Googl, with the forum in my signature it's not impossible to stay a member, but it's impossible to stay a member unless you can make quality posts (and there's a very, very, very strict definition for what counts as quality: minimum 100 words per post, plus every post has got to meet three rules: it needs to be on topic, assist the discussion and add something new to the thread...or the account gets deleted.) Once they've cleared that hurdle of demonstrating they can talk sense members get access to a lot of premium content in the VIP Lounge and in Private Groups. It's a good circle - contribute something useful and you get free access to useful stuff others have contributed. Focus = quality. That wouldn't work in DP. It's a lot more casual here, there's a lot more chitter-chatter and it's a lot less serious. And that works in this environment, horses for courses. My money says that Shaun won't lose any members who matter, he'll lose only a handful of link tarts ...and he's better off without them. He won't lose money either as the benefit of the premium membership is not limited to the occasional algo-approved dofollow link.
I have a better analogy to this,... It's like going to a bar to have a drink and not being able to smoke a cigarettes, since beer and cigs go hand and hand. What I do now if I want to drink and smoke? I go to the Indian Reservation casinos where I can drink and puff all I want, I guess, if we disagree with this we can either go to another popular webmaster forums or create our own and make our own rules. While spammingy is annoying, couldn't DP maybe raised the post count to 1000 before one can have a signature and making it a no-follow? This takes a lot of things of being a member let alone being premium member, why even become a premium member at this point when your site could be considered a spammy.
And then you have the problem of retards posting even more useless posts because they are trying to get to 1,000 posts. And if the only value of premium membership to you is for links, I definitely would suggest not getting it.
Shawn, couldn't you have a filtering / trust system to weed out those claimed "retards" ...or mark their accounts to not acquire further privileges? That would solve a lot of problems. Trusted users who do make quality posts would then have dofollow signatures on reaching the set goal. I've hardly ever used a signature in my many years here so it wouldn't affect me, but it would serve as an incentive for some of your members like dcristo and wwws. I'm reluctant to teach a grandmother to suck eggs, but the level of trust on our forum goes a bit like this: Level 1 (new user): No trust. All posts pre-moderated, no signature, no access to editing profile, no access to VIP areas, PMs limited to a maximum of 5 (to prevent PM spamming). No blog posting. No Groups access. They are promoted to next level based on age of account, number of approved posts etc. Members have to prove themselves within 30 days. Accounts not making any posts that are deemed of acceptable quality (as per our anti-fluff rules) are ... deleted. OK, that's drastic but you could move them to a usergroup created for fluff posters Level 2: No pre-moderation, but all other blocks still in place. Slightly higher PM allowance Promoted to next level based on number of approved posts Level 3: Free access to premium content in VIP Lounge, ability to join and create groups, low key signature with only ONE link allowed etc. It works well for us; the Sitepoint admins examined the system to see which aspects they could adopt on their own forums and, I believe, they did implement some of our practices. Would they not work here?
Not manually, no, but with your programming skills I've no doubt you could drum up something to weed out 99% of new-user fluff auto-magically (much of it has clear markers). Or you could explore ways to crowdsource the task. For example, all new user posts could have a "Thumb this down if you think it's fluff" button. You could show that button to a certain category or categories of members. I'm sure a lot of other ideas could emerge from long time users here. Would you be open to the idea of castrating imbecile fluff posters and deleting their accounts before they become "regular members"? You could simply prune all zero post accounts periodically. Anyone incapable of making anything except fluff posts - which keep getting deleted - will never manage to hold an account down. I take great delight in my monthly cull of the idiots, spammers, fluffers, lurkers .... everyone with zero posts! BTW, I'm a big supporter of your marketplace as an alternative to Flippa's and have often recommended on the EP forums that people should use it more!
You ban a few key RIPE assigned netblocks and a couple key countries (not gonna name them) and about 99% of the crap would stop. But then you'd have the other 1% which is still just as crappy.
The nofollow change is only one (very) small thing compared to other stuff we do. Â Just because one thing we do isn't as effective as other things we do, it doesn't mean we should do away with it. Â That's the nice thing, we can do multiple things, not just one. Â Â
I just read all 14 pages of this thread and want the 30 minutes of my life back. I persoanlly think for a large forum like this its an obvious move, and as for the premium members complaining, unless it was outlined as a benefit of your membership when you purchased it (which I'm very sure it wasn't) I don't see the problem.
Far from it Shawn. Obviously there's a lot you do on a daily basis to combat spam on a forum this size and I won't ask you for details, but the question is do you wish to do anything about fluff and those posters who are capable of nothing but fluff? Those are the cases that are most likely to fill their sigs with links, most disruptive, least useful and the ones who give mods the most work. A large chunk of the moderation work disappears, as we've found, if you are more selective with who you take in. For a forum like ours the average daily spam attempts is probably 50-100. On our forum we don't get that many in the average month and most days the mods have no moderation work to do - the average rate at which each mod issues infractions is less than one for the year. Place a big enough hurdle for people to jump, and remove those who can't jump, and you eliminate 99% of the problem. The issue Sitepoint had with that approach when we last discussed it with them - and which would be a consideration here as well - is existing fluff posters who are grandfathered in. That can be dealt with as well if there's resolve to eradicate the worst levels of fluff. Dofollow for only those members who've proven themselves and there's incentives for members to try harder to make sensible posts to earn the privilege. At the moment the main tool you have for distinguishing members is post count and that, as you've pointed out already, ain't going to do anything except incentivise people to drive up post count. The only alternative is using a quality yardstick and that's what I found works a treat.
Yeah... I would suspect the bulk of the people banned are banned for posting pointless crap. Â It's not even about grandfathering anyone in though... if someone with 10,000 posts starts posting pointless crap, they will be banned just as easily as someone new. I think the main issue is people can only see what's NOT been done already... and they can't see what HAS been done. Â Over 1,000,000 posts have been deleted, but people only see the ones that were missed or not gotten to yet. I'm in no way implying the current system is as good as it can be. Â Anyone who makes that claim about *anything* is a fool. Â Â I'm constantly (as in daily) building things to make stuff better (people just don't notice most of the things). Â And I have a bunch of ideas that I haven't yet had the time to develop. Even if I don't acknowledge every little thing, I actually read people's suggestions and weigh various aspects of each... how difficult would it be to develop, what sort of resource impact would it have (both in hardware and man hours), what sort of other issues would whatever it is potentially cause, etc, etc... I'm fairly methodical when it comes to new things and never make changes for sake of making changes. Â If you were logged in as me to the forum, things look quite different because my account is more or less a testing ground before things go live. Â It doesn't all have to do with spam/useless posts either... there's a LOT of stuff I do and test. Perfect example... when I'm logged in some of the graphics are SVG files instead of PNG. Â The main logo is 7.7k as a PNG file, but as an SVG file it's 1.9k and more importantly it's vector based so it's infinitely scalable. Try zooming in on this image to see what I mean... if your browse supports viewing it's source, you can see it's just an XML file describing the image. http://i.dpstatic.com/misc/dps_logo.svgz Errrr... not sure how I got sidetracked there. Â lolÂ
Logo = very clever, I didn't know that stuff like that even existed! (I'm not technical, I don't understand HTML, CSS etc., much less XML) OK, I have one for you I've been watching you and DP with interest for many years ....and long before either of us became "millionaires". I even made a little donation to your server fund in the earlier days when DP was much smaller and you had a few less eBay-caused (or other) grey hair. I don't doubt either your technical skills or how much you do behind the scenes here - from the "don't reply to say you sent a PM" to the banning of selling Diggs to the current nofollow on sigs, there's constant change and improvement happening. Further, for every one thing we notice there may be a thousand we don't know about. But... if I may volunteer a suggestion.... Change the official DP definition of "pointless crap", move it up a few gears and ruthlessly delete members who can't cut the mustard. That'll trim the membership numbers down, raise the average IQ here by several thousand and make the entire moderation and admin jobs much, much easier.
I'd like to... but the problem is that it actually makes more work, not less.  If you make a rule that the quality of posts need to be at least of a certain level, it's ultimately subjective and there really is no way around needing to manually review the 10,000 or so posts that come through each day. The reality is that human moderation doesn't scale well.  You have a certain amount stuff to review, divided by people doing the reviewing.  For sake of argument, let's say we *could* manually review every post that people post in a reasonable amount of time...  What if there were 100,000 posts/day or what if there were 1,000,000 posts in a day?  It just doesn't scale... it's like asking Google to manually review every page on the Internet and apply a subjective "quality score" that the reviewer gives it.  Obviously we aren't at that scale, but I don't like to build mechanisms or processes that don't scale... it's pointless in my mind.  I'd rather work on ways to do it programmatically. And we *are* doing a lot of stuff programmatically around 500 new users per week are deleted before they even get a *chance* to make their first post... based on the probability they are going to be a spammer or a low quality poster.  It's based on a lot of different things, but just one example would be if historically we see a lot of post from that subnet that ultimately received infractions.  If the ration of infraction worthy to non-infraction worthy posts from that subnet gets over a certain level, we just stop allowing those users automatically.  It's a little better than just blocking entire countries (which I've thought of), but it's also far more complex. We spin a ton of posts through anti-spam services like http://akismet.com/ in order to programmatically pre-moderate posts (hundreds of posts are automatically blocked that way). As I've mentioned, I have some other things I need to find the time to develop... I just haven't had the time for it all yet. But spending a ton of time to build a framework for something that ultimately does not scale well isn't something I'm going to spend a huge amount of time on.  Developing systems that could make stuff become "self-moderating" is more the stuff that interests me.  Obviously you will always need some degree of human moderation... but they should be the ones just generally watching over the system, not the ones that are pre-moderating every piece of content that comes through manually. The "Report Post" function that any user can do is a very rudimentary type of crowd sourcing of existing users.  They see something against our rules, it alerts mods.  But what if users could somehow self-moderate without allowing cheaters to take advantage of the system... building something like that is a more worthwhile way to spend development time imo.
I can see your concern. The beauty of what I'm suggesting is actually that it's highly scalable. Subjectivity can, to a large extent, be stripped out by the wording of your fluff definition. Every post on our forum (by a newbie) has to meet three criteria: It has to be on topic, add something new and be a certain length. You could add "no repetition of anything already said". There is very little subjectivity there and, further, it allows shifting a lot more of the moderation to something like the "Report Post" function: In DP you could even allow posts by such newbies to be viewed by only trusted members who have the option of marking it as fluff so it gets automatically deleted and never reaches a full public audience. On various sites I too use IPs to filter, and akismet etc; it's not a case of replacing them but adding to what those tools do. I have a smart guy, a very smart guy, as an admin of one of my larger forums (details available by PM if you wish). It was phpBB and he was an expert in it. He had developed various analytical tools, had huge IP block lists etc. He was passionate about killing spam and spammers and did a brilliant job. His tools could even tell spammers by the occupation they gave when registering and how many millisecond/seconds/minutes it took them to perform certain tasks etc. When I suggested a couple of years ago that we move the forum over to vB and I gave him my master plan to dump a lot of what he had already built ... he freaked. I eventually persuaded him and today he's a very happy camper. Sometimes less is more.
Anyway I don't see what the fuss is about, nofollow has been introduced by google, then they said it's not so useful anymore. All the while I don't think it matters much. But I guess you've done studies about that, me I prefer to continue thinking it doesn't matter if it has the nofollow thing or not. I like that the signature links for premium members are still visible by guest though, because this can bring traffic sometimes.
Spammers (and telemarketers) are like deseases, they are apart of the internet (and offline), if you let it be, it will consume you. To lower the spam rate, be on top of it and that's all one can do, spammers will do the same, it's a vicious circle until one gives up and it's never the spammers that does the giving up.
Hmm finally had some time to run over some of Shawn's posts . Scalability is indeed the name of the game in the industry once you've passed several barriers . By the looks of it you need a self learning AI that can nuke fluff as fast as it is posted . Unfortunately this type of technology exists only in 1 MIT lab and it is sill experimental. However if that AI was able to better play Civ5 after reading tutorials created by humans for humans it will certainly be able to act as a addition to the post report system . That should increase the efficiency of the moderator team by a very large percent . Unfortunately until such technology will be available for us we're still stuck with code and people . My guess is that if Shawn somehow get 6 dedicated mods that spend 4 hours minimum doing moderation then 10.000 post won't be that hard to filter with the current system . 100.000 thou are out of the simple human league - you'd need a command and control system just to handle moderation and that is a burden a forum owner doesn't wish to have . I don't think Sitepoint is a comparable example , yes the quality there is a bit better but then again Sitepoint lacks the gargantuan market that DP has . IMHO yhat alone is a factor that can't be ignored . .