Export an Org file to TaskJuggler

TaskJuggler is a project management tool. It provides an optimizing scheduler that computes your project time lines and resource assignments based on the project outline and the constraints that you have provided.

The TaskJuggler exporter is a bit different from other exporters, such as the HTML and @LaTeX{} exporters for example, in that it does not export all the nodes of a document or strictly follow the order of the nodes in the document.

Instead the TaskJuggler exporter looks for a tree that defines the tasks and optionally trees that define the resources and reports for this project. It then creates a TaskJuggler file based on these trees and the attributes defined in all the nodes.

You need to have Org's contrib/ in your Emacs load-path and to (require 'ox-taskjuggler) to add the TaskJuggler export. See also the docstring of org-export-backends on how to add new backends by default.

When ox-taskjuggler.el is loaded, it defines two new handlers within C-c C-e (org-export-dispatch):

Create your tasks as you usually do with Org mode. Assign efforts to each task using properties (it is easiest to do this in the column view). You should end up with something similar to the example by Peter Jones in http://www.contextualdevelopment.com/static/artifacts/articles/2008/project-planning/project-planning.org. Now mark the top node of your tasks with a tag named :taskjuggler_project: (or whatever you customized org-export-taskjuggler-project-tag to). You are now ready to export the project plan with C-c C-e J which will export the project plan and open a gantt chart in TaskJugglerUI.

Next you can define resources and assign those to work on specific tasks. You can group your resources hierarchically. Tag the top node of the resources with :taskjuggler_resource: (or whatever you customized org-export-taskjuggler-resource-tag to). You can optionally assign an identifier (named resource_id) to the resources (using the standard Org properties commands, see the Property syntax section of the manual) or you can let the exporter generate identifiers automatically (the exporter picks the first word of the headline as the identifier as long as it is unique—see the documentation of org-taskjuggler-get-unique-id). Using that identifier you can then allocate resources to tasks. This is again done with the allocate property on the tasks. Do this in column view or when on the task type C-c C-x p allocate RET <resource_id> RET.

Once the allocations are done you can again export to TaskJuggler and check in the Resource Allocation Graph which person is working on what task at what time.

The exporter also takes TODO state information into consideration, i.e., if a task is marked as done it will have the corresponding attribute in TaskJuggler (complete 100). Scheduling information is also taken into account to set start/end dates for tasks.

The exporter will also export any property on a task resource or resource node which is known to TaskJuggler, such as limits, vacation, shift, booking, efficiency, journalentry, rate for resources or account, start, note, duration, end, journalentry, milestone, reference, responsible, scheduling, etc.@: for tasks.

The exporter will handle dependencies that are defined in the tasks either with the ORDERED attribute (@pxref{TODO dependencies}), with the BLOCKER attribute (see @file{org-depend.el}) or alternatively with a depends attribute. Both the BLOCKER and the depends attribute can be either previous-sibling or a reference to an identifier (named task_id) which is defined for another task in the project. BLOCKER and the depends attribute can define multiple dependencies separated by either space or comma. You can also specify optional attributes on the dependency by simply appending it. The following examples should illustrate this:

* Preparation
  :PROPERTIES:
  :task_id:  preparation
  :ORDERED:  t
  :END:
* Training material
  :PROPERTIES:
  :task_id:  training_material
  :ORDERED:  t
  :END:
** Markup Guidelines
   :PROPERTIES:
   :Effort:   2d
   :END:
** Workflow Guidelines
   :PROPERTIES:
   :Effort:   2d
   :END:
* Presentation
  :PROPERTIES:
  :Effort:   2d
  :BLOCKER:  training_material @{ gapduration 1d @} preparation
  :END:

TaskJuggler can produce many kinds of reports (e.g., gantt chart, resource allocation, etc). The user defines what kind of reports should be generated for a project in the TaskJuggler file. By default, the exporter will automatically insert some pre-set reports in the file. These defaults are defined in org-export-taskjuggler-default-reports. They can be modified using customize along with a number of other options. For a more complete list, see M-x customize-group RET org-export-taskjuggler RET.

Alternately, the user can tag a tree with org-export-taskjuggler-report-tag, and define reports in sub-nodes, similarly to what is done with tasks or resources. The properties used for report generation are defined in org-export-taskjuggler-valid-report-attributes. In addition, a special property named report-kind is used to define the kind of report one wants to generate (by default, a taskreport).

For more information and examples see the Org-taskjuggler tutorial at http://orgmode.org/worg/org-tutorials/org-taskjuggler.html.