Is someone aware of any estimation technique for the time required for writing LR scripts
What i am thinking is something like if we have "n" as number of screens in a test cases (assuming there is one to one mapping between Test cases and the script ) then the time required is something like "n(n+1)" in minutes.
This time includes debugging as well as unit testing for the script in both Vuser as well as controller scenario for few vusers. I tmay also incluse the time required to make the script dynamic.
I am not sure if this formula works but just popped in my mind....
I do not know if you can quantify this. I am always working on my scripts. And not just minor tweaks. With LoadRunner 7.5 they have new functions that did not exist before so I have added all sorts of new functionality to the scripts. I have also changed home made functions to new LoadRunner functions.
Now if you are for thinking initial recording of the script then it would take you about as long as a manual tester going through the product. But after that, it depends on how good of a programmer you are and how well you know LoadRunner and its built in functions.
The current script set I am working on, I took over from another programmer back in October and I am still working on it.
Don't for get to include the complexity of the achitecture. A three tier system (e.g. Web front-end, Unix mid-tier and OS/390 backend) would be harder to execute/debug than say a one tier system (e.g. web & SQL box).