repost checker?

[Maverick]

Seems like it would be pretty straight forward to implement some kind of repost checker for posts in Off-Topic. This would simply scan the post for a particular URL which has been posted into the system before and warn the user that this is a repost.

Only limitation would be if it slowed down the submitting of a post due to the repost check taking a while. I don't know if AT uses any custom fusetalk code but it may be possible to do this without requiring a new version of fusetalk.

Thoughts?

 

[x04d DaY]

That might help, but people don't always post the same link, even if it is the same video/picture etc.
But still, that would reduce atleast 50% of the reposts.

 

[So]

This is actually not a bad idea, but it's extremely CPU intensive, maybe if CPU and disk perfomance continues to outpace forum growth for the next ten or so years.

 

[HBalzer]

Originally posted by: Maverick
Seems like it would be pretty straight forward to implement some kind of repost checker for posts in Off-Topic. This would simply scan the post for a particular URL which has been posted into the system before and warn the user that this is a repost.

Only limitation would be if it slowed down the submitting of a post due to the repost check taking a while. I don't know if AT uses any custom fusetalk code but it may be possible to do this without requiring a new version of fusetalk.

Thoughts?

I think there are plenty of repost checkers already. They make themselves known with their gay
R
E
P
O
S
T
posts

 

[MrBond]

The SA forums use something like this - any time you post a new thread with a URL in it, it checks it based on past threads and lets you know if any other threads have been posted that contain that link.

It works fairly well, it does have some problems with certain characters in URLs and will say it's been posted again when it's clear it hasn't.

Of course, it also helps that people can close their own threads there before X number of posts flood them with useless replies. That's not really an option here though, because we'd get people closing threads in the useful forums and it'd make it hard to bump a thread with more info/etc.

 

[tami]

we already have users telling other users that they've made reposts. it usually forces the latter to start searching more carefully. sometimes, a few things can slip through, because the search function is not working 100%. maybe we can address that easier issue instead of this, which is definitely more processor intensive.

 

[olds]

Originally posted by: HBalzer

Originally posted by: Maverick
Seems like it would be pretty straight forward to implement some kind of repost checker for posts in Off-Topic. This would simply scan the post for a particular URL which has been posted into the system before and warn the user that this is a repost.

Only limitation would be if it slowed down the submitting of a post due to the repost check taking a while. I don't know if AT uses any custom fusetalk code but it may be possible to do this without requiring a new version of fusetalk.

Thoughts?

I think there are plenty of repost checkers already. They make themselves known with their gay
R
E
P
O
S
T
posts

Seriously, wwybywb?

 

[BillyBobJoel71]

REPOST

 

[gsellis]

Straightforward? I would not say that. You have to create a server side app that is an expert engine that constantly monitors all incoming posts for duplication. So, it needs to index all the text in previous posts and compare to posts being submitted in real time. Not a minor feat there either. No thanks. I like Anandtech running at the speed it does now. That engine would suck up CPU cycles and make disk activity go way up.

 

[KoolDrew]

Originally posted by: gsellis
Straightforward? I would not say that. You have to create a server side app that is an expert engine that constantly monitors all incoming posts for duplication. So, it needs to index all the text in previous posts and compare to posts being submitted in real time. Not a minor feat there either. No thanks. I like Anandtech running at the speed it does now. That engine would suck up CPU cycles and make disk activity go way up.

:thumbsup:

 

[HBalzer]

Originally posted by: oldsmoboat

Originally posted by: HBalzer

Originally posted by: Maverick
Seems like it would be pretty straight forward to implement some kind of repost checker for posts in Off-Topic. This would simply scan the post for a particular URL which has been posted into the system before and warn the user that this is a repost.

Only limitation would be if it slowed down the submitting of a post due to the repost check taking a while. I don't know if AT uses any custom fusetalk code but it may be possible to do this without requiring a new version of fusetalk.

Thoughts?

I think there are plenty of repost checkers already. They make themselves known with their gay
R
E
P
O
S
T
posts

Seriously, wwybywb?

Seriously why does everyone keep asking me that? Is someone missing that used to post ~40 post a day? I need to start a thread here in forum issues to answer everyone who ask me that. Like I tell everyone else ask HBalzerThe2nd that in a few weeks.

 

[tami]

Originally posted by: HBalzer

Originally posted by: oldsmoboat

Originally posted by: HBalzer

Originally posted by: Maverick
Seems like it would be pretty straight forward to implement some kind of repost checker for posts in Off-Topic. This would simply scan the post for a particular URL which has been posted into the system before and warn the user that this is a repost.

Only limitation would be if it slowed down the submitting of a post due to the repost check taking a while. I don't know if AT uses any custom fusetalk code but it may be possible to do this without requiring a new version of fusetalk.

Thoughts?

I think there are plenty of repost checkers already. They make themselves known with their gay
R
E
P
O
S
T
posts

Seriously, wwybywb?

Seriously why does everyone keep asking me that? Is someone missing that used to post ~40 post a day? I need to start a thread here in forum issues to answer everyone who ask me that. Like I tell everyone else ask HBalzerThe2nd that in a few weeks.

if you are banned and recreate a username HBalzerThe2nd to get around your HBalzer ban, you'll be permabanned for doing so.

 

[Maverick]

Originally posted by: gsellis
Straightforward? I would not say that. You have to create a server side app that is an expert engine that constantly monitors all incoming posts for duplication. So, it needs to index all the text in previous posts and compare to posts being submitted in real time. Not a minor feat there either. No thanks. I like Anandtech running at the speed it does now. That engine would suck up CPU cycles and make disk activity go way up.

Rather than monitoring all the text in every post, it would be easier to have a separate table for URLs used in new posts (within 30 days or so). Every time a new post is created, any [L] tags are checked against the URL table to see if they already exist. If so, a repost message is sent to the user. Otherwise the URL in their new post is added to the URL table.
Once an entry in the URL table is older than 30 days or so, it gets deleted.

But something like this would probably have to be implemented on the fusetalk side as part of the normal post submission process.
Or the repost check can be an optional step for the user if they want to avoid getting flamed. This would keep the volume of repost check requests much lower.

Looks like the interest is there but this may not be technical feasible just yet.

 

[So]

Originally posted by: Maverick

Originally posted by: gsellis
Straightforward? I would not say that. You have to create a server side app that is an expert engine that constantly monitors all incoming posts for duplication. So, it needs to index all the text in previous posts and compare to posts being submitted in real time. Not a minor feat there either. No thanks. I like Anandtech running at the speed it does now. That engine would suck up CPU cycles and make disk activity go way up.

Rather than monitoring all the text in every post, it would be easier to have a separate table for URLs used in new posts (within 30 days or so). Every time a new post is created, any [L] tags are checked against the URL table to see if they already exist. If so, a repost message is sent to the user. Otherwise the URL in their new post is added to the URL table.
Once an entry in the URL table is older than 30 days or so, it gets deleted.

But something like this would probably have to be implemented on the fusetalk side as part of the normal post submission process.
Or the repost check can be an optional step for the user if they want to avoid getting flamed. This would keep the volume of repost check requests much lower.

Looks like the interest is there but this may not be technical feasible just yet.

Hmm..this is not a bad idea...it could be further optimised by ONLY checking when a new thread is started, not every post, since posting an old link in an existing thread isn't such a big deal.

 

[Cheezeit]

It would be much easier and great if people actaully put some thought into their post and checked before posting. seriously, some reposts are on the same front page as the op.

 

[HBalzer]

Originally posted by: Cheezeit
It would be much easier and great if people actaully put some thought into their post and checked before posting. seriously, some reposts are on the same front page as the op.

I agree, It?s one thing to repost something that is say back beyond 6 or 7 pages but to repost a topic that is on the 1st or 2nd page is just sheer laziness, ignorance, and stupidity.

 

[gsellis]

Not quoting...

But using even a url monitor will not catch the reposts. Lots of the repost are rehosted elsewhere and "rediscovered". OK, they would catch most of the news threads and would have caught the phallys repost in OT (I think, maybe not).

 

[dighn]

Originally posted by: Cheezeit
It would be much easier and great if people actaully put some thought into their post and checked before posting. seriously, some reposts are on the same front page as the op.

It would be better if people used descriptive titles/posts instead of going

"LOL Look at this"
funny video

 

[Lonyo]

Why not just have the users search for the URL in previou sposts?
It's not hard, you do this:
Copy URL
Click "Search" button
Paste URL
Select correct forum(s)
Click "Search"
See if any hits come up.

 

[Maverick]

Originally posted by: Lonyo
Why not just have the users search for the URL in previou sposts?
It's not hard, you do this:
Copy URL
Click "Search" button
Paste URL
Select correct forum(s)
Click "Search"
See if any hits come up.

Something even this simple that was done automatically when URLs were posted would drastically reduce the amount of reposts

 

[HBalzer]

I don't see why repost bother you people so much. Yea I can understand the aggravation when they repost a topic on the next 2 or 3 pages. After that it is all the repost Nazis that are annoying. If it?s a repost and you have nothing new to add don?t click on it.

 

[Rock Hydra]

Actually, I don't think reposts are all that bad within reason. I don't know that many people who search through old threads to find funny movies, flash files, news stories, etc. These reposted threads would target a new audience, since some people obviously weren't around/have missed the thread. Now, as stated above, if it's been within the last 50 or so threads, then yeah, it's rediculous.