views:

660

answers:

3

We are an all Unix shop (Solaris, Linux). This last product cycle I returned to a project lead capacity, and needed to produce a schedule. I asked what tools my managers would accept, and was surprised to hear "text files". My teammate and I gamely tried this, and probably worse, HTML tables, to track the tasks we wanted to size. It was pretty painful.

We then tried a few tools. MrProject is buggy, limited and crashes too frequently. My manager swears that Microsoft Project is inflexible. Whenever they needed to change a task, reassign a resource or rebalance, it generally hosed their plan. So I started looking around on the Internet for a Linux-capable project planning tool. One that sounded interesting is TaskJuggler. It's neat in that the inputs are declarative files. I feel like I'm building a makefile for a project.

However. I have a limited amount of time to devote to evaluating this tool and it seems pretty complex. Before diving into the next product cycle, I'd like to know if TaskJuggler is robust enough, flexible and capable of handling multi-month, multiple resource projects with frequent changes. So I'm calling on all engineers who have had experience with this tool to share their insights. Thanks!

+1  A: 

TaskJuggler's syntax is rather easy, but do take your time to read the documentation. My experience with TJ:

  • very powerful and expressive syntax
  • useful for detailed calculation of large projects

However in reality a manual planning takes into account many implicit constraints, which TJ requires to be made explicit in order to obtain realistic scenario's. This is of course true for every planning tool, but I found it rather cumbersome to add and edit manual constraints in large projects in TJ... Therefore I found it less suited for project tracking and rescheduling afterwards.

I now use OmniPlanner, which is a far simpler tool than TJ and MSProject but turns out to suit my needs (especially in tracking, analysis and reporting).

Dong Hoon
Thanks. I actually looked at OmniPlanner briefly, but the idea of hauling my personal iBook into work every day and possibly having to share it to let team co-leaders input their numbers put me off that.I'll try TJ on a more constrained project, but keep my expectations low for the larger ones.
Don Wakefield
+3  A: 

There is nothing free in project management, and managing a complex project with software is inevitably complex. The real question is, does the chosen tool help with this?

Task Juggler has a learning curve, and in the end is suitable for someone who doesn't mind reading the manual (an absolute necessity for this tool) and isn't tied to graphical input. Task Juggler requires that you think about your project and structure it in a meaningful way. It is helpful if you do a diagram in advance (many TJ users make mind maps and there is a tool out there somewhere to generate TJ input statements from a FreeMind mind map). It is also very helpful if you organize your input file in some meaningful way, making things easy to find.

That said, once you get going, creating a project with TJ is super fast. You don't need to bother with a million dialog boxes, you just tell TJ what you want in TJ's text language.

But all of that aside, what I like about TJ (and hated at first, coming from a legacy of other more traditional tools) is that it ensures that your schedule makes sense. OpenProj happily schedules resources at 300% and more. TJ will give you an error and make you fix it. Yes, it's annoying. But the end result is that you have a project schedule that makes sense and can actually be executed. Imagine that!

As I started out by saying, nothing's free. TJ requires study and some effort. The reward is rich and copious reporting, all the information you need to manage your project to cost and schedule, and the enforcement of a logical, reliable approach to scheduling and resource allocation. And it doesn't cost $499 or whatever MSP goes for --- it's free.

Bob Newell PMP
Thanks, Bob. It's nice to hear a largely positive story. I put TJ aside based on the slight response and limited time, but you lead me to reconsider.
Don Wakefield
+1  A: 

I have been using taskjuggler for the past 4/5 years now (4 projects with an average duration of a year or more). I find it very useful to create my initial estimates of

  • how long the project will take
  • When will each resource group be freed.
  • What if we added more resources with varying level of experience and efficiency to different domains of the project.

Typically the kind of stuff that upper management will ask you about your schedule can be generated much faster and at a more accurate granularity compared to doing something similar using MS Project or other GUI based tools.

Till recently I was using taskjuggler to get my initial schedule and using ms excel to track the project.

This is the first time I am using task juggler to actually track the project on a weekly basis. and so far the results look good.

Vijay
I noticed the 'initial' hedge throughout most of your comment. I'd be very interested to hear how the iterative use goes in your next project(s). Thanks for the info.
Don Wakefield