The cost of rewrite : Business Object

A business object requiring a change is a headache in project.

Lets say u need to add a new status field to an existing bean.

Run sql scripts and add these to migration scripts.

Then you have a completely new object. You must change all constructors.

Check member methods to change related member methods.

Check service methods.

Add tests to unit tests.

...

List is long. Often managers underestimate these kind of things.
"I only wanted one new field"
Maybe change seems simple but it breaks all former code.

Junior Idea : "Why do not we write a new program"

All juniors(including me) comes with the idea of rewriting the program in 1st month of employment.
And when they are rejected they are surprised.
They think they offer a better version but they were rejected :)

What one understands over time :

1)A full complete project running in an old language/style is much better
then an unknown version with new language/style.

2)Juniors are too fast to propose a rewrite. It is obvious that they did not see
whole system and is not aware of former problems.

3)A junior is a new worker. It is not certain how long he will work at
that place. So no company can depend on him.

Why projects fail?

In Programming world , you always see unsatisfied customers and developers.
After years of development I can understand a project will fail or not from beginning
if there are some signs like below.

1) Low price
Projects specs are announced and price is asked to companies.And then companies bid for lowest price.
What is lowest price? Can you afford to make a project will 2 juniors instead of 1 senior?
1 2 3 or whatever Junior is not one senior.
There is no replacement in programming.

2) Who knows the industry?
Programmers often underrate project domain. They think most important thing in project
is technical part. I never see a big project where domain knowledge is not important.
Programmers must respect domain knowledge and understand it as much as they can.

3)Unclear breakpoints.
At 1st stages people code fast and come to a stage.
At that stage ,customer says this is not what we want.

Will you continue with problematic logic?
Can you afford a re-write?

Program plan must have some milestones and specs to reply
these questions.



My simple rule for a project :

A project will fail or will be in a low quality :

If there is no one int the team who has done a technical similar project