Love the site, really. But one small suggestion for the next version.
When updating the view count, would it be possible to disallow multiple clicks on a post that occur from a single IP address within a five minute period?
What kind of difficulties would that sort of filtering cause?
-
-
What if the requests come from different people who happen to share the same IP? (Yeah, it's not super common, not statistically important levels anyway, but it's a valid case to consider and it's hard tellin' not knowin' for sure, eh?).
Most people do not refresh the browser continuously in order to watch a view counter step. Why would one do this? Well, maybe interviewees would do it so that their interview looks more popular than it is really is, but I doubt very much that interviewees have time or inclination to do this. I mean, do you think this happens?
The rational/logical stuff out of the way, yeah, this could be done and probably is already implemented in the new code base. I'm not sure though.
KFA Team, thoughts?
C -
Charles wrote:What if the requests come from different people who happen to share the same IP? (Yeah, it's not super common, not statistically important levels anyway, but it's a valid case to consider and it's hard tellin' not knowin' for sure, eh?).
Yes, I'd probably assume that on a site aimed at techies, the vast majority of your readers, will be posting from their own machine, with a unique IP. This may not be the same for some consumer based sites, but even then, I would say that the number posting from shared ip addresses is pretty small.
Charles wrote:
Most people do not refresh the browser continuously in order to watch a view counter step. Why would one do this?
Well a friend is working on a profession-based community site, which offers incentives for popular stories. He was looking into ways of preventing folk from clicking loads of times to raise the ranking of the stories. He mentioned this over a beer or ten, and it got me thinking.
Does C9 rank by views or replies, or as I suspect, a combination of both? If it's number 3, then I thought that some kind of filtering system would improve the accuracy of the ranking.
-
Ray6 wrote:
Yes, I'd probably assume that on a site aimed at techies, the vast majority of your readers, will be posting from their own machine, with a unique IP. This may not be the same for some consumer based sites, but even then, I would say that the number posting from shared ip addresses is pretty small.
You sure? What about people that browser from work? Or indeed from within Microsoft (all their internet access is proxied)
I'd say the majority of readers who are not reading from home are masked behind a proxy or gateway IP
-
blowdart wrote:

Ray6 wrote:
Yes, I'd probably assume that on a site aimed at techies, the vast majority of your readers, will be posting from their own machine, with a unique IP. This may not be the same for some consumer based sites, but even then, I would say that the number posting from shared ip addresses is pretty small.
You sure? What about people that browser from work? Or indeed from within Microsoft (all their internet access is proxied)
I'd say the majority of readers who are not reading from home are masked behind a proxy or gateway IP
Good point.
Mmm ... and you can't do it from user sessions because folk might not be logged in.
How about looking for an expiring cookie on the browser?
-
Charles wrote:What if the requests come from different people who happen to share the same IP? (Yeah, it's not super common, not statistically important levels anyway, but it's a valid case to consider and it's hard tellin' not knowin' for sure, eh?).
Most people do not refresh the browser continuously in order to watch a view counter step. Why would one do this? Well, maybe interviewees would do it so that their interview looks more popular than it is really is, but I doubt very much that interviewees have time or inclination to do this. I mean, do you think this happens?
The rational/logical stuff out of the way, yeah, this could be done and probably is already implemented in the new code base. I'm not sure though.
KFA Team, thoughts?
C
What about a IP and a Cookie. If IP same different cookie new person within a reasonable time frame. Cookies disabled ? count away. -
Ray6 wrote:Love the site, really. But one small suggestion for the next version.
When updating the view count, would it be possible to disallow multiple clicks on a post that occur from a single IP address within a five minute period?
What kind of difficulties would that sort of filtering cause?
what problem do you think this might solve ??
I read some ideas about who it could be done but not really any info on *WHY* it should be done.
as was said an IP may be (often will be) shared.
and if I want to get nasty I can use multiple proxy bots coming from different class C subnets such that I could generate 10,000 different clicks that each may appear to be uniqe.
but that assumes that this is a case of somene flooding c9 with multiple hits.
so what are you looking to fix?? -
figuerres wrote:
and if I want to get nasty I can use multiple proxy bots coming from different class C subnets such that I could generate 10,000 different clicks that each may appear to be uniqe.
Yeah, you could. But most people won't.
-
Also, why can't we get a Linux forum? The Coffeehouse, Linux Forums. Would prevent so much spamming of Coffeehouse with pointless distro news.
-
Lloyd_Humph wrote:Also, why can't we get a Linux forum? The Coffeehouse, Linux Forums. Would prevent so much spamming of Coffeehouse with pointless distro news.
Because the sirhomer/emet/corona_coders don't care about Linux, they just want to stir the pot and piss people off. They're just craving attention. -
Lol, and since when was SirHomer a troll?
-
My IP is used by a few hundreds of people. There are also people that are required to use proxy servers.
-
I don't know the details of how views are recorded on the current Channel 9, but on the new one it's done by an image that is set to be cached for 12 hours. That means for every device/browser in a 12 hour period the view count will be increased by 1.
To get something truely accurate would require a re-architecting of the interwebs.
For the most part, our numbers should be pretty close unless someone abuses it.
IMHO, ranking should never be done by views alone.
Thread Closed
This thread is kinda stale and has been closed but if you'd like to continue the conversation, please create a new thread in our Forums,
or Contact Us and let us know.