Re: MrProject Features/Target Market

On Thu, 22 Aug 2002 16:53:28 +0000 [email protected] wrote:

> I thought it might be useful for me to add a few 
> comments re:Features and Target Market.
[...]
> Now, I have some pretty strong feelings about the 
> direction and feature-set that should be 
> incorporated.

Great! Strong feelings and good manners (as you obiously have ;o)) make good
advocates. Experience makes good specification writers ...
Well, ok, i'm not a MrProject developer and therefore not really entitled
to say the above, but I do feel (strongly) that it is the input coming from
the trenches that can make the difference.
I am aware of the dangers, e.g. people from the trenches sometimes can get
'entrenched' in their view, but nevertheless :o)  
> 
> Not so many years ago, there existed numerous 
> software tools that were vastly more powerful 
> than even the most capable contemporary tools 
> today. Those tools ran on the mainframe and they 
> did an incredibly good job of baselining and 
> tracking the time, resources, and cost dimensions 
> of large projects of great complexity - and of 
> multiple small projects. The names of those tools 
> were MSCS/COPES (Management Scheduling and 
> Control System by McDonnell Douglas), Project/2 
> by Project Software Development Inc., CIPREC by 
> IBM, PREMIS/PICOM by ???, PMS-4 by IBM was 
> probably the first of the truly powerful systems.

What are the most striking features of such systems?
Which do you still remember and why?
Do they still apply today?

[...]
> Planner (P3) at the high-end (industrial 
> strength, if you like) and MS Project at the low 
> end.

Yeah, rock bottom.

[...]
> On to a few specifics:
> 
> Re: Target Audience. I did not see any mention of 
> the possibility of pursuing the large 
> Engineering/Construction companies - nor of 
> government use for their major 
> acquisition/development projects/programs.
[...] 
> It seems the current view is more toward the 
> lower-end user that places less demand on the 
> system.

I think the reason here is to avoid overspecialization at the beginning and
start with something rather symple. The optimum would be a system that can
be primed to a particular industry with a few config changes by the user.

> Re: Single-User system. I noticed the comment 
> that MS Project is a single user system. Well - 
> yes and no - but that really is not the issue 
> here.
[...]
My support on this one. I guess it would be better to prepare the stubs for
a system which is multi-user and distributed. Changing later could be costly.
 
> Re: Task Duration(s). I noted that we are 
> planning to allow 3 types of durations for an 
> individual task - pessimistic, assumed, and 
> optimistic. This is a network model that reflects 
> PERT as opposed to CPM. I wonder why it is that 
> we chose this approach?
> 
> You see, PERT is, by its nature, MUCH more 
> complex - both from a system User perspective - 
> and from the perspective of those who will 
> receive the outputs. It is a probabilistic model, 
> and as such, it offers useful information on 
> projects in which there is VERY large variability 
> with durations - but at the cost of being MUCH 
> more complex. 
> 
> CPM (Critical Path Method), on the other hand, is 
> deterministic and makes use of only a single 
> estimated duration for each task. It requires 
> only a single azimuth throughout the network 
> rather than PERT's three azimuths.
> For every type of project I have ever been 
> associated with, CPM offered a much more 
> practical and useful output than did PERT. I 
> suggest you reconsider this direction.

Ok, but does it give me an error estimate of the delivery date?
Anyway, I don't think that one should exclude the other, both have
advantages and applications.

Ciao

-- 
Michele Ravani                  [email protected]
"Those who live hoping, die singing" My Gran