How do you deal with this?

_Postal_

Hi every one,

I am currently in a company as a Software Test engineer.

The project manager intends on putting live a web application that has not been tested.

Due to extensive reworks over the last month nothing has been completed until to-day. The dead line for deployment is this evening.

There is no possible way to test this application with in the available time. Not even if we had the reset of the week.

Parts of the project seem incomplete from what I have managed to get my hands on so far to-day.

Can I get some feed bak please. I find this rediculous. What would you do / expect in this situation?

Regards,
_Postal_

John_Mc

Refuse to sign off on it and strongly recommend that it does not go live until fully tested and verified. If the manager chooses to proceed regardless, then you cannot be held accountable.

Take accurate notes from any discussions between you regarding this just to be sure.

_Postal_

Thank you for your fast reply.

I have done this all ready to no avail.

I needed to make sure there was no other way about this... Im new the position.

seamus

Yeah it's completely ridiculous, but since you're not the project manager, it may not be your place to alert anyone "higher up" that there's a problem.

First and foremost cover your own ass. Send an email to the project manager stating that you cannot recommend going ahead this evening. If you are required to provide any QA sign-off, don't provide it.

What you need to do now is prevent anyone from turning around tomorrow and blaming you for everything turning to ****.

If this is something which could put your employer in jeopardy, then it might be prudent to talk to the PM's boss telling him that think it would be better to miss then deadline than go live.

John_Mc

_Postal_ wrote: »

Thank you for your fast reply.

I have done this all ready to no avail.

I needed to make sure there was no other way about this... Im new the position.

There's not much you can do then unfortunately. Your project manager clearly likes living on the edge!

_Postal_

nuts... & Thank you all for your replies.

Evil Phil

seamus wrote: »

First and foremost cover your own ass. Send an email to the project manager stating that you cannot recommend going ahead this evening. If you are required to provide any QA sign-off, don't provide it.
What you need to do now is prevent anyone from turning around tomorrow and blaming you for everything turning to ****.

Get a read receipt too, if you can. Print both your email and your read receipt out (if you have one) and keep copies in work and at home. That way if you end up in a meeting where all eyes turn to you as to why untested code was released into production you can simply produce these copies and say, I pointed this out, this is why I refused to give QA sign-off on this.

You could possibly cc somebody higher up than your project manager too, but only you can make the call on this.

As seamus said, cover your own ass. That way when the **** inevitably hits the fan it won't be your job on the line.

p

Sometimes it's better to get a not-perfect piece of software out, and fix it later, rather than perfect it the first time. Sometimes this is for business reasons, sometimes promotional reasons and sometimes silly reasons, but either way these are not your concern. You can test as best as possible in the time given, and as others said, note your concern so that it doesn't come back to bite your ass.

This stuff will happen, best thing is to do you job as well as possible in the time given, and not let it get to you.

John_Mc

p wrote: »

Sometimes it's better to get a not-perfect piece of software out, and fix it later, rather than perfect it the first time. Sometimes this is for business reasons, sometimes promotional reasons and sometimes silly reasons, but either way these are not your concern. You can test as best as possible in the time given, and as others said, note your concern so that it doesn't come back to bite your ass.

This stuff will happen, best thing is to do you job as well as possible in the time given, and not let it get to you.

Yes, I agree - it's about balancing the risk of software defects against missing deadlines and commercial pressures. In this case though, the OP has said the software hasn't been tested at all so it's impossible to weight the risk of releasing it...

Webmonkey

p wrote: »

Sometimes it's better to get a not-perfect piece of software out, and fix it later, rather than perfect it the first time. Sometimes this is for business reasons, sometimes promotional reasons and sometimes silly reasons, but either way these are not your concern. You can test as best as possible in the time given, and as others said, note your concern so that it doesn't come back to bite your ass.

This stuff will happen, best thing is to do you job as well as possible in the time given, and not let it get to you.

In a way this is what Microsoft does. Obviously though this is a not "Just Enough Testing". OP, cover yourself as the previous posters have said.

Evil Phil

Webmonkey wrote: »

In a way this is what Microsoft does. Obviously though this is a not "Just Enough Testing". OP, cover yourself as the previous posters have said.

They don't release anything onto Microsoft.com without it being properly tested first.

Yes, imperfect software is released all the time and that can be due to business pressures but at least it's had some testing. Releasing untested code onto a production server is crazy.

Webmonkey

Evil Phil wrote: »

They don't release anything onto Microsoft.com without it being properly tested first.

Yes, imperfect software is released all the time and that can be due to business pressures but at least it's had some testing. Releasing untested code onto a production server is crazy.

This is what I'm referring to http://www.satisfice.com/articles/gooden2.pdf

Obviously they have excellent testing but it's impossible to catch all bugs so full testing is never really an option only for hard real time systems for example.

Obviously the OP's case is a bit different, more testing should obviously be done there.

fergalfrog

Let us know how you get on. Did it go live yesterday evening?

If it's not fully tested perhaps try and log all errors etc so that at least you are using a live audience for testing.