4 ms·

Hi Patrick :-) So you want to tackle the spam issue. The problem with your system as far I see it is: 1) it can be easily gamed (as other people have pointed

by toumhi 15y ago

Hi Patrick :-) So you want to tackle the spam issue. The problem with your system as far I see it is: 1) it can be easily gamed (as other people have pointed out) 2) there's not enough incentive to use it So what? You have emails and a second email-like communication system? I think email is unlikely to be killed any time soon. Its versatility makes its force. Still, it's used for many use cases, and the way to use less email is to have better alternatives for specific use cases (for example, file attachments - todo items - meeting arrangements). Better solutions actually exist for each of the use cases but it's then a matter of getting enough velocity to have enough people using it (for e.g email attachments has probably decreased with dropbox and file sharing services).

padobson 15y ago
Part of my problem might be that I don't experience this particular problem either. However, it is something I see the target audience complaining about a lot. I agree that there might be some problems with gaming the system - any system can be gamed, we hackers know this better than anyone. So what are some ways the system could be made more secure? A LolCat page could get 250k likes on Facebook easily, but could it consistenly get 5,000 likes a week for a year? Maybe staying power should be more important than overall popularity. What are some better metric to measure importance?
- mds101 15y ago
  >A LolCat page could get 250k likes on Facebook easily, but could it consistenly get 5,000 likes a week for a year? Maybe staying power should be more important than overall popularity. Exhibit A: The Cheezburger Network. That said, if you change it to work such that it's personalized for every user, so that it only shows emails from people who are important to them, rather than emails from people the system considers to be important, then maybe people will appreciate seeing emails from the LolCat page of their choice. Also, a better way to measure importance would be to go through the user's sent items history and map out the people they email most frequently. I once wrote a classifier which classified email contacts as important or not important based on the sent items history of the user and it worked pretty well for me. You could also use Bayesian classification of the contents of the emails the user sends and the emails to which the user replies and use that data to determine if the user would like to see an email, based on it's contents (rather than the sender).

by toumhi 15y ago

padobson 15y ago

Part of my problem might be that I don't experience this particular problem either. However, it is something I see the target audience complaining about a lot. I agree that there might be some problems with gaming the system - any system can be gamed, we hackers know this better than anyone. So what are some ways the system could be made more secure? A LolCat page could get 250k likes on Facebook easily, but could it consistenly get 5,000 likes a week for a year? Maybe staying power should be more important than overall popularity. What are some better metric to measure importance?

mds101 15y ago
>A LolCat page could get 250k likes on Facebook easily, but could it consistenly get 5,000 likes a week for a year? Maybe staying power should be more important than overall popularity. Exhibit A: The Cheezburger Network. That said, if you change it to work such that it's personalized for every user, so that it only shows emails from people who are important to them, rather than emails from people the system considers to be important, then maybe people will appreciate seeing emails from the LolCat page of their choice. Also, a better way to measure importance would be to go through the user's sent items history and map out the people they email most frequently. I once wrote a classifier which classified email contacts as important or not important based on the sent items history of the user and it worked pretty well for me. You could also use Bayesian classification of the contents of the emails the user sends and the emails to which the user replies and use that data to determine if the user would like to see an email, based on it's contents (rather than the sender).