Rebroadcasting - anonymous or endorsed?

[Ravi Tejasvi]

Currently the boost (rebroadcast) is supposed to just forward the message to the other peers just by resetting the message priority (Option 1). Should we go ahead with that or follow what twitter does with retweet (Option 2)?
 As in, let's say the original message is "Hello" by user Ravi. And suppose user Sam boosts it, by option 1, just Hello by Ravi will be propagated via Sam,
 but in option 2, the message would become "@Ravi: Hello" by user Sam. 

(optional)

Going with option 2 would mean to add a container to the original message just having the recent booster. I.e  initially
<message author="Ravi"> Hello</message>

after boost by Sam

<message author="Sam"><boost><message author="Ravi"> Hello</message></boost></message>

and suppose this is boosted again by option 2 by Rita then,

<message author="Rita"><boost><message author="Ravi"> Hello</message></boost></message>

keeping track of only the original author. 

(/optional) 

So should we go with option 1 or option 2 or with both i.e one being anonymous boost and other endorsed boost (Two buttons). Please share your views on this.


Also if user Ravi was following user Sam then all the posts by Sam will be anonymously boosted by Ravi, to propagate the good content. Ravi can decide to not to follow Sam anytime he chooses. 

[Ian Clarke]

Just to add a little to this:

One of the goals of Tahrir is to ensure that "better" messages will get propagated more widely.  So, for example, if you are following another pseudonym, your node should automatically "boost" their messages (this would definitely be the "silent boost" - option #1 below).  I think we need this functionality no-matter what.

The question is whether a message that is explicitly boosted (by clicking the "boost" button on the message) should be silently boosted in this way, or whether it should be something closer to how retweets work, where you are retransmitting a message with your own signature, but quoting someone else's message.

Or perhaps we should provide both options (the main problem here being that simply explaining the difference to people will require several sentences - not ideal since a good UI doesn't require an instruction manual...).

Thoughts?

Ian.

--
You received this message because you are subscribed to the Google Groups "Tahrir Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tahrir-developm...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
Ian Clarke

Blog: http://blog.locut.us/

[Kevin Wang]

The option 1 is a must.  For option 2, the problem is, how do we decide the popularity of one message?  As the message (@user name : message content) is just a text, anyone can change it during the boosting.  It kind of like republishing the message rather than retweet it.

Actually, Twitter doesn't provide option 2 at the very beginning, if I recall it right.  The retweet part is done by the third party client, which just concat user's new comment with the original message.  And when the message is over 140 characters, the message will simply be truncated.

So I suggest we only do the option 1 at current stage, user can decide to "retweet" the message manually.

--

----------------------------------

Yours sincerely,

Kevin Wang

[Ian Clarke]

On Tue, Jun 18, 2013 at 5:25 PM, Kevin Wang <kev...@gmail.com> wrote:

The option 1 is a must.  For option 2, the problem is, how do we decide the popularity of one message?  As the message (@user name : message content) is just a text, anyone can change it during the boosting.

Actually not if the boosted message includes the signature of the original author, which would be embedded as an attribute of the XML tag that the message goes in - eg.

<message author="tom" signature="A5...">Totally agree with this <message author="fred" signature="r3W...">Joe Smith is great</message></message>

 

 It kind of like republishing the message rather than retweet it.

Isn't retweeting the same as republishing in the context of Twitter?

 

Actually, Twitter doesn't provide option 2 at the very beginning, if I recall it right.  The retweet part is done by the third party client, which just concat user's new comment with the original message.  And when the message is over 140 characters, the message will simply be truncated.

Yes, originally Twitter had no built-in support for retweets, which effectively meant that it only had option #2, but then they integrated the concept into their UI so now it looks a bit more like option #1.

 

So I suggest we only do the option 1 at current stage, user can decide to "retweet" the message manually.

Perhaps, except it might be difficult to retain the digital signature of the original message without explicit support in the UI.

 

Ian.

[Nidhi Rastogi]

I may be the first one to question, why we need a silent boost. From what I understood, silent boost means anonymously broadcasting a message to one's followers. Quora does something similar where a user can use ones credits to promote an answer. This appears in the list of questions of interest to other users. I don't' know how this algorithm works though.

Rebroadcasting will at least tell the followers who sent the message else if this turns into a spamming technique, they'll never learn who to un-follow.

So while rebroadcasting is necessary, anonymously doing that requires further conviction. And yes, how to explain the nuances to the users.

[Ian Clarke]

I think we definitely need an automatic silent boost - this is where Tahrir will boost messages automatically where there is some reason to believe that the message is of good quality, such as if the user is following the author.  The goal of this is to ensure that good quality messages get propagated further than poor quality messages.  We can control the "strength" of the boost by resetting the priority to different values (the priority of messages drops every time they are retransmitted, unless they are boosted).

The question is whether a manual silent boost should be possible, and my question is - why not?  If Tahrir can silently boost automatically on behalf of a user, then why shouldn't the user be able to silently boost manually?

Ian.

[Kevin Wang]

But if the boosted message includes the name of the original author and all users who retweeted it, won't it eventually reach the limit of 140 characters after being retweeted for several times (e.g.  @userA: // @userB: // @userC: blablabla)?  The message will be filled with user name instead of valued content, which is one problem of current Twitter.

Retweeting is somehow different from republishing, depends on who is the original author.  For example, if user B is going to retweet this message from user A, the original user will be A no matter how many times the message is retweeted, the users who retweet is just like forwarders.  The end receiver of this message doesn't care about who is in the path of retweeting, instead only need to know who is the original author, who is the one that broadcast this message to him.

e.g.

@userA: blablabla... (retweeted by @userB)

@userA: blablabla... (retweeted by @user

​C

)

​    userA will never be changed​

Republishing is quite like pasting the text of the message.  It is like quoting one's saying.  If user B republishes A's message, the original author will be B.  Then user C republishes it, the original author will change to C again.  Thus things will be different when you want to track who is the source of this message.

e.g.

@userA: // @userB: // @userC: blablabla

In the future, if you want to distinguish whether one user can be trust or not by algorithm, this can be trouble.  I don't know whether Tahrir will implement this in the future or have already implemented this, but in my mental design, the message from a user who have higher trust value, will be granted higher priority value when being published.  The "trust value" can be calculated by an average "quality" of messages coming for this user.  Thus spammer will have lower chance to win as they will have much low priority.

----------------------------------

Yours sincerely,

Kevin Wang

--

[Ian Clarke]

On Tue, Jun 18, 2013 at 7:18 PM, Kevin Wang <kev...@gmail.com> wrote:

But if the boosted message includes the name of the original author and all users who retweeted it, won't it eventually reach the limit of 140 characters after being retweeted for several times (e.g.  @userA: // @userB: // @userC: blablabla)?  The message will be filled with user name instead of valued content, which is one problem of current Twitter.

To solve this I think we would impose a limit of the message nesting depth, perhaps 2.  Clearly it would be ridiculous to have 7 or 8 embedded messages.  Tahrir doesn't have a 140 character limit, although we probably need to think about whether we should have any limit at all (we'll need some limit to prevent denial-of-service attacks).

 

Retweeting is somehow different from republishing, depends on who is the original author.  For example, if user B is going to retweet this message from user A, the original user will be A no matter how many times the message is retweeted, the users who retweet is just like forwarders.  The end receiver of this message doesn't care about who is in the path of retweeting, instead only need to know who is the original author, who is the one that broadcast this message to him.

e.g.

@userA: blablabla... (retweeted by @userB)

@userA: blablabla... (retweeted by @user

C

)

   userA will never be changed

Republishing is quite like pasting the text of the message.  It is like quoting one's saying.  If user B republishes A's message, the original author will be B.  Then user C republishes it, the original author will change to C again.  Thus things will be different when you want to track who is the source of this message.

Yes, I think we agree - we should only have the most recent booster and the original author.

 

In the future, if you want to distinguish whether one user can be trust or not by algorithm, this can be trouble.  I don't know whether Tahrir will implement this in the future or have already implemented this, but in my mental design, the message from a user who have higher trust value, will be granted higher priority value when being published.  The "trust value" can be calculated by an average "quality" of messages coming for this user.  Thus spammer will have lower chance to win as they will have much low priority.

Yes, there are lots of opportunity to use "web of trust"-type ideas here, but right now we need to focus on what is realistic functionality for the alpha (ie. the minimum that will be useful).

Ian. 

[Nidhi Rastogi]

I agree that re-broadcasting should not take away the credits from the original author of the message and should have his signature @author whoever re-broadcasts it.

I also think if a message is getting rebroadcast-ed (and not boosted) should increase and not decrease in priority since it's increasing in popularity with every re-broadcast.

--

[Ravi Tejasvi]

I also think if a message is getting rebroadcast-ed (and not boosted) should increase and not decrease in priority since it's increasing in popularity with every re-broadcast.

Err that's a naming error. Here if the priority is set to low means that it is more important.

 
And if User A follows User B then when silent boost is present, the posts of B are "promoted" via A. It will still have the original author's id. And these posts will come in "all posts", and the posts by whom A is following also appears in "following posts".

Let's say C follow B, B follows A. 
One thing we have to see is that whether we have to let C (who doesn't follow A) know that the posts appearing by A in "All posts" are because B is following A, or in other words B promoted it. This would be to stop spams as this can help the user to stop following any user.
But wouldn't anonymous boost lose it's meaning by doing so? 

--

Regards
Ravi Tejasvi

[Kevin Wang]

My understanding is that, the anonymous part is about low level things, such as you cannot trace back the IP or location of the source, but the whole Tahrir thing is a social product anyway.  If you do not include the original author (where you know who is saying that so that you may want to follow him directly) and last retweeter (for spam purpose and help user understand why he is receiving this message), how do user interact with strangers and expand their social circle?  

On Wednesday, June 19, 2013, Ravi Tejasvi wrote:

I also think if a message is getting rebroadcast-ed (and not boosted) should increase and not decrease in priority since it's increasing in popularity with every re-broadcast.

Err that's a naming error. Here if the priority is set to low means that it is more important.

 
And if User A follows User B then when silent boost is present, the posts of B are "promoted" via A. It will still have the original author's id. And these posts will come in "all posts", and the posts by whom A is following also appears in "following posts".

Let's say C follow B, B follows A. 
One thing we have to see is that whether we have to let C (who doesn't follow A) know that the posts appearing by A in "All posts" are because B is following A, or in other words B promoted it. This would be to stop spams as this can help the user to stop following any user.
But wouldn't anonymous boost lose it's meaning by doing so? 

On 19 June 2013 10:44, Nidhi Rastogi <nidhi...@gmail.com> wrote:

I agree that re-broadcasting should not take away the credits from the original author of the message and should have his signature @author whoever re-broadcasts it.

I also think if a message is getting rebroadcast-ed (and not boosted) should increase and not decrease in priority since it's increasing in popularity with every re-broadcast.

On Tue, Jun 18, 2013 at 8:26 PM, Ian Clarke <ian.c...@gmail.com> wrote:

On Tue, Jun 18, 2013 at 7:18 PM, Kevin Wang <kev...@gmail.com> wrote:

But if the boosted message includes the name of the original author and all users who retweeted it, won't it eventually reach the limit of 140 characters after being retweeted for several times (e.g.  @userA: // @userB: // @userC: blablabla)?  The message will be filled with user name instead of valued content, which is one problem of current Twitter.

To solve this I think we would impose a limit of the message nesting depth, perhaps 2.  Clearly it would be ridiculous to have 7 or 8 embedded messages.  Tahrir doesn't have a 140 character limit, although we probably need to think about whether we should have any limit at all (we'll need some limit to prevent denial-of-service attacks).

 

Retweeting is somehow different from republishing, depends on who is the original author.  For example, if user B is going to retweet this message from user A, the original user will be A no matter how many times the message is retweeted, the users who retweet is just like forwarders.  The end receiver of this message doesn't care about who is in the path of retweeting, instead only need to know who is the original author, who is the one that broadcast this message to him.

e.g.

@userA: blablabla... (retweeted by @userB)

@userA: blablabla... (retweeted by @user

C

)

   userA will never be changed

Republishing is quite like pasting the text of the message.  It is like quoting one's saying.  If user B republishes A's message, the original author will be B.  Then user C republishes it, the original author will change to C again.  Thus things will be different when you want to track who is the source of this message.

Yes, I think we agree - we should only have the most recent booster and the original author.

 

In the future, if you want to distinguish whether one user can be trust or not by algorithm, this can be trouble.  I don't know whether Tahrir will implement this in the future or have already implemented this, but in my mental design, the message from a user who have higher trust value, will be granted higher priority value when being published.  The "trust value" can be calculated by an average "quality" of messages coming for this user.  Thus spammer will have lower chance to win as they will have much low priority.

--

Regards
Ravi Tejasvi

--
You received this message because you are subscribed to the Google Groups "Tahrir Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tahrir-developm...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--

[Ravi Tejasvi]

Hmm that's correct. That was an error by my side. :) Original author must be included. 

Regards
Ravi Tejasvi