RE: [ydn-delicious] subwebs and better link recommendations

2006-10-09 Thread Larson, Timothy E.
Amir Michail wrote:
 * when entering a bookmark, you supply not only tags but also virtual
 inlinks and virtual outlinks; for example, when bookmarking
 TeXmacs, you might supply LyX as a virtual inlink and several TeXmacs
 resources pages as virtual outlinks.   
[snip]

If simply having pre- and post-link title expansion text is seen as too
complex for users, I _know_ this will be.

Tim
-- 
Tim Larson
West Corporation, Interactive TeleServices
Eschew obfuscation!


 
Yahoo! Groups Links

* To visit your group on the web, go to:
http://groups.yahoo.com/group/ydn-delicious/

* Your email settings:
Individual Email | Traditional

* To change settings online go to:
http://groups.yahoo.com/group/ydn-delicious/join
(Yahoo! ID required)

* To change settings via email:
mailto:[EMAIL PROTECTED] 
mailto:[EMAIL PROTECTED]

* To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]

* Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
 




Re: [ydn-delicious] subwebs and better link recommendations

2006-10-08 Thread Amir Michail
On 10/8/06, Joshua Schachter [EMAIL PROTECTED] wrote:
 I've frequently thought that there should be a way to tightly bind a
 bunch of items together (i think of them as a project, all researched
 in one swoop).

 I'm not sure how this would be different from virtual inlinks/outlinks,
 or even what the distinction might be...


I suppose you mean bind a bunch of items more tightly than through
commonly used tags.  Virtual inlinks/outlinks are more fine-grained
than commonly used tags for grouping and there is no need to invent
new tags (and remember what they mean).

Incidentally, a potential problem with virtual inlinks/outlinks is
that many users may not know when to use an inlink and when to use an
outlink. So maybe supplying related links might be good enough.

But I think the potential of virtual links to improve personalized
recommendations in a way that leverages the collective intelligence of
users (rather than by using more sophisticated collaborative filtering
algorithms) is interesting.

Also, I have doubts about personalization methods that depend on
similarities among users (e.g., collaborative filtering, social
networks, etc.). The method I proposed does not assume that other
people share your interests along multiple dimensions. It's more
fine-grained.

Amir


   _

   From: ydn-delicious@yahoogroups.com
 [mailto:[EMAIL PROTECTED] On Behalf Of Amir Michail
   Sent: Saturday, October 07, 2006 9:02 PM
   To: ydn-delicious@yahoogroups.com
   Subject: [ydn-delicious] subwebs and better link recommendations
   
   

   Hi,
   
   What do you think of the following?
   
   * when entering a bookmark, you supply not only tags but also
 virtual
   inlinks and virtual outlinks; for example, when bookmarking
   TeXmacs, you might supply LyX as a virtual inlink and several
 TeXmacs
   resources pages as virtual outlinks.
   
   * these virtual inlinks and outlinks allow you to create your
 own
   subweb that you can use to browse your bookmarks; for example,
 when
   browsing the LyX bookmark in your subweb, you will see TeXmacs
 as a
   virtual outlink
   
   * moreover, the combined virtual inlinks and outlinks of all
 users
   provide an alternative view of the web, perhaps with more
 interesting
   linking
   
   * when browsing links, you can click more like this or fewer
 like
   this; the system keeps track of both of your liked and disliked
 links
   
   * virtual inlinks and outlinks along with liked/disliked links
 can be
   used to enhance personalized recommendations
   
   * for example, to compute the personalization score for some
 link Y
   with respect to some user P's liked links U_1, ..., U_m and
 disliked
   links V_1, ..., V_n, we could do the following:
   
   compute scores for Y's virtual inlinks from all users: say S_i
 for
   X_i = Y (the stronger the implication, the higher the score)
   
   compute scores for Y's virtual outlinks from all users: say S'_i
 for
   Y=X_i (the stronger the implication, the higher the score)
   
   compute the personalization score for Y with respect to user P
 as follows:
   
   sum S_i over U_i - sum S'_j over V_j
   
   Amir
   

   



 [Non-text portions of this message have been removed]





 
Yahoo! Groups Links

* To visit your group on the web, go to:
http://groups.yahoo.com/group/ydn-delicious/

* Your email settings:
Individual Email | Traditional

* To change settings online go to:
http://groups.yahoo.com/group/ydn-delicious/join
(Yahoo! ID required)

* To change settings via email:
mailto:[EMAIL PROTECTED] 
mailto:[EMAIL PROTECTED]

* To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]

* Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
 




Re: [ydn-delicious] subwebs and better link recommendations

2006-10-08 Thread Mislav Marohnić
On 10/8/06, Amir Michail [EMAIL PROTECTED] wrote:

 * when entering a bookmark, you supply not only tags but also virtual
 inlinks and virtual outlinks; for example, when bookmarking
 TeXmacs, you might supply LyX as a virtual inlink and several TeXmacs
 resources pages as virtual outlinks.


certainly interesting, but few would use it because of the complexity :-/

I'm pretty much tech-savvy and a del.icio.us power user, but what you
described would take too much time even for me... posting should never be
more than it is now: title, summary, tags. I left Ma.gnolia just because of
the extra 1 click everywhere


[Non-text portions of this message have been removed]



 
Yahoo! Groups Links

* To visit your group on the web, go to:
http://groups.yahoo.com/group/ydn-delicious/

* Your email settings:
Individual Email | Traditional

* To change settings online go to:
http://groups.yahoo.com/group/ydn-delicious/join
(Yahoo! ID required)

* To change settings via email:
mailto:[EMAIL PROTECTED] 
mailto:[EMAIL PROTECTED]

* To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]

* Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
 





Re: [ydn-delicious] subwebs and better link recommendations

2006-10-08 Thread Amir Michail
On 10/8/06, Mislav Marohnić [EMAIL PROTECTED] wrote:

 On 10/8/06, Amir Michail [EMAIL PROTECTED] wrote:
  
   * when entering a bookmark, you supply not only tags but also virtual
   inlinks and virtual outlinks; for example, when bookmarking
   TeXmacs, you might supply LyX as a virtual inlink and several TeXmacs
   resources pages as virtual outlinks.

  certainly interesting, but few would use it because of the complexity :-/


It's different, but I don't think it's complex. As with tags, you
invest some effort at bookmarking time to help you find stuff later.

  I'm pretty much tech-savvy and a del.icio.us power user, but what you
  described would take too much time even for me... posting should never be
  more than it is now: title, summary, tags. I left Ma.gnolia just because of
  the extra 1 click everywhere


This is optional anyway and the system could suggest some virtual
links for you to use (much as it now suggests tags for you to use).

Amir

  [Non-text portions of this message have been removed]




 
Yahoo! Groups Links

* To visit your group on the web, go to:
http://groups.yahoo.com/group/ydn-delicious/

* Your email settings:
Individual Email | Traditional

* To change settings online go to:
http://groups.yahoo.com/group/ydn-delicious/join
(Yahoo! ID required)

* To change settings via email:
mailto:[EMAIL PROTECTED] 
mailto:[EMAIL PROTECTED]

* To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]

* Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
 


Re: [ydn-delicious] subwebs and better link recommendations

2006-10-08 Thread Nitin Borwankar

On Sun, 8 Oct 2006 14:01:54 +1000, Amir Michail [EMAIL PROTECTED]
said:
 Hi,
 
 What do you think of the following?



Aside from tag-related links and user-related links, this would be a way
for a user to provide *explicit* user-related links, things that a user
thinks *should* be related to this link.

This is conceptually very close or identical to the kinds of collections
enabled by Rollyo, Wink, Squidoo, Kaboodle and others (Yahoo MyWeb ? ).
So, adding a lightweight, easy-to-add explicit user-defined clustering
on a bookmark might be a good idea.  If the system showed clusters that
other users have created around this bookmark, that might help as well. 
It appears different enough from other notions of related (
http://tinyurl.com/e854x ) to be interesting if added in a lightweight
way to del.icio.us.

You could possibly jigger the existing delicious tag system and fake
this by creating large sets of special case tags to keep all this glued
together with private semantic tags, but that seems to suggest this
should really have intrinsic system support.

Nitin Borwankar, Tag Schemer
http://tagschema.com



 
 * when entering a bookmark, you supply not only tags but also virtual
 inlinks and virtual outlinks; for example, when bookmarking
 TeXmacs, you might supply LyX as a virtual inlink and several TeXmacs
 resources pages as virtual outlinks.
 
 * these virtual inlinks and outlinks allow you to create your own
 subweb that you can use to browse your bookmarks; for example, when
 browsing the LyX bookmark in your subweb, you will see TeXmacs as a
 virtual outlink
 
 * moreover, the combined virtual inlinks and outlinks of all users
 provide an alternative view of the web, perhaps with more interesting
 linking
 
 * when browsing links, you can click more like this or fewer like
 this; the system keeps track of both of your liked and disliked links
 
 * virtual inlinks and outlinks along with liked/disliked links can be
 used to enhance personalized recommendations
 
 * for example, to compute the personalization score for some link Y
 with respect to some user P's liked links U_1, ..., U_m and disliked
 links V_1, ..., V_n, we could do the following:
 
   compute scores for Y's virtual inlinks from all users: say S_i for
 X_i = Y (the stronger the implication, the higher the score)
 
   compute scores for Y's virtual outlinks from all users: say S'_i for
 Y=X_i (the stronger the implication, the higher the score)
 
   compute the personalization score for Y with respect to user P as
   follows:
 
   sum S_i over U_i - sum S'_j over V_j
 
 Amir
 
 
  
 Yahoo! Groups Links
 
 
 
 
 
 
 
 
-- 
  Nitin Borwankar
  [EMAIL PROTECTED]



 
Yahoo! Groups Links

* To visit your group on the web, go to:
http://groups.yahoo.com/group/ydn-delicious/

* Your email settings:
Individual Email | Traditional

* To change settings online go to:
http://groups.yahoo.com/group/ydn-delicious/join
(Yahoo! ID required)

* To change settings via email:
mailto:[EMAIL PROTECTED] 
mailto:[EMAIL PROTECTED]

* To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]

* Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/