No, there is no way to specify specific dates without establishing
constraints. Nor should you want to. There are spcific exceptions, of
course, where you do need to establish constraints and that is why Project
allows you to enter data manually in the start and finish fields. But in
general the ONLY date you should ever input into Project is the Project
Start Date and perhaps any major deadlines - everything else should be
calcualted for you by Project. In fact, a primary function of Project is to
do exactly that and *not* just to passively display a schedule you have
already determine elsewhere.
You say you have specific start and finish dates for your tasks but do you
really? You may have dates that someone has determined they THINK the tasks
can occur on, they may want them to occur on those dates really, really
badly, and you most probably have dates that are deadlines by which tasks
must finish in order to successfully meet the business requirements that
gave rise to the project but how do you know (yet) if those dates are
actually possible? The boss may have even said "fidgeting the widgets MUST
happen during the third week of September or you're fired!" but even that
mandate does not mean it is actually going to be possible to do it. After
all, the widgets might not be ready until October. MS Project will take
what you do know as fact - the date the project can get underway, the
network relationships between the tasks that drive their sequencing, the
time it will take to do the work of each individual task, and the
availability of the resources who will actually do the work and then it
calculates the schedule that is possible under those conditions. It gives
you a "reality check" on the dates you've been given. If it matches what
you require, great! Alas it probably won't on the first pass. But now,
because you let Project do its thing, you have some very valuable
information. First you know that those schedule dates you were given simply
won't work and you'll be looking for a new job if you try to work it as
handed to you. Secondly, you have a model where you can experiment - "if I
can get an extra engineer assigned we can do both of these tasks at the same
time instead of having to do them in sequence because the one guy we've
already got can't work on both of them at once and if I can get some OT
approved for the widget assembly tasks we can get that done a week earlier
and that in turn means we CAN fidget the widgets when the boss has mandated
because they'll be ready for the fidgeting process three weeks earlier than
we would have had before!"
So to answer your last question - you made a good start by identifying and
entering the tasks. But then you blew it by going back to input dates- that
you should never, ever, do. Instead, you should have looked at the process
itself and set up the links between the tasks that model the flow of product
or information through the project, estimated the length of time each task
will take given the resources you have to do them, and assigned the
resources. You then examine to calculated dates for the individual tasks to
see if they are close to what you needed. If not, you'll have to adjust
things that out in the real world you actually can control and that can
effect the sequencing and/or durations of the individuals tasks.
