This isn’t to say that for-profit companies are bad, it’s just that if they can make money by tracking your behavior and serving you ads, then that will take precedence over giving your project a good user experience.
And it's not as if they exactly roll out the red carpet for all of their advertising users either. One runs across complaints all the time. So expect to be treated just a little worse than those other guys who make Google & Yahoo money.
They do... That's why you will get ads in the middle of the most popular shows. (advertiser's business over user experience again) Ratings are per show, not per channel really.
I said the primary user, not the ONLY user. TV, search engine results, magazines, news papers, radio, pretty much anything that's ads support has two (maybe more) users. Primary is the advertisers since they actually pay for things, and secondary is the viewers. The delicate balance is satisfying the viewers while also satisfying the advertisers.
Networks cancel shows not because nobody watches them, but because nobody will advertise on them, and they sure as hell don't care if the viewers who do watch it want to keep watching.
Yes that is true. But that isn't the point of the question.
The implication of the parent is that since Google (by analogy TV companies) make money from advertising, they have less incentive to take care of users.
Ratings determine how much a network can charge the advertisers for commercial time. Ratings are a measure of viewer satisfaction. This isn't coincidental.
Actually, ratings don't really measure viewer satisfaction. They measure the number of people watching. Those aren't really the same things.
The advertising based system on television is part of the reason why shows must have broad appeal, and often cater to a lowest common denominator. Shows which may require more effort from the viewer may get a smaller audience, but that audience may be much more satisfied. Unfortunately, since television's model is based only on total viewing audience, it doesn't matter how satisfied the viewer really is, just as long as they keep watching.
Gross viewer satisfaction by some definition of 'viewer' & 'satisfactions.' Very similar metric to box office sales where consumers pay directly. Better?
None of this is relevant to the point being discussed. The point is that just because company A gets paid by Advertiser B that wants to reach Customer C does not mean that company A does not have the incentive to serve customer C properly.
Gross viewer satisfaction by some definition of 'viewer' & satisfactions.
I don't think that's a nitpick. I'll sometimes watch a TV show, but still not be wholly satisfied with it. He's pointing out a problem with the metric. With the current economics, companies that produce TV would be most pleased if a whole ton of somewhat dissatisfied people watch a show, versus a few that love a show. The economics encourage mediocrity.
The analogy fails because networks have a finite amount of air time to sell every day. Shows with low ratings get dropped. Google and others can literally create advertising space. Applications with relatively low usage cost them little, but they are also unlikely to invest in improving it.
No they don't. They have limited attention, but their space is, for all intents and purposes, infinite. Networks have 24 hours of programming, Google has millions of page views, application users, etc.
But, even if something gets very small amounts of attention, it is still incredibly cheap to serve ads (because that's all automated) and, as a result, make money. The user experience only has to be good enough to keep enough people using it that you break even on the infrastructure.
The world of online advertising has not been rewarding to those that provide not very popular services cheaply. This has certainly never been the case for Google. The only services doing well from advertising on a large scaleare the very popular ones
This thread is awful. Every comment is completely on its own. No context. I say 'Google has the equivalent of online space and of course I have to prove that to be true in every possible context.
It is the earlier post by zedshaw that made a general claim that when the provider gets paid by an advertiser, a user cannot count on being looked after (more then in direct payments). I gave an example of industry that for decades has been proving this statement wrong. I get pounced, red herrings swinging.
If you are continuing the argument of the GGparent comment as you seem to be, you are arguing that TV is completely different from web because TV shows need to meet a certain threshold to stay on air. This is why TV people care about ratings/customer satisfaction.
It is the earlier post by zedshaw that made a general claim that when the provider gets paid by an advertiser, a user cannot count on being looked after (more then in direct payments). I gave an example of industry that for decades has been proving this statement wrong. I get pounced, red herrings swinging.
In the eyes of many, myself included, the example of TV proves your point wrong. TV content producers seem to be motivated and geared to produce mindless dreck for the LCD viewer. Hits happen occasionally and unpredictably, so that's not actually what they try to do. Excellent reviews from critics doesn't mean you can charge advertisers more. Mediocrity is the rule of the day.
Overall it looks nice. The most interesting/controversial bit to me is the following (taken from librelist.com):
Nobody Can Be Taken Off
The “libre” in the name means freedom for everyone to discuss their opinions, so nobody will have the right to boot you off or enforce arbitrary “list Nazi” rules simply because you disagree with them.
That of course doesn’t mean you should act like a jerk, but the librelist.com philosophy is that healthy communities can survive and need “trolls”.
Can anyone provide a defense (from Zed or anyone) for the idea that a healthy community needs trolls? I can understand the need for dissent and argument, but it seems to me that trolling is categorically different from those.
Trolls are the predators of internet communities. They prey on the sick, the weak, and the perennially outraged (mostly this last group) and, if they are successful, drive them away from the community. A community without perennially outraged people is better than a community with them.
The question is whether this benefit is worth the potential cost, which is probably something that almost entirely dependent on the nature of the community and the personalities of the trolls involved.
Edit: By "perennially outraged", I am referring to people who are constantly angry or miserable for no reason other than that such an emotional state is "what they do". They aren't trolls themselves, any more than a particularly unpleasant customer is a troll. They're just constantly unhappy and vocal about it.
Generally there's three classes of trolls (this is just my opinion):
Griefers are what you describe. People who are more interested in gaming the implied social norms of the community in order to destroy it. These people suck, but totally blocking them eliminates the second class:
Dissenters are people who just disagree with you, and considering nearly every open source project is started by opinionated dissenters it's just bound to happen. I believe that without these people though your project dies because your ideas stagnate and you never really deal with actual criticism until it's too late. Also, they may just be the vocal part of your user base that wants something you don't.
Finally, moderators who use their moderation powers to "indirectly troll". By this I mean people who unilaterally deny people's communications because it would interfere with their hidden agendas. If you go look at nearly every failed community, at the heart is someone like this.
What I think, and what we'll find out, is that what really drives people away is not the presence of all of these, but the dominance of one of these types of "trolls". My idea is to just balance them out and give the community a real vote on what is what.
I think there's one more, people are like Griefers except way worse.
I think usenet can speak to the problem of trolls ruinging the community. Which is why web forms came to dominate, despite usenet's superiority in technical features and access (most people had email clients before web browsers).
Usenet was ruined by a small number of people who were trying to ruin others reputations by posting all sorts of insane diatribes to the group. Plus the spammers. Uggg... the spamemrs.
We have different definitions of "troll." Your second category does not fit into mine; I do not consider someone who is genuinely trying to communicate to be a troll.
Your assumption is that it's impossible to differentiate category one from category two. I see no reason for that to be true.
Nope, I never said it was impossible to differentiate, I just said that the differentiation shouldn't be controlled by a few people who potentially won't understand the difference.
Here's a question for you: If you moderate a list, do you let subscribers see what you've moderated and why? If it's for the benefit of the mailing list community, then why not?
You should make this clear on librelist.com; right now, it appears that there will be no mechanism to deal with trolls.
If I did moderate a list (which I don't), and there was an ability for subscribers to see the raw feed, I'd allow that. I am curious if your approach will work. It's similar to how things operate here, but I wonder if it can work with no moderators. (HN does have editors who monitor flagged articles and comments.)
I'm curious if also that could bring up another model: lists can be moderated, but there's "Sousveillance" on the moderators with feedback rating from the community. That way, you get moderators, but people who are member get to see if you're just crushing dissenting opinion. People could then see that your moderation style is rated at "Nazi" and just not join.
The suckless mailinglist is a mailing list that seems to accept, if not encourage, the second type of trolling. It would be interesting to build some statistics from it's archives to see when and who abandoned the list. I think you are right, Zed, in that getting rid of the first type get's rid of the second type and empowers the third type. I also agree that the second type is very good for the community. Just look at the development that goes on in the suckless community, some really great stuff.
How about the ability for the list creator/moderator to enable/disable the 'troll' flagging functionality for their mailinglist.
This allows the creation of lists whose entire purpose is trolling each other (i.e. trolls trolling trolls trolling trolls...). Or in general if I think that the 'troll' flag is 'censorship,' then I have the option to disable people (on my list at least) from using it.
I think who they end up driving away is mostly determined by the kind of community and the kind of trolls.
Most of the places on the Internet where I hang out, the trolls like the community they are a part of and are rather uninterested in destroying it. They do their thing primarily to get an emotional reaction out of people, and the easiest people to get an emotional reaction out of are the people who are constantly emotional. Because the overlap between permanently unhappy people and the "good" people (the latter group consisting of, e.g., core devs, or mods, or various other "central" members of the community) is (for most types of communities that I'm interested in) pretty small, that's not really a problem.
On the other hand, turning 4chan loose on a Neopets mailing list would probably end badly for the community, yes.
4chan users are not trolls. They're griefers and sociopaths. Like you said, trolls are generally part of a community, or make an attempt to get a rise out of people through détournement, not through outright antagonism.
Over at kuro5hin.org we experienced a 'trollocaust' as it were, wherein the trolls successfully drove all the emotional beautiful souls off to another site (hulver.com). If the critical mass of trolls grows too large, community can't really be sustained.
If we define the emotional state of a person relative to the sort of messages they send to the list, rather than their "true" emotional state, which is more relevant to the list, the the trolls create "perennially outraged" people. I don't think that argument works.
I guess the premise is that they are (objectively speaking) a valid part of any community.
Perhaps in some way it is better to have "troll trolls" than people who appear to be normal but are actually very trollish. I remember a community where a bunch of people would just answer all questions with:
"OMG WTF? NOT THIS AGAIN. Use the search FFS."
Now this was more or less encouraged behaviour as Admins and Mods did this but these guys took it too far. So it eventually was trolling all but in name.
To be honest i'm not sure I believe all that but at least he has a "troll rating" system.
Look at my previous reply on the subject, but I'm hoping tons of really smart people can think of a solution that is not so binary as: moderation=="list nazi moderates all dissenting opinion away" OR troll=="full frontal goatse porn for all!".
In other words, with the librelist code, it IS possible to have a middle land where the community can voice their opinion of what's spam or troll without eliminating real dissenting opinions and debate.
Nothing to allow tyranny of the minority or majority. It’s all about free speech and open communication.
However, this is balanced with the above spam marking, and potentially a “troll rating” that’s similar to a spam rating.
Additionally, keep in mind that if done right this could make it easier to have a moderated list. Rather than put some small set of brains in charge of keeping it clean, we just get the users to help keep it clean. Then you have a list system that is designed for a community to debate and argue while having what they want for communication without needed too much blocking human intervention.
This worries me a little too. What about the guy who posts several messages an hour all consisting of only of swear words or who posts answers to newbie questions with links to Goatse or any of the infinite other ways I'm sure exist to totally disrupt a list if there's no way to control anything.
Of course presumably there will still be people with the authority to remove really bad examples like that. But will they have the time to look after the many many lists which will probably soon exist?
The plan is to do it craigslist style. Each email will have a "hate" link you can hit to go mark it as spam or troll. Too many of these and the sender is blocked (for spam) or throttled (for trolls).
On the inverse side, I think I'll also track your voting rates, and if you tend to rate too many people as trolls or spam then your voting will also be throttled.
But, it's also just a plan. The goal is to create something that lets communities moderate themselves without too much tyranny of the majority of minority. That's why the code is open, so that people with ideas about how to do that can help make it happen.
Ok, that makes a bit more sense. It's not that no one can be censored, it's just that the power to do so doesn't rest with a small group of admins. That sounds much better, hope it works out!
A big difference between craigslist flagging and inbox flagging is that by the time a user can flag a message in email, the message has been delivered to everybody. You can't have a few people remove crap messages from everybody's inbox (right?), so if I can post without reputation, and/or signup in some automated way, the whole thing melts down.
Maybe deliver to people in a reputation order and slight delay to allow the damage to be minimized?
"Damage"? Yeah, it's impossible to stop all spam or stuff you don't want to see. It's only possible to reduce it over a longer period of time. In fact, any gear we cook up on the project using Lamson could probably be circumvented with Lamson itself if not thought out carefully.
So yeah, you'll get some spam. Install a spam filter and just accept that's the price you pay for being on the internet, and we'll do our best to reduce it.
Sounds just like "free software" to me--you're free from the constraints of commercialization, but not from the constraints of the programmers' social ideologies.
I'm not saying this to troll, I can just imagine lots of situations where I may want to set up a mailing list, but few where I want that list to become some sort of anarchic debating society where someone else's software handles the moderation for me.
Well, the code's right there, so if you don't trust anyone else's moderation software, then you are free to grab it and run your own show. In fact, I'll even put up docs on how to get your own going.
Maybe I'll call it the "tin foil hat" release of Librelist. :-)
Many systems allow a small minority to tyrannically control the group, applying censorship where nobody actually wants it.
Many systems may allow this behavior, but that doesn't mean it actually happens. At least I've never personally encountered it on a mailing list. Is it really that big a problem?
On the other hand, maybe the current state of comp.lang.lisp is a good example of what a mailing list with no moderation might look like. There are still good conversations, but there is so much spam and trolling that I don't bother reading it anymore -- it took too much work to sift through all the junk. Even worse, the trolls create an environment that is not exactly inviting to new users.
They make you log in just to see archives, which ruins any promotional and educational value your archives have.
That's not exactly true. At Yahoo, I believe this is up to the list owner. (I know my most active list there allows viewing of archives without logging in.)
At Google, it's weirder: if you're not logged into Google at all, you can often view the archives without being prompted for a login. But if you are logged into Google elsewhere, they demand you either reconfirm your login with Groups, or logout entirely, before allowing you to view to the archives.
(That's still a problem -- enough of an annoyance in fact that I wouldn't consider hosting any more groups there until it's fixed. But it seems mainly an eccentricity of their cross-site login policies, and does still offer a route to non-logged-in viewing.)
>You also subscribe to a list by simply sending your first message to list@librelist.com. It will then confirm you and send your original email on to the list. No special subscribe addresses, difficult workflows, or endless help references. Just send an email.
Totally unacceptable IMHO, and unlike how any other mailing list functionality works. If you send a message to the list, it should arrive there, without confirmation from you, and you shouldn't be subscribed unless you explicitly want that. Otherwise how can you participate in a conversation on the mailinglist without subscribing or allow cross-mailinglist posts?
Nothing to allow tyranny of the minority or majority. It’s all about free speech and open communication.
However, this is balanced with the above spam marking, and potentially a “troll rating” that’s similar to a spam rating.
How do you reconcile these two when an unpopular opinion can be marginalized as troll or spam?
This is why I’m proposing that Librelist be run similar to freenode.net in that it should be developed by the same community and a not-for-profit operation.
If it works out to be a friendly list interface with a group feeling, this would be an absolute winner project.
As a year of graduates we use a yahoogroups list since 2000, and we never changed it. Yahoogroups is currently running on 1995 technology, and google groups get spammed, let alone the web pages look like a pharmacy store. Also both lack the "we are a group" feeling.
So yes, since we're looking for a better interface and couldn't find one hitting 2010, if done properly this would be a definite winner.
Just curious, do most people here use the email client as the primary interface for the group sites?
I personally use the web interface for busy listings just to keep things separate from my actual email, and email client for only very low-volume listings. Am I missing out on something?
This is off topic, but are there any mailist that archive nothing, anonymize every poster's ID except spam or uninvited, delete unnecessary headers, and do not log anything?
Not that I know of, then again, if you want that, uh, why are you on a mailing list that's public in the first place. Sounds like you want a PGP key ring and a great Mutt alias setup.
Great idea. I've started to smell the "funk" recently around Google Groups; Zed pinned down exactly what it was. This will be a very interesting project moving forward.
This strikes me as hitting the nail on the head!