Story Web Browsing History Browser — Manual

• Anything is automatically derived from browsing logs. During browsing, there is no special user activity – like bookmarking – required.
• Orientation by means of the terms looked up via Google searches.
• Any information about the browsing history fully remains private. No social networking.
• Runs as a Web server, providing a Web user interface.
• Implemented in SWI Prolog, utilizing the new Places database interface to browsing history information of Firefox 3.
• GNU General Public License.

Contents

1. Installation
   1. Required Further Software
   2. Setting up the System
   3. Providing Places Databases
2. Usage
   1. Invocation
   2. Password Protection
   3. Page Types
      • Control
      • Searches
      • Sessions
      • Full History

1. Installation

1.1 Required Further Software

$ sudo port install xft2

1.2 Setting up the System

1. Unpack the distribution archive. A directory story is created.

   $ tar xzvf story.tgz

2. Move the story directory to some place. Write access to the story directory is not required. By default, the application assumes story is placed directly in your home directory. If another location is used, set the environment variable STORYDIR, for example by a line like the following in ~/.bash_profile:

   export STORYDIR=/usr/local/software/story

3. Create the user specific directory ~/.story with configuration files and stores for temporary data with the command

   $ story/bin/install-story-skel.sh

   This command creates a password file ~/.story/passwd with an entry for a single user of the Web server interface of the system. The user name is by default the value of $USER, but can also be supplied as optional argument to install-story-skel.sh. The password is prompted by install-story-skel.sh.

   If ~/.story already exists, then install-story-skel.sh aborts without any effect.

   Configuration settings are in ~/.story/config.pl. This file can of course be manually adapted to specific needs.

1.3 Providing Places Databases

The places.sqlite file can be found in your Firefox profile folder, for example on Linux

    ~/.mozilla/firefox/xxxxxxxx.default/places.sqlite

    ~/Library/Application Support/Firefox/Profiles/xxxxxxxx.default/places.sqlite.

A small example of a places database is provided with the distribution as

    story/doc/examples/places.sqlite.

History Time Span. Obviously, history databases are much more useful if the history spans longer time than the default 9 days (changed to 180 days in recent versions of Firefox 3). You can set this in the Firefox preferences: Edit -> Preferences -> Privacy: History, Remember visited pages for the last ... days.

Conversion from a Firefox 2 profile. Since you might not want to use Firefox 3 regularly while in alpha status, you can convert history data from a Firefox 2 profile folder, by just invoking Firefox 3 once. The places.sqlite file is then automatically generated, and you can copy it from the Firefox profile folder to some other place. To ensure that your Firefox profile folder is not permanently affected by the one-time invocation of Firefox 3, you can reconstruct it from a backup after quitting Firefox 3, or you can run Firefox 3 as a different user with a copy of your original profile folder.

2. Usage

2.1 Invocation

1. Start the Web server with

   $ story/bin/story-server.sh

   Optionally a port number can be given as argument to story-server.sh. The default port is 2210.

2. Open the root page of the server in a Web browser. For the default port, its URL is:

   http://localhost:2210/

2.2 Password Protection

2.3 Page Types

The server provides pages of four types:

• Control
• Searches
• Sessions
• Full History

Control Page

A Control page allows to perform actions like loading places SQLite database files and changing configuration. It also shows information about the status of the system.

The server root points to a Control page.

Searches Page

The Searches page provides a kind of index or abstract of the entire history by just showing a choice of search terms of performed Google searches and associated date information. Visits are grouped into sessions, as specified with a Control page.

If there already is a browser window displaying the Searches page, this window is re-used, effecting just a re-positioning in that window. Hence we speak of the (instead of a) Searches page. This is implemented with the HTML target attribute.

Sessions Page

Full History Page

Its upper part is like a Searches page, and its lower part like a Sessions page, but for all sessions in the entire history. The image shows the main differences to these page types.