really what it is all about. But I fear that one persons troll can be another persons pain in the neck.

### Flame Warriors

Best classification can be found at Flame Warriors. Personally, I've always thought of myself as a Diplomat.

[Edit Note: On second thought, on LtU I'm definitely the Evil Archivist, as I keep tabs on Everything you Say].

...funny stuff. :)

### A traditional forum troll is

A traditional forum troll is fishing for flames, not information.

### trolls

It depends: nowadays they are little uglish mischief creatures with fuzzy hairdos. In the scandinavian myths and J.R.R.Tolkien works, they are gigantic and stupid human-like beasts with a taste for human flesh...

i believe most trolls in web forums resemble neither of them. though some surely behave like the latter... :)

### Definition of "Troll"

The classical definition of a "troll" is one goes into a forum and makes statements which are controversial or offensive (to the members of that forum), with the intent of provoking a response. The troll's motive may be amusement, attention-getting, or a deliberate attempt to disrupt the normal workings of the forum.

Whether any trolls (under the classical definition) are to be found here, I dunno.

Sometimes, passionate advocates for an unpopular position can be labelled as trolls. Often times, differences of opinion as to whether an individual is a troll or not are derived from differences of opinion as to the purpose of the forum.

As an example, I will point to a fellow well-known to many of us, topmind. Top is a tireless advocate of relational database technology, and in particular a methodology he calls "table oriented programming". He is also a frequent critic of OO. In addition to maintaining a (website) on the topic, he is a frequent visitor to forums such as c2.com and comp.object (places largely dominated by OO proponents), where he is quite good at stirring things up.

Is topmind a troll (as is often alleged by others in these forums?) A lot depends on what you think the purpose of the forum is.

c2.com is, somewhat officially, neutral on the subject of what programming paradigm is the best. Unofficially, c2 was for a long time the home of the design patterns and XP communities, two topics intimately related to OO, and has (until recently) had a very strong Smalltalk bias. It's a safe observation that many c2ers consider the site a "pro-OO" site, and as such view topmind's contributions as an unwelcome intrusion. (There are also issues with the technical quality of top's stuff, but that's a different issue). Based on that (and based on the sheer volume of material he posts--top seems unwilling to let anyone else have the last word), top has been frequently labelled a troll by members of c2; a few edit wars have sprung up over his material. However, top is still quite welcome there (there is only one individual who Ward has banned at c2; that person was banned for far more serious misconduct than mere "trolling").

comp.object may be a different story. I don't read it, so I don't know if it's an officially neutral or pro-OO forum. Certainly, if a forum exists to promote OO (rather than merely discuss it, pro or con), the presence of a db zealot would be highly unwelcome.

Returning to LtU. Officially the site is language neutral, although with an emphasis on research. Unofficially, this site is dominated by advocates of functional programming; and there is a strong bias here against the dominant industrial OO languages. By "bias" I don't refer to any editorial policy, but the center of mass of the opinions of the participants here. Which is OK--there are lots of forums out there where one can discuss how to configure EJBs, and why virtual protected inheritance might somehow be useful. :) Interestingly enough, neither of the individuals which have been recently accused of trolling are anti-FP, though both have some controversial opinions on the topic.

### FP

Personally, I try to get more people interested in OOP to post, etc. I don't see myself as an FP guy.

The problems I am referring to are when people make statements that seem contentful, but on analysis don't say anthing, when they mention "facts" that are completely wrong, and when someone corrects them, use a bait and switch tactic etc.

Topmind, as far as I remember, was never banished here as a troll, even though some of the discussions were heated.

### My 2 cents...

The classical definition of a "troll" is one goes into a forum and makes statements which are controversial or offensive (to the members of that forum), with the intent of provoking a response. The troll's motive may be amusement, attention-getting, or a deliberate attempt to disrupt the normal workings of the forum.
One of the other classic "troll" behaviors is deliberately and consistently feigning ignorance or making intentionally obtuse comments. The obvious problem here is that there's no way to tell whether the individual is really trolling, or if one more explanation, elucidation, definition (or whatever) might do the trick... I think this kind of trolling is what people are most often accused of in ongoing debates, because people start to think, "How can this guy still not see what I'm saying?!" when in fact people are just talking past one another... Over the years, I know I've been guilty of this at times, and I'm certainly sorry for it.

As far as actual changes or remedies to LtU, I'm not sure if I'd really support any particular changes to the forum mechanics. As painful as it is, I think having this type of discussion every couple of years might be enough to keep problems in check.

I suppose one change that might be OK would be to allow Ehud to remove the ability to reply to selected comments. That would be a pretty conservative extension, and I think it would match pretty clearly existing practice. From time to time Ehud steps in and says, "Enough of this." It might be nice to be able to enforce that. But I'm not sure it's necessary.

I agree that the quality and thoughtfulness of some of the discussion has been a bit lower lately, and I don't like to see ad hominem attacks or extensive discussion of individual posters' behavior, or any of the other obviously anti-social stuff you guys have identified. That's the basics...

But frankly, I guess I'm less annoyed by some of the long-standing advocacy wars (static/dynamic typing, for example). I've learned a lot from those things over the years, and I feel like, as a community, we've made some progress. So, even if it's painful or boring at times, I don't really think "agreeing to disagree" is that fruitful. So I'm less interested in trying to contain or manage those discussions, and more interested in just trying to promote a high quality across the board.