Prototype for about:me page
Margaret Leibovic
I'm an intern working at the Mountain View office this summer, and I've
been working on developing an about:me page for the browser. The purpose
of this page is to give users a statistical analysis of their browsing
activity, primarily for their amusement. A full description of the
project can be found here:
https://wiki.mozilla.org/Firefox/Sprints/about:me.
I created an add-on to share a prototype of this page, which you can
download here:
https://services.forerunnerdesigns.com/extensions/get/abo...@test.mozilla.com/0.1/.
Please install it and let me know what your think! (You need Firefox 3.5
or above.)
I'm aware that the page can be a little slow to load if you have a large
history, but it's just a prototype, so I'll work on fixing that later :)
I'm currently grouping websites by domain, rather than by individual
page, mostly to avoid redundancy between pages with different urls but
the same content (i.e. mail.google.com and gmail.com). However, this
reduces the level of detail of the individual entries, which may or may
not be a good thing, depending on whether you visit the same page all
the time on a given website (such as a specific reference page on
developer help site) or lots of different pages on the same website
(such as lots of different facebook profile pages). If you have any good
ideas about how to address this issue in a way that would make users
happy, please let me know!
Thanks!
Margaret
Boriss
https://services.forerunnerdesigns.com/extensions/get/abo...@test.mozilla.com/0.1/
The grouping by domain or URL problem is quite interesting - some
feedback on this would be really helpful!
- Boriss
On Jun 10, 3:38 pm, Margaret Leibovic <mleibo...@mozilla.com> wrote:
> Hi all,
>
> I'm an intern working at the Mountain View office this summer, and I've
> been working on developing an about:me page for the browser. The purpose
> of this page is to give users a statistical analysis of their browsing
> activity, primarily for their amusement. A full description of the
> project can be found here:https://wiki.mozilla.org/Firefox/Sprints/about:me.
>
> I created an add-on to share a prototype of this page, which you can
> download here:
> https://services.forerunnerdesigns.com/extensions/get/abou...@test.mozilla.com/0.1/.
Boriss
I noticed a couple small problems with how some domains show up on the
top 10 list:
- Right now, there are separate entries if the user visits a page on
different country code domains (google.com and google.co.uk show up as
independent entries).
- http://twitter.com/ is showing up in the list as "Twitter /
Maintenance"
- Facebook shows up as "http://www.facebook.com/" rather than
"Facebook" or "Facebook | Home"
- While my top 10 seem roughly ok for the most part, something weird
may be going on... I notice that my email, something I visit multiple
times a day, isn't there - but "firefox logo - Google Image Search" is
On Jun 10, 4:03 pm, Boriss <flyingtoas...@gmail.com> wrote:
> The URL to get the addon should be:
> https://services.forerunnerdesigns.com/extensions/get/abou...@test.mozilla.com/0.1/
Mike Connor
On 10-Jun-09, at 4:35 PM, Boriss wrote:
> This is coming along awesomely!
>
> I noticed a couple small problems with how some domains show up on the
> top 10 list:
>
> - Right now, there are separate entries if the user visits a page on
> different country code domains (google.com and google.co.uk show up as
> independent entries).
>
> - http://twitter.com/ is showing up in the list as "Twitter /
> Maintenance"
That's probably the last title we had for that URL...
> - Facebook shows up as "http://www.facebook.com/" rather than
> "Facebook" or "Facebook | Home"
Redirects suck, I bet wikipedia and other sites that have some path
for their homepage have the same problem.
> - While my top 10 seem roughly ok for the most part, something weird
> may be going on... I notice that my email, something I visit multiple
> times a day, isn't there - but "firefox logo - Google Image Search" is
Hmm, frecency? If your mail is always in a tab, it won't show up
here. I wonder if we want to have some sort of "time this tab is
frontmost" boost to frecency/visits to help this case in general...
not sure.
-- Mike
Dave Townsend
>
> On 10-Jun-09, at 4:35 PM, Boriss wrote:
>
>> This is coming along awesomely!
It is indeed awesome. It is both nice and yet deeply disturbing to know
that I have been using bugzilla at 4am in the morning.
>>
>> I noticed a couple small problems with how some domains show up on the
>> top 10 list:
>>
>> - Right now, there are separate entries if the user visits a page on
>> different country code domains (google.com and google.co.uk show up as
>> independent entries).
>>
>> - http://twitter.com/ is showing up in the list as "Twitter /
>> Maintenance"
>
> That's probably the last title we had for that URL...
There might be something weirder going on. I have
http://tinderbox.mozilla.org/Firefox listed as "Nomadix"?!?
>
>> - Facebook shows up as "http://www.facebook.com/" rather than
>> "Facebook" or "Facebook | Home"
>
> Redirects suck, I bet wikipedia and other sites that have some path for
> their homepage have the same problem.
>
>> - While my top 10 seem roughly ok for the most part, something weird
>> may be going on... I notice that my email, something I visit multiple
>> times a day, isn't there - but "firefox logo - Google Image Search" is
If I understand this, the top 10 is the top 10 domains. Then each result
shows I guess the most popular page in that domain. Is your email on the
same domain as something else in the list?
Blair McBride
history as data (using total visits, not frecency). I also grouped by
domain, and found the results both interesting and appropriate.
http://skitch.com/unfocused/bu4ut/places-treemap
So yes, grouping by domain gets my vote. Perhaps with a secondary
contextual display of the top URLs within that domain.
- Blair
keith_w
>
> I'm an intern working at the Mountain View office this summer, and I've
> been working on developing an about:me page for the browser. The purpose
> of this page is to give users a statistical analysis of their browsing
> activity, primarily for their amusement. A full description of the
> project can be found here:
> https://wiki.mozilla.org/Firefox/Sprints/about:me.
This line above got underlined and turned into a clickable URL, but
the one three lines below didn't.
Any idea why? Actually, it looks like the URL got crippled after "/get/
abou..." and therefore is not a valid url.
keith whaley
kei...@dslextreme.com
> I created an add-on to share a prototype of this page, which you can
> download here:
> https://services.forerunnerdesigns.com/extensions/get/abou...@test.mozilla.com/0.1/.
Robert Buecheler
On FF 3.0.10 and 3.0.11 I can see the original link that Margaret
Leibovic provided and follow it withouth problems.
Replies by Boriss and keith_w suggest Gecko based agents: Seamonkey
and Minefield.
They seem to "truncate" long URLs (actually URIs containing the "@"
sign get three chars before the sign replaced by three dots by those
agents). I am not sure, but it might be only the visual part that gets
truncated. The underlying link is probably still intact and
"clickable".
--
Robi
> _______________________________________________
> dev-apps-firefox mailing list
> dev-apps...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-apps-firefox
>
Keith Whaley
> I don't know what causes it, but I suspect your news reader.
> On FF 3.0.10 and 3.0.11 I can see the original link that Margaret
> Replies by Boriss and keith_w suggest Gecko based agents: Seamonkey
> and Minefield.
> They seem to "truncate" long URLs (actually URIs containing the "@"
> sign get three chars before the sign replaced by three dots by those
> agents). I am not sure, but it might be only the visual part that gets
> truncated. The underlying link is probably still intact and
> "clickable".
>
I hear what you're saying, Robert, but...unless the URL address turns a
contrasting color and/or becomes underlined, it is not clickable, and
will not open.
If my SeaMonkey acted that way all the time I'd be most happy, but it
doesn't.
Is that a SeaMonkey (or Gecko) problem?
With URLs that are non-responsive like that I have to copy and paste to
get it into a Navigator URL field (what DO they call that little box?)
THEN open it.
Thanks,
keith whaley
David McRitchie
> and Minefield.
> They seem to "truncate" long URLs (actually URIs containing the "@"
> sign get three chars before the sign replaced by three dots by those
> agents).
That is Google, does wonders with email addresses, message-id, and code,
supposedly protects people from spam, but it just happens to help as
a Google stranglehold on newsgroups in combination with not being able
to use the message-id to find a message in newsgroup archives held by Google,
and breaking links that rely on message-id.
Gijs Kruitbosch
> <snip>
>>>
>>> I noticed a couple small problems with how some domains show up on the
>>> top 10 list:
>>>
>>> - Right now, there are separate entries if the user visits a page on
>>> different country code domains (google.com and google.co.uk show up as
>>> independent entries).
>>>
>>> - http://twitter.com/ is showing up in the list as "Twitter /
>>> Maintenance"
>>
>> That's probably the last title we had for that URL...
>
> There might be something weirder going on. I have
> http://tinderbox.mozilla.org/Firefox listed as "Nomadix"?!?
That's a wifi provider thing though, so it might well be a redirect too...
~ Gijs
Andreas
this sentence I wrote the last, but I have to say it the first: This
post is not for promoting the extension I made. It's for giving you new
and maybe interesting ideas of what could be sometime make it into
Firefox itself!
I've done something similiar to your extension.
In my diploma thesis which I just finished I examined how to help the
user finding already visited websites again, using the local history. I
created an extension called "WebReview"
(http://dragontalk.opendfki.de/wiki/WebReview and
https://addons.mozilla.org/firefox/addon/11315).
Currently, the built-in methods of Firefox don't support the user very
much. Think about the history sidebar, where everytime you enter a
search string, all sorting and filtering is reset. You can't search in a
specific date range for a keyword.
Sorting by "most visited" just gives me an endless list of urls without
further information. I'd say that is to much information on the one hand
and to less on the other.
And you don't see how these pages belong to each other, which makes it
hard to refind information that is distributed over several pages.
So, the first thing that happened at the beginning of my thesis was that
I was amazed how many valuable information is stored in the Places
database of Firefox. I wondered why not letting the user access and use
them?
For the regular visited websites, I created a page called about:mystart
which is designed as a personal and adaptive homepage (and page for a
newly created tab) for the user. For a better recognition, every page is
represented by it's screenshot, that have been taken by the extension.
The data for building this page is gathered from the local history.
It shows sites you visit every day and - and that's very cool - it
proposes you sites that you would visit this current day (based on the
last access and the visit frequency, see below). That means, that
section will change every day. Just think about visiting a news magazine
every monday. Instead of manually entering the url (or clicking a
bookmark), it is automatically proposed every monday.
Furthermore, about:mystart shows you a list of just closed tabs to
restore them quickly without navigating to the menu bar and the most
visited pages.
But these pages aren't just the "most visited" from Firefox.
In the first few weeks of my thesis, I found out that a simple visit
counter is not enough for a large database with entries which are some
months old. It doesn't tell you how often (over a longer interval) you
really visited that site. A site with 10 visit could be visited on 1 day
only, or even on 10 days once a day. And re-visiting pages on the same
day happen quite often.
So what I did is invented a value called DayVisits. These are visits on
a daily basis. No matter if you visit a page only once or 10 times on
one day, only one DayVisit will be counted. The advantage is, if you
have a site with 25 DayVisits, you know for sure that you have visited
it exactly on 25 days. That gives you more meaningful data over a longer
distance.
But DayVisits alone are not enough. They don't tell you, how regularly
you visited a site. Therefore I added a visit frequency, also
daily-based. (Unfortunately the frecency value of Firefox is artifical,
so it does not allow that.)
DayVisits and Frequency are calculated from the Places database at first
start retroactively and refreshed everytime you visit a page!
You can see a screenshot of about:mystart at
http://dragontalk.opendfki.de/wiki/WebReview/Features/MyStart
I have to admit that it's not super eye-candy yet (!), but there's
barely time for killer designs within the few months of a thesis.
It was more important that it actually works and you can see something.
Another feature of WebReview is a more enhanced chronicle, which you can
see at http://dragontalk.opendfki.de/wiki/WebReview/Features/Chronicle
It groupes all visited pages by domain and subdomain (which works as a
quick-filter when you click on an entry) and allows you to search in a
specific date range or with a minimum/maximum of (Day)Visits. For better
recognition, I used the screenshots here too. With these search
criteria, you can narrow down your results a lot more. A deletion of
pages is also possible.
The third thing I worked on is an enhancement of the visualization of
the browsing session. You may know about the forward/backward-problem
when clicking on links while you have predecessor and successors,
eventually you can't navigate to some of the pages you just visited.
It's very great that the Places database stores the relationship between
two pages, so I could access it and visualize the browsing session in a
tree-like structure:
http://dragontalk.opendfki.de/wiki/WebReview/Features/Graph
It shows you every page you visited in a particular session.
You can navigate not only in the current session, but also every session
that is stored in the Places database, no matter how old it is, even
before the installation of the extension.
This extension wasn't just about making something that you use once in a
while. There are quite a lot of extensions that show you a visualization
of the current session.
But one of the major goals was to provide access to data even before
installation of the extension. So, you don't depend on when you
installed it. You can just use it as it was always there (assumed you do
not wipe your history). I've found no other add-on that does that.
So, this was a very long post, but I hope I could explain why and how I
did some things and give something back of the great work I got with
Firefox.
If you have questions or comments or suggestions, please don't hesitate,
I really appreciate them!
If I can help, just let me know.
Greetings,
Andreas
Marco Bonardo
> Hi all,
>
> this sentence I wrote the last, but I have to say it the first: This
> post is not for promoting the extension I made. It's for giving you new
> and maybe interesting ideas of what could be sometime make it into
> Firefox itself!
I suggested Andreas to post here, since he analyzed our history usage
for quite some time, and has some interesting ideas to share.
Margaret could maybe spend some time talking with him in #places channel
(his nick is TheOne), to gather some useful experience about working
with history stats.
An ideas exchange would definately be welcome.
Marco
Andy Edmonds
also been working on analyzing places. In the mainline analysis, I'm
posting R code to the mozilla wiki for looking at the stats, summary
at http://surfmind.com/muzings/?p=339
Specifically for about me, I think search volume stats would also be a
cool bit of activity summarization. I noted such on the about:me wiki
page (https://wiki.mozilla.org/Firefox/Sprints/about:me) and also
posted SQL for more advanced metrics on search success by provider:
https://wiki.mozilla.org/User_talk:Andyed
The domain vs URL issue is a big one and regularly causes Andreas'
"about:mystart" page to be less than optimal for me. The best idea I
have for dealing with this is to take a combination of frecency and
*hub* behavior. The page that generates the most referrals to other
pages on that domain is likely the most representative URL. So sum
(from_visit) for a self join to the history table on the same domain.
Perhaps you could even get away without the same domain restriction
and visit self-join and just sum from_visit for all the place_ids from
a the specific domain. That's not the fastest query, but while you
can do really neat stuff with moz_historyvisits self joins, it's
rather expensive!
Can't wait for 3.5 to go gold so I can get more folks using
about:me!
On Jun 16, 5:33 am, Marco Bonardo <mak77NONSPA...@supereva.it> wrote:
...
Peter Lairo
> (his nick is TheOne)
I sure hope that is in reference to Neo from The Matrix, and not that
other one.
--
Regards,
Peter Lairo
The browser you can trust: www.Firefox.com
Reclaim Your Inbox: www.GetThunderbird.com
Islam: http://www.jihadwatch.org/islam101/
Israel: http://www.jewishvirtuallibrary.org/jsource/myths2/
Church of the Flying Spaghetti Monster: http://www.venganza.org/
"So, why don't we ever talk about the sun's contribution to global
warming? Well, because we can't regulate it, tax it, or make it feel
guilty for what it's doing" (www.WhatYouOughtToKnow.com)