Start-to-Finish Example

[Sinister]

Could someone give me a few examples of the use of a Start-to-Finish
dependency?

Also, how should a successor behave if you increase the successor's
duration? I'm curious to see before and after shots.

 

[John]

Could someone give me a few examples of the use of a Start-to-Finish
dependency?

Also, how should a successor behave if you increase the successor's
duration? I'm curious to see before and after shots.

Sinister,
I never cared much for the start-to-finish link type because it attempts
to predict the future. Nonetheless, one example might be for a test plan
that must be submitted for review x number of days before the test
starts. For example, let's say something is being designed and built for
an eventual qualification test. The test plan will take 30 days to
prepare and must be submitted 90 days before the test is scheduled to
begin. If the test itself is task ID 10 then the test plan could have a
predecessor of: 10SF-90d. Project will then schedule the start date of
the test plan so it can be completed and submitted 90 days before the
test.

With regard to a successor. The Duration of a successor task has nothing
to do with its link to the predecessor. The predecessor task defines the
start date of the successor and the duration defines the successor
finish date.

Hope this helps.
John
Project MVP

 

[davegb]

Sinister,
I never cared much for the start-to-finish link type because it attempts
to predict the future.

We must see scheduling very differently, John. I believe that every
schedule I've ever done, with or without SF links, was an attempt to
predict the future. Never have gotten very good at it! Every one I ever
did changed at least a little.

Nonetheless, one example might be for a test plan

 

[JackD]

A few seconds with MS project and you should be able to see how it works
before and after.
I'd only use SF for tasks which have some fixed time associated with the
task - for example, you need to submit paperwork 3 weeks prior or something
like that.

-Jack Dahlgren

 

[John]

We must see scheduling very differently, John. I believe that every
schedule I've ever done, with or without SF links, was an attempt to
predict the future. Never have gotten very good at it! Every one I ever
did changed at least a little.

Dave,
No I don't think we see it differently. Every plan someone develops is a
prediction. It's just that the SF relationship puts more "faith" in the
prediction. For instance, in my example Project is telling us when to
start the test plan based on something that is going (we hope) to happen
almost 3 months in the future. Every other link type in Project has what
I consider a little more reasonable approach to the predicting. That is,
present tasks are not generally tied to something in the future. I would
rather use links that are a little more predictable. For example, have
the test plan start so many days or so much percent after the design
phase starts. With this approach the test plan may be ready more than 90
days before the test but in my mind starting the test plan based on
where we are today is better then basing it on where we hope to be at
some future date. It's a fine line distinction but it just feels better
and more intuitive.

John

 

[davegb]

Could someone give me a few examples of the use of a Start-to-Finish
dependency?

Also, how should a successor behave if you increase the successor's
duration? I'm curious to see before and after shots.

So far, all the examples of SF relationships have been with long lag
times. It might be a little clearer with an example of a direct SF
relationship.

I take my car to the mechanic and tell him/her I have to have it back
at 5 pm. I need the brakes and the radio fixed. If the mechanic planned
his/her work, they would estimate the time it takes to do each task and
then back up to a start time. I.e., if they estimate it will take an
hour to do the brakes and 90 min to do the radio, they would schedule
the start at 2:30 pm. It's really just scheduling backward. It's high
risk because there's no margin for error unless someone adds some extra
time, like in this case, starts at 2 pm, just in case. It's kind of
like Just-In-Time scheduling done in manufacturing.

The best applications I can think of occur when there is a good reason
why something has to be done at the last minute. Like shipping the
lettuce. Definitely don't want it on site a week early!

Hope this helps in your world.

 

[davegb]

Dave,
No I don't think we see it differently. Every plan someone develops is a
prediction. It's just that the SF relationship puts more "faith" in the
prediction. For instance, in my example Project is telling us when to
start the test plan based on something that is going (we hope) to happen
almost 3 months in the future. Every other link type in Project has what
I consider a little more reasonable approach to the predicting. That is,
present tasks are not generally tied to something in the future.

Again, my experience is different. The vast majority of projects I've
done are dependent on a future date, the date the project is due. While
I may schedule from a start date for other practical reasons, my target
is the end date, and if my schedule shows the project finishing later
than that date, I revise my schedule, i.e., attack the CP, to try to
achieve that end date. I guess I don't see much difference between that
and saying I have to complete the test plan by a certain date because
someone needs it 90 days from now.

I would

rather use links that are a little more predictable. For example, have
the test plan start so many days or so much percent after the design
phase starts. With this approach the test plan may be ready more than 90
days before the test but in my mind starting the test plan based on
where we are today is better then basing it on where we hope to be at
some future date. It's a fine line distinction but it just feels better
and more intuitive.

Maybe it's just a much finer line in my mind than in yours! As for
the intuitive part, that's so individual that it's hard to say. To my
mind, there's no difference between saying I have to start on date x to
finish on date y than it is to say I have to complete the test plan 90
days before the test. Just a difference in perception I think. We
probably end up doing pretty much the same things at the same times,
but with very different plans for getting there. This supports
Eisenhower's quote, "Plans are meaningless, planning is everything". It
doesn't matter which way we plan it, we'd both probably bring the
project in on time, which is the part that does matter.

 

[Sinister]

Thank you both. Very interesting points, and the example was great. Thanks.

 

[John]

Thank you both. Very interesting points, and the example was great. Thanks.

Sinister,
You're welcome.

John