Suggestion to prevent the spammers

Archived. Please post suggestions to our new community page!
Locked
tigerdude77
Posts: 13
Joined: Tue Nov 03, 2009 6:47 am

Suggestion to prevent the spammers

Post by tigerdude77 » Tue Nov 24, 2009 8:51 am

I love the way SearchTempest allows me to search the ENTIRE country, and the results are pretty darn good IMO. Using that same concept, isn't there some way for ST to run a daily utility to search for ads throughout the entire country, and if the same CAR ad appears dozens of times - REMOVE THAT AD everywhere. I flag these ads as fast as I can - even all over the country, but it's time consuming. I'm in IT and I know how easy this would be. I'm not sure what the magic number should be, but any identical ad that is listed more than 10 times should be deleted everywhere. Maybe block that sender from placing more ads until they can prove that they are legit. When searching for identical ads (CARS ONLY), simply sort the list of ads by title and date only (yes I know it's a lot), then proceed thru the entire list removing all the duplicate titles. That would also remove all those ads that are listed almost daily/weekly and the old ones are still there for weeks.

I still think ALL car ads should require a phone number. I tend to flag if I email them and there is no phone number and they choose not to respond with a phone number. If they want to sell their car, they should want to include their phone number. What idiot would not want to do that?

If you need help writing the program to remove the duplicates, let me know. I've been doing this stuff for ~ 35 years, and it's a piece of cake.

Thanks again for you product. Works great for me.

User avatar
Tempest Nathan
Site Admin
Posts: 1371
Joined: Fri Apr 18, 2008 10:06 pm
Contact:

Re: Suggestion to prevent the spammers

Post by Tempest Nathan » Thu Nov 26, 2009 10:08 pm

I can definitely see how something like this would be helpful. Unfortunately it's not as easy as it sounds. First of all, all the results from searchtempest actually go directly from craigslist to the searcher's computer. The searchtempest server never actually accesses craigslist, and so has no direct access to the results. Even if it were possible to access the listings though, many completely different vehicle ads will have the same title. (Think 2005 Ford Mustang GT say - across the US, how many of those do you suppose are listed on craigslist?) Even in the same day, there would be a lot of legitimate duplicates. Without actually opening every ad and parsing the text, which would put too much load on both my end and theirs, there's no way to tell the legit dups from the spammers. PLUS even if we wanted to parse the ads themselves, spammers will often make minor changes, often automatically, to the text of their ads to evade the auto-blocks craigslist itself already has in place.

As I said, I can absolutely see the value, and would be happy to hear more if you can see opportunities I've missed. For now though, I have found the ratio of user benefit to time invested to be greater elsewhere. ;)

Locked