I think this is a good idea and an interesting experiment. I also thought of a possible technical concern with rot13ing the names that I wanted to share. This may be obvious, but I felt it was probably safer to say it anyway.
Rationality Quote 1, posted on October 1st.
Ryvrmre Lhqxbjfxl
Rationality Quote 3, posted on October 1st.
yhxrcebt
Rationality Quote 6, posted on October 1st.
Nyvpbea
If you upvoted these three quotes, checked who the sources of your favorites are, and promised yourself that you would not change my opinion or upvote them after looking, and you then check back in the thread tomorrow, you might see:
Rationality Quote 10, posted on October 2nd.
Ryvrmre Lhqxbjfxl
Think that it is really a good quote, and then realize you remember who Ryvrmre Lhqxbjfxl is before you made your voting decision because you checked him yesterday.
I didn’t want to just find a problem and leave it there because it is an interesting experiment, so I thought of a number of ways to avoid this particular problem:
1: Tell people not to check in the middle in case they remember a particular encryption-decryption on a later check.
2: Disallow duplicate quotes.
3: Use a system that doesn’t have encryption-decryption pairs.
4: Have all names linked externally so there is no accidentally readable encryption decryption pair.
As far as I can tell, 4 is probably the best. It’s unlikely that you will accidentally click and read a link you didn’t mean to click, but you might happen to accidentally derot13 something mentally if you see it multiple times. As long as the link offers no hints as to the author, it should be a secure system.
3 is rather complex for this type of experiment, although it would also be secure.
2 is no fun, because sometimes there are multiple quotes from a single person that you want to share.
1 is also a possible idea, but it seems like it might be more frusturating than 4. However, it would be easier to implement.
This is admittedly a not a very substantial problem and will probably not have changed the results, and if it did would probably have not changed the results by very much, but since you were interested in an experimental idea I liked, I wanted to discuss possible ways to help the experimental protocol.
An improvement on 4 would be to link to the actual source of the quote, when possible—that is, to some webpage that lists the quote along with the author. When this is not possible, rot13.com can accept short phrases (such as an author name) in the URL itself: e.g. http://www.rot13.com/index.php?text=Ryvrmre%20Lhqxbjfxl
There is an ugly hack: people could post the name of the author as a reply to the quote, (along with a karma dynamo,) and people could downvote the name (and upvote the dynamo). Downsides:
1) The first few people to see the quote would still see the name.
2) People who have changed their options from the default to not hide sufficiently downvoted comments would still see the name.
3) wedrifid would be deprived of the opportunity to participate fully in the communal project along with the rest of us, as would anyone else awesome enough to live on the edge of using up all their downvote ammunition.
Bonus: voting on the value of the quote could be on a tab within the downvoted thing, so one would not even know the LW hive vote upon first reading.
This would cease to be an ugly hack if it were possible to add comments that would automatically be compressed without being downvoted.
I think this is a good idea and an interesting experiment. I also thought of a possible technical concern with rot13ing the names that I wanted to share. This may be obvious, but I felt it was probably safer to say it anyway.
Rationality Quote 1, posted on October 1st.
Ryvrmre Lhqxbjfxl
Rationality Quote 3, posted on October 1st.
yhxrcebt
Rationality Quote 6, posted on October 1st.
Nyvpbea
If you upvoted these three quotes, checked who the sources of your favorites are, and promised yourself that you would not change my opinion or upvote them after looking, and you then check back in the thread tomorrow, you might see:
Rationality Quote 10, posted on October 2nd.
Ryvrmre Lhqxbjfxl
Think that it is really a good quote, and then realize you remember who Ryvrmre Lhqxbjfxl is before you made your voting decision because you checked him yesterday.
I didn’t want to just find a problem and leave it there because it is an interesting experiment, so I thought of a number of ways to avoid this particular problem:
1: Tell people not to check in the middle in case they remember a particular encryption-decryption on a later check. 2: Disallow duplicate quotes. 3: Use a system that doesn’t have encryption-decryption pairs. 4: Have all names linked externally so there is no accidentally readable encryption decryption pair.
As far as I can tell, 4 is probably the best. It’s unlikely that you will accidentally click and read a link you didn’t mean to click, but you might happen to accidentally derot13 something mentally if you see it multiple times. As long as the link offers no hints as to the author, it should be a secure system.
3 is rather complex for this type of experiment, although it would also be secure. 2 is no fun, because sometimes there are multiple quotes from a single person that you want to share. 1 is also a possible idea, but it seems like it might be more frusturating than 4. However, it would be easier to implement.
This is admittedly a not a very substantial problem and will probably not have changed the results, and if it did would probably have not changed the results by very much, but since you were interested in an experimental idea I liked, I wanted to discuss possible ways to help the experimental protocol.
An improvement on 4 would be to link to the actual source of the quote, when possible—that is, to some webpage that lists the quote along with the author. When this is not possible, rot13.com can accept short phrases (such as an author name) in the URL itself: e.g. http://www.rot13.com/index.php?text=Ryvrmre%20Lhqxbjfxl
There is an ugly hack: people could post the name of the author as a reply to the quote, (along with a karma dynamo,) and people could downvote the name (and upvote the dynamo). Downsides:
1) The first few people to see the quote would still see the name.
2) People who have changed their options from the default to not hide sufficiently downvoted comments would still see the name.
3) wedrifid would be deprived of the opportunity to participate fully in the communal project along with the rest of us, as would anyone else awesome enough to live on the edge of using up all their downvote ammunition.
Bonus: voting on the value of the quote could be on a tab within the downvoted thing, so one would not even know the LW hive vote upon first reading.
This would cease to be an ugly hack if it were possible to add comments that would automatically be compressed without being downvoted.