I’ve just begun using Imendio Planner (formerly known as Mr. Project) for project management at work. I’ve never used such kind of software before so I don’t really know how people usually work with it, but I do have some idea of what I need.

The program, though, works somewhat differently as I expected (if you are curious about the specific problem I have with it just scroll to the bottom of this post). I talked about it with Albert (a coworker), David and Pere and they all agreed with me that the program would be much more useful with that feature.

Afterwards I posted a question to the planner mailing list exposing this issue. People at the mailing list were very kind and commented possible solutions and workarounds.

After a few emails from different people, a poster wrote (I reproduce it here for your convenience since this email comes from a public mailing list):

<rant>
It’s funny that what’s perfectly obvious to someone new to project management is ignored by most experts.

I have to agree with Josep that any tool that does project management has to make resource constraints a first-order problem, and not an afterthought. I’ve yet to see any widely used tool that does this; Primavera, MS Project, etc don’t do it, they all require you to go through hoops to do resource leveling and most of the time you end up having to create artificial temporal dependencies such as the suggestion given to Josep. These dependencies are at best a workaround, but in practice they can be detrimental to the project because they obscure the possibility of doing tasks in parallel by the “simple” addition of resources to a task, in practice, as the project progresses you figure out that the link is artificial and often end up doing the task in
parallel anyway, thereby throwing your hands in the air and saying: “screw this project plan”, or even worse, spending another couple of hours redoing the whole thing (and then an hour’s meeting explaining the “new and improved” project plan to the team.

Which is why there exist better methodologies than Critical Path analysis (yeah, I know, I repeat myself) such as Critical Chain management, that take into account resource constraints from the beginning.
</rant>

To be honest, several people replied to him with good arguments. My intention is not to assess which view is right. Probably it is not, but it could be, at least potentially, one of those situations in which “experts are biased” because of too much experience, or too many preconceptions, …

Experts often are right, but be aware because there are several cases in which they could in fact be wrong:

• Experts can be biased because of experience, academic knowledge, and so on…
• Sometimes experts don’t question opinions of other experts because it could turn out that they were wrong.
• Sometimes an expert achieves so much reputation that colleagues don’t even think he/she could be wrong.
• Even worse, experts tend to elect other experts just because they hold the same opinions as themselves.
• Experts could have obscure reasons to be wrong on purpose (possibly, of course).
• Experts may be corrupt.
• (if you can imagine more reasons please comment on it)

So trust experts, but also trust yourself.

Technicalities

The problem I had is that my project had tasks which didn’t have any logical connection between them and which were assigned to the same work resource (a person).

The software schedules these tasks concurrently, which is not wrong for me, but it calculates time to complete both tasks by looking independently which one will end up later. This would mean that the worker would have to work twice as fast, or 16 hours a day (instead of eight).

In my opinion, the program could either schedule them sequentially by calculating which tasks means a harder constraint to the project as a hole (because of possibly having dependent tasks) or just do something less intelligent, like assigning half the resource to each task and double the duration for both. There are more strategies that the program could adopt but these two come up immediately.

Proposed solutions were:

1. Addition of a dependency between the two tasks. I didn’t like this much because this dependency only arises because of the resource problem, but it’s not a real dependency. If I decided to assign one of those tasks to another resource I would need to manually remove that dependency.
2. Assigning half the resource to each task manually. The problem with this is that it’s difficult to implement in a case as simple as having to tasks, A and B, which need an amount of work of 5 and 8 hours. Since I would assign half the resource to each, that would mean working 10 days on both concurrently, and working 6 more days (only half-time) to task B.

I’m possibly overlooking something but this feature looked as something obvious to all people I’ve talked to about it. Please don’t hesitate to comment on this.

This entry was posted on Wednesday, March 9th, 2005 at 03:33 and is filed under Programari lliure, Cultura, Filosofia barata, Ciència, [en]. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.