Hi,
Thats a tought question but I could explain how we
approach the stuff.
I am a project manager and devloper and actually we have a
huge project on which we decide to use .NET.
The first problem we get is that we come from VB5 users
and when we see the introduction of .Net we were thinking
of converting how old project to .NET. That was a really
bad idea so we decide to rebuild safelly the whole thing
again.
Frist of all we start to analyse the whole .NEt concept by
how self with some getting started example and so on, and
we see quite fast that it was a huge effort and as in most
industry we tend to reduce development time as much as
possible the we decide to intent a 5 days training to be
point to essential only. After that training we say WOwwww
so many good things inside that tool.
Now it depends a lot how far you would go in your .NET
approach. YOu can either think about simple things or more
evolutive and modular stuff.
Then we decide to go through a 2 weeks analysis on how we
can implement the feature that our customer needs after
this we get a clear idea on what we will use for the
implementation but how to do it is another story.
We start then to sit on a black board and write
specification of our application feature and start to
split it in small component that will defined our classes.
One thing was sure, as a new user consider by adding a by
2.5 factor the developement time compare to the tools you
get used.
What we decide to do then is to use Visio and UML for
havinh a clear view of our things will be organise and
thats not a piece of cake.
by my own experience, its better to spend more time than
expected on specification that will fit your need at the
end compare to redo it again and again because you did not
specify it a bit deeper.
Another thing important as well, especially as new user
in .net, do not hesitate after writing a few specification
to take one or two days to translate your basic specs in
few test line of codes just to see if you are in a good
direction, do not try to make it good but just to verify
what you want to do can be done in the way you though.
According to my few onth experiment do not hesitate to
devide your modules in smaller modules than you was doing
before, this by keepimg in mind that testing small entity
integfration works bettter that huge one.
You know I am deep in that actually, specifying and
specifying and specifying and still know there is some
points in .-NET that I do not catch and put a break on my
specs definition.
Your were talking about cost, it depends so much of your
project size... in our case doing the application we are
working on actually tooks 3 month in VB5 and actually you
can be sure it will take more than 6 for us in .NET even
if we have a clear idea how it should work.
Of course we could also do it in 3 month too by using
implementing it in a similar way we was doing in VB5 but
there is no sens doing that ...
here was the step we are actually hope it helps you a bit
to start.
WOUld love to share your approach as well so let me know
how you are doing
There si one good book as well for project analasis
in .net called something like .NET Architecture Analysis
from Microsoft Press
Regards
Serge
-----Original Message-----
Hi
Any Ideas and Inputs on Cost Estimation/Planning stage of .NET Projects,
i.e. things like Function point Analysis, COCOMO and on full planning stages
etc etc.
Any Frameworks, feasible methodologies.. Etc etc ???????
[I believe experience speaks a lot on this cost estimation, but I thought
I'll get some fresh ideas and thoughts]
I'm Totally Blank [!!] on these areas, so any help will be highly
appreciated.
As these stuff are occupying most of my time and hindering me from my
normal day to day activities
[forgive me if it's Off-Topic]
Regards,
Logu K
.
