testing, review, etc 1

[VE]

My main job function is writing Crystal Reports. The company I work for has really old, disconnected systems so I actually have to do a lot of work in MS Access (though I'm now learning SQL which I like a lot better) to force the data into a format I can work with.

The data originates in Excel, Access, Paradox and MS SQL. I’m in a situation where I’m actually rather pleased that I CAN do this, but very irritated that I have to.

I used to work for a manager who was very hands on and understood what I was working with. He also would test my reports and help me find problems that I may have overlooked.

That manager left about a year ago, and now I have no "tester". I am also being asked to develop and publish new complicated reports on a much faster timeline, generally with several related reports being created/published simultaneously. There is no other person in the company with even a similar skill set for me to work with.

Of course this means that now users are finding the problems. My credibility is getting lower and lower. I kind of feel like I'm being set up.

How common is this situation? Am I just being a big baby about this or was I just spoiled before?

Any advice as to what I should to make this situation better?

Thank you
VE

 

[RiverGuy]

If a user requests a report, he or she should be more than happy to test it for you. If you are able to get this into more of a process, then it might make you look better. If you present something as production-ready and its found to contain flaws, that looks a lot worse than if you release it to UAT and request input from the user.

 

[Glenn9999]

Heartbreakingly common, I'm afraid. Usually, having another set of eyes to look, test, and think of test scenarios that you might not have thought of is a luxury, though it's generally a smart thing.

Dilbert is not a fictional cartoon. It is a documentary.

 

[VE]

Thank you.

I think what Riverguy said is the missing link. I actually had an unwritten process when I had the other manager, when the new one took over it got lost (most likely because it was very informal and not written down).

For my sanity and integrity I need to write down a set of steps including a list of questions for requester when they ask for the report.

I think a semi formal report definition as the first step, a thorough test by the requester (as many rounds as needed) then some sort of formal OK by the requester when it is complete.

That should also cut way down on the insanely complicated requests I get, if they are responsible for OKing it before publication then they are not going to be as likely to insist on things that I tell them upfront are difficult.

Glenn9999, I suspected that might be the case. Blah. Oh well.

Thank you
VE

 

[AlexCuse]

Too true, Glenn.

or is it

[small]----signature below----[/small]
Majority rule don't work in mental institutions

My Crummy Web Page

 

[philhege]

When you create a report per a user's request, there should be an explicit set of functionality that defines the report. Get the user to sign off on the functional specifications. Use this document as a reference to create a test script. Schedule time for the user to come in and follow the script, verifying that the report works as they specified.

Two "must-haves": the script, and the user's time to test the report, ESPECIALLY in your current environment. You should never, ever release an application to production without testing, evaluating, and correcting problems found during the test(s). Your boss should understand this and make this process an iron-clad policy at the company.

If any one of these actions is missing or you get pushback, find another job, or your situation will force you to find another job eventually.

Good luck!

Phil H.
Some Bank
-----------
I used to be happy programming. Now I'm miserable project planning. Hey, at least the money's good.