[Epiphany] Re: [Usability] User defined metadata (was: epiphanytoolbar/bookmarks)

> Definately!
> However, not everything on gnomedesktop is flame. :) 
> http://gnomedesktop.org/comments.php?op=showreply&tid=11890&sid=1135&pid=11844

I was not implying everything is a flame :)

> This posting describes a very real problem of the topic approach: He
> can't classify his bookmarks with the current system. So he has a topic
> "Jobs", but how does he classify a bookmark as "far" or "near" or puts
> it into one if his defined job categories (like "consulting")? 
> Automatic metadata (like fetched from the website) won't help at all,
> because as much as this is improved, it won't be able to set things
> depending on the user, like his location. He could create more topics
> and then put a bookmark into the "Jobs", "Near Companies" and
> "Consulting" topics but this would be weird and the topic list would
> quickly grow into unusability.
> While this might not be important for most users, it might be a rough to
> just say "deal with it", considering that it's a pretty heavy
> regression. The solution probably is to allow the user to add own
> metadata to a topic. This isn't very easy though, if you just allow the
> user to add his own keywords, this will be anything but userfriendly
> (you have to use your keyboard and remember which keywords you used
> before). While replying to this posting, I wondered how this could be
> done in a simple, yet flexible way (you can read my reply there if you
> want to skip this longer email).

This is quite an extreme case ... not sure how many real users will need
it but ... It was an interesting challenge, so I did a quick self test
on it.
I thought a more generic solution to the far/near problem was to allow
quality metadata in epiphany. Which in practice means to allow the user
to rate bookmarks like in rhythmbox or (read only) on some web sites.
Not sure if this actually necessary but it's an interesting feature at
(I didnt code it but I tried to simulate ...)

Approximative results:


Step 1: Creating the bookmarks hierarchy in mozilla and adding 3 far and
3 near Consultant companies. ~ 3 minutes
(I'll do describe the pain it was to organize them, I didnt remember it
was so difficult ;))
Step 2: Open one by one 3 near sites. First time 40s, after some
training 20s


Step 1: File 6 bookmarks, with title: Consulting - Name of the company
in the default Work category. ~2 minutes. (Another option could have
been to put the Consulting word in an ipotetical notes textview.)
Step 2: Open bookmarks editor, select Work topic (maybe necessary if you
have other bookmarks with Consulting in it), type con. Double click on
the bookmarks one by one. 20s first time, 10s after some traning.

I'll add that I felt very hard both editing the tree and accessing all
the submenus levels. In epiphany I had to perform only very simple
actions, clicking and typing basically.
(You can imagine how fun is to Bookmarks->Jobs->Near->Consulting for
each site you want to visit.)

The timing is very unscientific and it's mostly a joke. But the example
is supposed to show a way to obtain the same result with simpler
actions. Am I missing something ?


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]