Forgetful Browsing

Sam Foster sfoster at
Wed Jul 29 18:03:51 UTC 2015

Another way to look at it might be as an "undo" for a browsing action. This
has the advantage of being a familiar metaphor. Like when you click some
link that ends up somewhere you don't want in your history or cookie trail.
Or search for something and realize afterwards you should've done that
search in a private window. Maybe this ends up being the same as "move to
private window". What constitutes one step of undo might get complicated
with redirects, but on the face of it this seems a simpler problem to solve
(but maybe it amounts to the same thing?)


On Fri, Jul 17, 2015 at 6:34 AM, Ehsan Akhgari <ehsan.akhgari at>

> FWIW, from a technical perspective, this is extremely difficult to
> implement.  We typically don't remember which tab the information that we
> write to the disk has come from, so going back and cleaning things up after
> the fact requires a ton of technical work, and adding complexity to Gecko.
> Additionally, the page may have live resources (for example, a live channel
> that is downloading information from the network) and moving such live
> resources to act as if they were opened in PB mode in the first place is
> also a huge project of its own.
> On Fri, Jul 17, 2015 at 9:12 AM, Staś Małolepszy <stas at> wrote:
>> (This is also available at,
>> with webm illustrations.)
>> I'd like to suggest a new feature in Firefox which combines Private
>> Browsing with the Forget Button and emphasizes Mozilla's commitment
>> to user's privacy by extending the scope of forgetting to partners.
>> tl;dr
>>   The user should be able to retroactively create private browsing
>>   contexts from already open tabs.  The browsing history for those tabs
>>   will be forgotten by the user agent when the context is closed.
>>   Additionally, the search query will be forgotten by the search engine
>>   if one of the ship-by-default engines was used.
>> ## Problem Definition
>> We know that our users abuse private browsing to have two email
>> accounts open side-by-side or to do a quick search (often times on
>> behalf of someone else) that will not be remembered by the browser nor
>> by the search engine.
>> Switching modes is a hard task, however.  Users have to remember to
>> start by opening a new private browsing window in the first place.
>> The idea to use a private window often comes as an afterthought.  The
>> only solution then is to use the Forget Button which is a drastic
>> measure:  all tabs and windows are closed and all history from
>> X last minutes is deleted.
>> If private browsing is an a priori method then the Forget Button is
>> a posteriori.  There's a need for a solution which sits in the
>> middle and puts users in control of what should be forgotten during the
>> entire lifecycle of the task.
>> ## Retroactive & Selective
>> Forgetful Browsing combines the best of two worlds.  It's forgiving to
>> the user because it's retroactive and less drastic because it's
>> selective.
>> For the purpose of this document I chose Tab Groups as the
>> vehicle for demoing the main concepts behind Forgetful Browsing.
>> ## How it works
>> At any point in time the user can make any set of tabs private.  In
>> order to do so, the user creates a new tab group in the Tab Groups view
>> or right-clicks on the tab title and selects Move To… > New Private
>> Group.
>> Upon closing of a forgetful tab, the entire browsing history from the
>> tab is deleted as if the tab was created private.  All cookies created
>> or modified during the lifetime of the tab are also deleted.
>>   Note: The cookie behavior is different from regular private tabs.
>>   In Forgetful Browsing modified cookies are also cleared which may
>>   cause the user to become logged out from her usual websites.
>> Additionally, the user agent makes an API request to search engines and
>> ad networks participating in the Forgetful Browsing movement which
>> makes them forget this part of the user's browsing history.
>> ## Why It Works
>> Forgetful Browsing allows users to rectify the decision about browsing
>> privately at any point in time.  There's no need to decide upfront.  It
>> reduces the user's cognitive load and enforces the feeling of being in
>> control of her online experience.
>> It also helps establish a cooperative relationship between the user and
>> content providers:  it's opt-in, selective and occasional.  It works
>> similar to the Do Not Track header, but the it's easier to respect for
>> content providers and has an immediate effect visible to the user.
>> Last but not least, Forgetful Browsing provides a validation to the
>> user as to why certain search engines are shipped by default with the
>> user agent.  These search engines are the ones that care about the
>> user's privacy and allow users to delete parts of their browsing
>> history.
>> (Search engines are a start;  the network of privacy-aware websites and
>> services can grow with time.  All that is required is an API entry
>> point which the user agent can use to make the request on behalf of the
>> user.)
>> -stas
>> _______________________________________________
>> firefox-dev mailing list
>> firefox-dev at
> --
> Ehsan
> _______________________________________________
> firefox-dev mailing list
> firefox-dev at
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the firefox-dev mailing list