Tracks Community Wiki

Linux Debian

2010-03-20T16:21:53+00:00

General Debian GNU/Linux installation

A detailed howto written by c00i90wn can be found at c00i90wn’s homepage here there explains how to install it on Debian or Ubuntu.

(These instructions were written by Thomas Themel)

Installing tracks on Debian is a bit more complicated than it should be, mainly because Debian splits ruby into a multitude of packages, while authors of ruby packages seemingly tend to assume that they are all installed.

Installing Tracks requires Ruby on Rails, which in turn requires RubyGems.

To install RubyGems, you need to install the following Debian packages:

ruby1.8
libzlib-ruby

libyaml-ruby
When these and their dependencies are set up, it should be possible to install !RubyGems:
sokrates:/home/themel/ruby/rubygems-0.8.4# ruby setup.rb all ---> bin [...] Successfully built RubyGem Name: sources Version: 0.0.1 File: sources-0.0.1.gem sokrates:/home/themel/ruby/rubygems-0.8.4#
For installation of the rails GEM, more Debian packages are required:
libdrb-ruby
liberb-ruby

rdoc
RubyGems can then install Ruby on Rails:
sokrates:/home/themel/ruby/rubygems-0.8.4# gem install rails Attempting local installation of 'rails' Local gem file not found: rails*.gem Attempting remote installation of 'rails' Updating Gem source index for: http://gems.rubyforge.org Install required dependency rake? [Yn] y Install required dependency activerecord? [Yn] Install required dependency actionpack? [Yn] Install required dependency actionmailer? [Yn] Successfully installed rails, version 0.9.5 Installing RDoc documentation for rails-0.9.5... WARNING: Generating RDoc on .gem that may not have RDoc. lib/rails_generator.rb:34:46: Skipping require of dynamic string: "#{path}/#{name}_generator.rb" Installing RDoc documentation for rake-0.4.15... Installing RDoc documentation for activerecord-1.6.0... Installing RDoc documentation for actionpack-1.4.0...
lib/action_controller/scaffolding.rb:87:37: Skipping require of dynamic string: "#{model_id.id2name}" Installing RDoc documentation for actionmailer-0.6.1... sokrates:/home/themel/ruby/rubygems-0.8.4#

To run tracks, we need another Debian package:
libwebrick-ruby
The final issue that I ran into is that the ruby MySQL client seems to expect the MySQL server socket in its default location in /tmp when the specified database host is localhost. Debian doesn’t place the server socket in /tmp/mysql.socket, but rather in /var/run/mysqld/mysqld.socket. This causes the error
No such file or directory - /tmp/mysql.sock
when starting tracks. To fix this, add a socket line to the appropriate environments in your config/database.yml:
development: adapter: mysql database: tracks host: localhost username: root password: socket: /var/run/mysqld/mysqld.sock
Happy GTDing!

Thomas Themel

Radagast has also had some success with Debian and Tracks, and posts detailed instructions here.

Ubuntu Dapper Drake

There are some very clear and straightforward instructions for Ubuntu Dapper Drake here.

Ubuntu 6.10 (Edgy Eft) Installation

Here are some clear instructions for getting rails and Edgy to play well together.

Debian Subdirectory Installation

Install notes for Debian Lenny, including specific details and patches required for installing in a subdirectory here

OpenSolaris

2010-02-12T15:36:38+00:00

Category:Installation
Submitted by carylt

Installing Tracks 1.7 on OpenSolaris 2009.06 using the MySQL 5.0 database in 10 steps

All the commands described in this document are to be executed by a user with the Primary Administrator role (The user created when installing OpenSolaris is granted this role automatically).

Step 1 – Download the Tracks 1.7 zip file, unzip it in your preferred location and change to that directory.

This document is written assuming the Tracks software is installed in /opt
$ cd /opt $ pfexec unzip tracks-current.zip $ cd tracks-1.7

Step 2 – Configure the OpenSolaris publisher repositories needed to install the MySQL Ruby Gem package

OpenSolaris 2009.06 is delivered with Ruby 1.8.7, which contains Ruby and RubyGems, but rather than recompiling ourselves the MySQL Ruby Gem it’s easier to simply install it directly from the OpenSolaris Contrib repository.

So we will need to add this repository to the list of available publishers of packages:



$ pfexec pkg set-publisher -O http://pkg.opensolaris.org/contrib Contrib

Now we can simply install the ruby-mysql package

Step 3 – Install ruby-mysql from the OpenSolaris contrib repository



$ pfexec pkg install ruby-mysql

Step 4 – Install, Start and Configure MySQL 5.0

Install MySQL 5.0:
$ pfexec pkg install SUNWmysql5

Start MySQL 5.0:
$ pfexec svcadm enable mysql

Creating a new user account:
$ /usr/mysql/5.0/bin/mysql -u root mysql> CREATE DATABASE tracks17; mysql> GRANT ALL ON tracks17.* TO 'mysql'@'localhost' IDENTIFIED BY 'mysql';

Step 5 – Workaround a known Ruby bug (present amongst others in OpenSolaris and BSD)

Edit the file /opt/tracks-1.7/vendor/rails/activesupport/lib/active_support/inflector.rb
Go to line 275, and add a “rescue” statement so that part of the code reads:



def transliterate(string)

Iconv.iconv('ascii//ignore//translit', 'utf-8', string).to_s

rescue

end

Step 6 – Install Rake



$ pfexec gem install rake

Step 7 – Configure the Tracks database section

Open file /opt/tracks-1.7/config/database.yml and edit the production section to look like:
production: adapter: mysql host: localhost username: mysql password: mysql database: tracks17

Step 8 – Configure the Tracks site section

Open file /opt/tracks-1.7/config/site.yml

Change the line saying salt: “change-me” to a string of your choice.
Change the line saying time_zone: “UTC” to your time zone.

You can use the command
$ /var/ruby/1.8/gem_home/bin/rake time:zones:local

to see all available timezones on your machine and test that your Ruby installation is working as intended.
If it doesn’t, chances are you missed Step 5.

Step 9 – Populate your database with the Tracks 1.7 schema



$ pfexec rake db:migrate RAILS_ENV=production

Step 10 – Start the application



$ pfexec script/server -e production

Now point your browser to http://127.0.0.1:3000 check that the application is working and create an admin user for Tracks.

Note: To ensure the Tracks application starts automatically at boot time, save the following startup script as /etc/rc3.d/S99tracks

File:S99tracks.txt

S99tracks

2010-02-12T15:35:23+00:00

Installation Index

2010-02-12T14:50:59+00:00

Category:Installation

General requirements

The most up-to-date instructions can be found in the manual on the homepage” You can run Tracks on your own (local) computer (MacOS X, Windows and Linux are all supported), or on your web host’s server, in which case you will have access to Tracks from any internet connected computer.
Specific instructions for various platforms:

OpenSolaris

Recommended Web Hosts

2010-02-09T04:00:36+00:00

Here are some web hosts recommended by users which allow you to install Tracks and therefore use it from any internet-connected computer:

TextDrive: Where bsag hosts the Tracks project, her blog and her own private version of Tracks. They have all you need to install Rails applications, and great technical support. (note: textdrive has been absorbed by Joyent – don’t know if that changes anything or not…)
Slicehost: Built for developers. We offer 256/512/1024 Xen VPS Slices with choice of OS (Ubuntu, CentOS, Gentoo). The 256 is perfect for hosting Tracks and we’re big fans. Full root access – run Apache, Lighttpd, Mongrel, Webrick – your choice. Please contact us or pop into the chatroom with any questions.
Site5: Installed and works without any problem. It allows users to optimize their sites many search engine optimization experts will agree.

FreeBSD

2009-12-24T10:34:19+00:00

Category:Installation

Tracks 1.7 installation on FreeBSD 8.0 with sqlite3 as database

Prerequisite: install this packages from ports:



sqlite3

ruby+oniguruma

ruby18-aspectr

ruby18-bdb

ruby18-dbd_sqlite3

ruby18-dbi

ruby18-fcgi

ruby18-gems

ruby18-iconv

ruby18-mysql

rubygem-actionmailer

rubygem-actionpack

rubygem-activerecord

rubygem-activeresource

rubygem-activesupport

rubygem-addressable

rubygem-data_objects

rubygem-deprecated

rubygem-do_mysql

rubygem-do_sqlite3

rubygem-extlib

rubygem-rack

rubygem-rails

rubygem-rake

rubygem-redcloth

rubygem-sqlite3

May be some of them aren’t needed (like the mysql-packages).

Set up Tracks

Download the Tracks zip file and unpack it to a suitable location, e. g. /usr/local

cd into the Tracks home directory and set up the sqlite database (only, if you don’t already have a tracks database!)
rake db:migrate RAILS_ENV=production

Test yor Tracks installation

After completing successfully, you can start Tracks from within the tracks home directory
script/server -e production

Now you should be able to access your Tracks installation by your browser (http://myserverip:3000)

Keyboard Shortcuts

2009-11-23T14:39:15+00:00

For some browsers/OSes, Alt will suffice in place of Alt+Shift.

Alt+Shift+t – Home
Alt+Shift+p – Projects
Alt+Shift+k – Tickler
Alt+Shift+c – Contexts

Alt+Shift+d – Completed Tasks
Alt+Shift+r – Manage recurring actions
Alt+Shift+l – Calendar of due actions
Alt+Shift+s – Show/Hide Notes
Alt+Shift+o – Show all notes

Alt+Shift+u – Preferences
Alt+Shift+i – Import and export data

Alt+Shift+n – Show/Hide Form (Add Action/Project/Context)

Database schema

2009-10-11T10:47:49+00:00

for people who do not have shell access or and are not able to use rake db:migrate the is the following sql to create databases

Please not that this the the schema for 1.8devel, but it should work fine for 1.7

-- MySQL dump 10.11
--
-- Host: localhost    Database: tracks_prod
-- ------------------------------------------------------
-- Server version    5.0.84

/*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */;
/*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */;
/*!40101 SET @OLD_COLLATION_CONNECTION=@@COLLATION_CONNECTION */;
/*!40101 SET NAMES utf8 */;
/*!40103 SET @OLD_TIME_ZONE=@@TIME_ZONE */;
/*!40103 SET TIME_ZONE='+00:00' */;
/*!40014 SET @OLD_UNIQUE_CHECKS=@@UNIQUE_CHECKS, UNIQUE_CHECKS=0 */;
/*!40014 SET @OLD_FOREIGN_KEY_CHECKS=@@FOREIGN_KEY_CHECKS, FOREIGN_KEY_CHECKS=0 */;
/*!40101 SET @OLD_SQL_MODE=@@SQL_MODE, SQL_MODE='NO_AUTO_VALUE_ON_ZERO' */;
/*!40111 SET @OLD_SQL_NOTES=@@SQL_NOTES, SQL_NOTES=0 */;

--
-- Table structure for table `contexts`
--

DROP TABLE IF EXISTS `contexts`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `contexts` (
  `id` int(11) NOT NULL auto_increment,
  `name` varchar(255) NOT NULL,
  `position` int(11) default '0',
  `hide` tinyint(1) default '0',
  `user_id` int(11) default '1',
  `created_at` datetime default NULL,
  `updated_at` datetime default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_contexts_on_user_id` (`user_id`),
  KEY `index_contexts_on_user_id_and_name` (`user_id`,`name`)
) ENGINE=InnoDB AUTO_INCREMENT=64 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `notes`
--

DROP TABLE IF EXISTS `notes`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `notes` (
  `id` int(11) NOT NULL auto_increment,
  `user_id` int(11) NOT NULL,
  `project_id` int(11) NOT NULL,
  `body` text,
  `created_at` datetime default NULL,
  `updated_at` datetime default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_notes_on_project_id` (`project_id`),
  KEY `index_notes_on_user_id` (`user_id`)
) ENGINE=InnoDB AUTO_INCREMENT=52 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `open_id_authentication_associations`
--

DROP TABLE IF EXISTS `open_id_authentication_associations`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `open_id_authentication_associations` (
  `id` int(11) NOT NULL auto_increment,
  `issued` int(11) default NULL,
  `lifetime` int(11) default NULL,
  `handle` varchar(255) default NULL,
  `assoc_type` varchar(255) default NULL,
  `server_url` blob,
  `secret` blob,
  PRIMARY KEY  (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `open_id_authentication_nonces`
--

DROP TABLE IF EXISTS `open_id_authentication_nonces`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `open_id_authentication_nonces` (
  `id` int(11) NOT NULL auto_increment,
  `timestamp` int(11) NOT NULL,
  `server_url` varchar(255) default NULL,
  `salt` varchar(255) NOT NULL,
  PRIMARY KEY  (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `preferences`
--

DROP TABLE IF EXISTS `preferences`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `preferences` (
  `id` int(11) NOT NULL auto_increment,
  `user_id` int(11) NOT NULL,
  `date_format` varchar(40) NOT NULL default '%d/%m/%Y',
  `week_starts` int(11) NOT NULL default '0',
  `show_number_completed` int(11) NOT NULL default '5',
  `staleness_starts` int(11) NOT NULL default '7',
  `show_completed_projects_in_sidebar` tinyint(1) NOT NULL default '1',
  `show_hidden_contexts_in_sidebar` tinyint(1) NOT NULL default '1',
  `due_style` int(11) NOT NULL default '0',
  `admin_email` varchar(255) NOT NULL default 'butshesagirl@rousette.org.uk',
  `refresh` int(11) NOT NULL default '0',
  `verbose_action_descriptors` tinyint(1) NOT NULL default '0',
  `show_hidden_projects_in_sidebar` tinyint(1) NOT NULL default '1',
  `time_zone` varchar(255) NOT NULL default 'London',
  `show_project_on_todo_done` tinyint(1) NOT NULL default '0',
  `title_date_format` varchar(255) NOT NULL default '%A, %d %B %Y',
  `mobile_todos_per_page` int(11) NOT NULL default '6',
  `sms_email` varchar(255) default NULL,
  `sms_context_id` int(11) default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_preferences_on_user_id` (`user_id`)
) ENGINE=InnoDB AUTO_INCREMENT=4 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `projects`
--

DROP TABLE IF EXISTS `projects`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `projects` (
  `id` int(11) NOT NULL auto_increment,
  `name` varchar(255) NOT NULL,
  `position` int(11) default '0',
  `user_id` int(11) default '1',
  `description` text,
  `state` varchar(20) NOT NULL default 'active',
  `created_at` datetime default NULL,
  `updated_at` datetime default NULL,
  `default_context_id` int(11) default NULL,
  `completed_at` datetime default NULL,
  `default_tags` varchar(255) default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_projects_on_user_id` (`user_id`),
  KEY `index_projects_on_user_id_and_name` (`user_id`,`name`)
) ENGINE=InnoDB AUTO_INCREMENT=190 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `recurring_todos`
--

DROP TABLE IF EXISTS `recurring_todos`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `recurring_todos` (
  `id` int(11) NOT NULL auto_increment,
  `user_id` int(11) default '1',
  `context_id` int(11) NOT NULL,
  `project_id` int(11) default NULL,
  `description` varchar(255) NOT NULL,
  `notes` text,
  `state` varchar(20) NOT NULL default 'active',
  `start_from` datetime default NULL,
  `ends_on` varchar(255) default NULL,
  `end_date` datetime default NULL,
  `number_of_occurences` int(11) default NULL,
  `occurences_count` int(11) default '0',
  `target` varchar(255) default NULL,
  `show_from_delta` int(11) default NULL,
  `recurring_period` varchar(255) default NULL,
  `recurrence_selector` int(11) default NULL,
  `every_other1` int(11) default NULL,
  `every_other2` int(11) default NULL,
  `every_other3` int(11) default NULL,
  `every_day` varchar(255) default NULL,
  `only_work_days` tinyint(1) default '0',
  `every_count` int(11) default NULL,
  `weekday` int(11) default NULL,
  `completed_at` datetime default NULL,
  `created_at` datetime default NULL,
  `updated_at` datetime default NULL,
  `show_always` tinyint(1) default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_recurring_todos_on_user_id` (`user_id`),
  KEY `index_recurring_todos_on_state` (`state`)
) ENGINE=InnoDB AUTO_INCREMENT=59 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `schema_migrations`
--

DROP TABLE IF EXISTS `schema_migrations`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `schema_migrations` (
  `version` varchar(255) NOT NULL,
  UNIQUE KEY `unique_schema_migrations` (`version`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `sessions`
--

DROP TABLE IF EXISTS `sessions`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `sessions` (
  `id` int(11) NOT NULL auto_increment,
  `session_id` varchar(255) default NULL,
  `data` text,
  `updated_at` datetime default NULL,
  PRIMARY KEY  (`id`),
  KEY `sessions_session_id_index` (`session_id`)
) ENGINE=InnoDB AUTO_INCREMENT=46388 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `taggings`
--

DROP TABLE IF EXISTS `taggings`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `taggings` (
  `id` int(11) NOT NULL auto_increment,
  `taggable_id` int(11) default NULL,
  `tag_id` int(11) default NULL,
  `taggable_type` varchar(255) default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_taggings_on_tag_id_and_taggable_id_and_taggable_type` (`tag_id`,`taggable_id`,`taggable_type`),
  KEY `index_taggings_on_taggable_id_and_taggable_type` (`taggable_id`,`taggable_type`),
  KEY `index_taggings_on_tag_id` (`tag_id`)
) ENGINE=InnoDB AUTO_INCREMENT=12747 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `tags`
--

DROP TABLE IF EXISTS `tags`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `tags` (
  `id` int(11) NOT NULL auto_increment,
  `name` varchar(255) default NULL,
  `created_at` datetime default NULL,
  `updated_at` datetime default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_tags_on_name` (`name`)
) ENGINE=InnoDB AUTO_INCREMENT=1026 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `todos`
--

DROP TABLE IF EXISTS `todos`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `todos` (
  `id` int(11) NOT NULL auto_increment,
  `context_id` int(11) NOT NULL,
  `project_id` int(11) default NULL,
  `description` varchar(255) NOT NULL,
  `notes` text,
  `created_at` datetime default NULL,
  `due` datetime default NULL,
  `completed_at` datetime default NULL,
  `user_id` int(11) default '1',
  `show_from` datetime default NULL,
  `state` varchar(20) NOT NULL default 'immediate',
  `recurring_todo_id` int(11) default NULL,
  `updated_at` datetime default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_todos_on_user_id_and_state` (`user_id`,`state`),
  KEY `index_todos_on_user_id_and_project_id` (`user_id`,`project_id`),
  KEY `index_todos_on_project_id` (`project_id`),
  KEY `index_todos_on_context_id` (`context_id`),
  KEY `index_todos_on_user_id_and_context_id` (`user_id`,`context_id`)
) ENGINE=InnoDB AUTO_INCREMENT=4939 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;

--
-- Table structure for table `users`
--

DROP TABLE IF EXISTS `users`;
/*!40101 SET @saved_cs_client     = @@character_set_client */;
/*!40101 SET character_set_client = utf8 */; CREATE TABLE `users` (
  `id` int(11) NOT NULL auto_increment,
  `login` varchar(80) NOT NULL,
  `crypted_password` varchar(40) default NULL,
  `token` varchar(255) default NULL,
  `is_admin` tinyint(1) NOT NULL default '0',
  `first_name` varchar(255) default NULL,
  `last_name` varchar(255) default NULL,
  `auth_type` varchar(255) NOT NULL default 'database',
  `open_id_url` varchar(255) default NULL,
  `remember_token` varchar(255) default NULL,
  `remember_token_expires_at` datetime default NULL,
  PRIMARY KEY  (`id`),
  KEY `index_users_on_login` (`login`)
) ENGINE=InnoDB AUTO_INCREMENT=4 DEFAULT CHARSET=latin1;
/*!40101 SET character_set_client = @saved_cs_client */;
/*!40103 SET TIME_ZONE=@OLD_TIME_ZONE */;

/*!40101 SET SQL_MODE=@OLD_SQL_MODE */;
/*!40014 SET FOREIGN_KEY_CHECKS=@OLD_FOREIGN_KEY_CHECKS */;
/*!40014 SET UNIQUE_CHECKS=@OLD_UNIQUE_CHECKS */;
/*!40101 SET CHARACTER_SET_CLIENT=@OLD_CHARACTER_SET_CLIENT */;
/*!40101 SET CHARACTER_SET_RESULTS=@OLD_CHARACTER_SET_RESULTS */;
/*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */;
/*!40111 SET SQL_NOTES=@OLD_SQL_NOTES */;

-- Dump completed on 2009-10-09 16:28:05 

Web host

2009-10-11T10:44:38+00:00

Installing Tracks on a webhost’s server

Database schema

ActionDependencies

2009-09-23T21:11:50+00:00

Category:Development

Adding dependencies to Tracks

This is meant to become a proposal for adding dependencies to Tracks. Please feel free to comment and update with your ideas.

Motivation

Actions in Tracks are supposed to be next actions. In reality actions could become dependent on other actions to be finished first. Tracks need to be able to handle this. Please note that Tracks is not a planning tool. Tracks is a tool for managing GTD-like next actions

Scope

The scope of this proposal is dependencies between actions.

Database

Todo state machine model

A pending state is added to the Todo state machine, and a new event block is introduced for switching from active to pending state.

Single vs multiple dependencies

This seem to be one of the main questions for debate, i.e. if an action should be able to depend on more than one action. Trying to summarize the discussion it boils down to the question:

Do the benefits of multiple dependencies outweigh the added complexity?

GUI

The dependency information is used in the GUI for displaying only actionable active actions. The views affected are: New/Edit Action, Home, Project, Tickler, Context, Tag and their mobile counterparts.

Adding the dependencies information

Use cases are:
1. Adding a predecessor to an existing action.
2. Adding a predecessor to a new action.
3. Removing a predecessor for an existing action.
4. Showing the predecessor of an action.

The add and edit action forms have an added field for adding predecessors named “Depends on”. Ideally, it would work as a Gmail-like multi value autocompletion field.

RB: I don’t know if the autocomplete will work. You need to know exactly the wording of the action. But other options are perhaps even more cumbersome to use (i.e. first select project and then the action)

HB: Advanced auto completion fields like Gmail’s and Firefox’s address field does not actually require exact wording. They start matching words inside sentences. Also, drag and drop as Yoichi mentioned may work to add dependencies, especially in project view.

Displaying the dependencies information

Use cases are:
1. The user only wants to see the actionable actions.
2. If an action is completed, any pending actions should be displayed.
3. The user wants to add or delete dependencies and see pending actions.

1. The user only wants to see actionable actions

In the Home and Context views, the GUI does not have to be visibly changed to implement a first version. The user will experience that only actionable actions are displayed.

2. If an action is completed, any pending actions should be displayed

In the Home, Context, Project and Tag views, if an action is completed any pending actions will become active.

3. The user wants to add or delete dependencies and see pending actions.

The Project view is probably the best view to view and edit dependencies and pending actions.

Example project view:

Let’s say that we have the following actions in our Add dependencies to Tracks project:

1. Create GUI proposal
2. Collect opinions
3. Implement revised GUI proposal
4. Learn Ruby
5. Learn ORM

The second action is dependent on the first action. The third action is dependent on the second, fourth and fifth action.

Using dependencies when displaying Active actions we get 1, 4 and 5, since they are not dependent on other actions. So, what about 2 and 3? How shall we display them? We could use a new section called “Blocked actions in this project” or rename the Deferred section. After the Context link the > and < links represents unique successors and uncompleted predecessors respectively.

The Create GUI proposal action is placed on top because it has the greatest number of successors (unblocks largest number of actions).
The Collect opinions action is placed on top of the blocked actions because it has the least number of predecessors (most likely to be unblocked first).

RB: Why do you want to show < and the >. I think the depends-on relation is enough. How relevant is information on the > relation?

HB: Especially for the active actions I think it is relevant to see the number of pending actions for an action. For the pending actions, it may be more relevant to see the predecessors.

Today we use due date, show date and entry order for calculating the display order for actions in the Home, Tickler and Project page. The current display order in the project page is as follows (TBC):

RB: I was thinking of adding a plus to a next action that has one or more dependent actions

HB: You mean in order to be able to expand/collapse the “depends on”/”pending” actions? I am pretty ok with that also, but see that more as a generic way to display more information about an action. If it is simple to implement, go for it, otherwise I think it would be enough with the simple visual cues given by < and > symbols with tool tips.

RB: I think it could be a good way to start with the dependencies with a container with the ‘blocked’ actions and later-on bring the plus to it. Indeed to be able to hide the non-next actions in the view, but be able to show them for review / edit

HB: Is this related to your concern of showing actions more than once? Can this information be display in a tool tip or pop up instead?

RB: with n-m yes. With 1-n, the plus is not easy to implement. So, just phasing here…

Active actions

Displayed in Home and project views.

Condition: show date=null or show date<=today.
Sort order: due date, entry order

Deferred actions

Displayed in Tickler and project views.

Condition: show date > today.
Sort order: show date, due date, entry order

Dependencies gives us one further condition for displaying actions.

Active actions in project view

RB: this is the same heading as before. Is this the project view versus the home page view above?

HB; The intention was rather project page in “No dependencies” vs “Dependencies” view

RB: ok, changed the title

This is quite straightforward. Filter on uncompleted predecessor count.

Condition: (show date=null or show date=today) and uncompleted predecessors=0
Sort order: due date, descending number of successors, entry order

Deferred and blocked actions

Treat a blocked action as having the latest show date of its predecessors. This makes the blocked action to be sorted into a place where it typically makes sense.

RB: I’m not sure what this means for the home page.

HB: The home page only displays active actions, so it will mostly affect the Project, Tickler and Tag views I guess?

Condition: show date > today or uncompleted predecessors > 0
Sort order: show date, due date, predecessor count, number of successors, entry order

RB: I’d like in a second phase
1. Create GUI proposal (+)

– Collect opinions – Implement revised GUI proposal
2. Learn Ruby
3. Learn ORM

RB: I find this very complex. Can’t we just use < and > and use the tooltip to sum up the dependencies. This is probably why a favor the plus :-)

HB: For users or developers (or both)? I hate it when users don’t understand how user friendly my apps are :-). Ok, my intention with multiple < and > was except displaying the number of dependencies, also the possibility to link or display information on the different actions. What you’re after is perhaps something more along the Notes toggle icon? So, what about some sort of “link” icon instead of the plus sign?

RB: in this case both :-) I think it is to much information in a few icons.

Open Issues

Do the benefits of multiple dependencies outweigh the added complexity?

Where do you see the most complexity in implementing a many-to-many relationship?

RB:
1. Technically: depending on how you show the relationships, it could mean that you display an action twice. All current views work on the assumption that an action is only shown once
2. Tracks is not a planning tool: the more complex relationships you allow, the more people expect Tracks to help plan a project. This is not what we want with Tracks.
3. I believe that in most (if not all) cases you can suffice with a 1-n relationship. This would mean changing Tracks minimally for dependencies. Less added complexity is always better imho
4. the gui would just get very complicated

HB:
1. Can you give me an example?
2. Ok, I trust your experience here.
3. I agree with less complexity. However, I think it is pretty common that an action can have multiple (independent) dependencies. Such as described in the example below, or let’s say “Buy bread”, “Buy butter”, which are independent actions doable in parallel, and “Make sandwich” which are dependent on both earlier actions.
4. Do you mean for the user or the developers? I think the most common use case would be to add a couple of actions to a project, in pretty much the order intended for execution, thus being able to add dependencies to existing actions. Without changing the GUI much (let’s say that we in the first step, skip any expandable sections, links to dependencies etc.), I still think most users would benefit from just using the dependency information when filtering out active next actions. The actual dependencies information would still be accessible and editable in the edit view.

RB:
1. I was thinking about a view with a plus to expand all dependent actions. If you have 1-n, the 1 will just expand the n actions. No problems here. If you have n-m, you can have action A which is dependent on actions B and C. On the home page, B and C will have a plus which will both show A when expanded.
2. Just trying to keep the scope (and be able to handle maintenance)
3. probably true, but let’s start simple. We can always enhance when the simple case works well
4. both, but again from the plus-point-of-view

Closed issues

Is it common to have dependencies between projects?

RB: yes, but I think we should leave that out of scope. You can always realize this in a tree with a 1-n relation between one action from project A and all actions from project B.
Naming
RB: I dislike the terms predecessor and successor. I find it very confusing looking at the action form in the mockup screen shot. Could we use “depends on”?

I also think we need to start this very simple, so a 1-n relation between predecessor and successor. Or an action depends on one other actions. That actions has zero or more actions that depend on it.

HB: I like “depends on” better in the GUI, but I think pred./succ. is sometimes more precise when discussing technical solutions.

RB: Ok, that sounds good

Tickets

It is partly related to (at least) the following tickets:
#105
#300 (most activity in this one)
#721 (duplicate)
#475
#185
#360

Some overlap with
#82
#91

Note: #903 tracks a GUI mockup suggestion.

Code

Code at git://github.com/bohrax/tracks.git