First Blog
Recently, I attended a testing conference named CAST 2007. I've been blogging since 2001 in my private blog, but I've never really published most of my testing thoughts for the general public, for a number of reasons. One being that non-disclosure law makes me feel kind of nervous, and I don't like criticism without a point. I've felt like I have the best job in all of testing for years, so protecting my company was more important to me than sharing with any community at large. This seems really selfish when I see it in writting. I'm not like many of the people active in the testing community. I have no problem with authority. In fact, I've even written "Best Practices" documents for creating automation that is based on written test cases. I spend months plotting out the most complete new feature test case documentation only to have it become out of date almost instantly. I have experiences that every tester has, and a few unique to me.
I was inspired enough by one of the sessions I attended to finally come out of the blogging closet, and take a chance that someone might benefit from my testing rantings, and that someone might even be me.
I want to gossip for a moment about James Bach. I've been testing for seven years now. I've seen at least 25 different speakers about testing. He is my personal favorite, and the weird thing is I don't know exactly why. It is disturbing to me that I find what he says both so charming and correct. Is it because what he says makes me feel good? Or is it because I believe he is right? Is it because he likes people and I am "people"?
I know why it isn't. It isn't because he's "My Type" and the redheaded tester girl has a romantic crush on him. First off, he has a beard. No offense to James, his lovely wife and family, or any other of my furry faced friends, but beards are a huge NO for dating in Lanetteland. Luckily, I have a wonderful boyfriend (clean shaven, of course), so the fact that many computer loving folk have beards won't be a problem. Secondly, he likes video games, and it sounds like he does quite a bit. Thirdly, he has no real problem with confrontation and discussing the meaning of the word "is" or any other philisophical discussion for as long as you may like. I bet you could even ask him, "What if the word D-O-G really spelled cat?" and he'd think about it and discuss it with you for a long time. I have the patience of a three year old waiting for a cookie.
The reason may be because he believes in the value of human minds in software. He is one of the people in the software industry who does not seem to believe that I could easily be replaced by robots were they programmed a bit more correctly.
I feel like every day it is my job to answer the following for any number of people: What can you give me that automation can't?
I care. I want to be testing. I love finding bugs. It makes my day if I find a way to corrupt or crash something. I would love to compete against your automation when it comes to finding bugs that matter. I know I can do it faster, cheaper, and better. I've loved the products I work on with an insane amount of passion long before I ever worked on it. I have a stake in the reputation of the software. I USE it. I know it. I feel what it is like to use it. Automation will never do more based on intuition and curiosity, but I will. I've always felt respected and appreciated in my career because I work hard, I never stop learning, and I have some talent. I also have unique ideas (some better than others in practice, admittedly). Why, all of the sudden, do I feel that years of history and reputation mean little? If I am not a person who likes to write code all day, I am of no use.
Let me explain. I have set up, maintainted, and ran automated systems twice. I'd rather stab my own eyes out slowly than go through false failures and create and maintain bad automation all day every day again. It isn't that I can't do it. It's that I won't and shouldn't do it. It is a waste of my time and talent. I'm not great at it and I don't love it.
So why is bad automation still considered better than the best human testing in the eyes of so many?
I guess you can say I am out of the blogging closet because I can no longer fight this battle, that feels like a losing battle, alone.
I was inspired enough by one of the sessions I attended to finally come out of the blogging closet, and take a chance that someone might benefit from my testing rantings, and that someone might even be me.
I want to gossip for a moment about James Bach. I've been testing for seven years now. I've seen at least 25 different speakers about testing. He is my personal favorite, and the weird thing is I don't know exactly why. It is disturbing to me that I find what he says both so charming and correct. Is it because what he says makes me feel good? Or is it because I believe he is right? Is it because he likes people and I am "people"?
I know why it isn't. It isn't because he's "My Type" and the redheaded tester girl has a romantic crush on him. First off, he has a beard. No offense to James, his lovely wife and family, or any other of my furry faced friends, but beards are a huge NO for dating in Lanetteland. Luckily, I have a wonderful boyfriend (clean shaven, of course), so the fact that many computer loving folk have beards won't be a problem. Secondly, he likes video games, and it sounds like he does quite a bit. Thirdly, he has no real problem with confrontation and discussing the meaning of the word "is" or any other philisophical discussion for as long as you may like. I bet you could even ask him, "What if the word D-O-G really spelled cat?" and he'd think about it and discuss it with you for a long time. I have the patience of a three year old waiting for a cookie.
The reason may be because he believes in the value of human minds in software. He is one of the people in the software industry who does not seem to believe that I could easily be replaced by robots were they programmed a bit more correctly.
I feel like every day it is my job to answer the following for any number of people: What can you give me that automation can't?
I care. I want to be testing. I love finding bugs. It makes my day if I find a way to corrupt or crash something. I would love to compete against your automation when it comes to finding bugs that matter. I know I can do it faster, cheaper, and better. I've loved the products I work on with an insane amount of passion long before I ever worked on it. I have a stake in the reputation of the software. I USE it. I know it. I feel what it is like to use it. Automation will never do more based on intuition and curiosity, but I will. I've always felt respected and appreciated in my career because I work hard, I never stop learning, and I have some talent. I also have unique ideas (some better than others in practice, admittedly). Why, all of the sudden, do I feel that years of history and reputation mean little? If I am not a person who likes to write code all day, I am of no use.
Let me explain. I have set up, maintainted, and ran automated systems twice. I'd rather stab my own eyes out slowly than go through false failures and create and maintain bad automation all day every day again. It isn't that I can't do it. It's that I won't and shouldn't do it. It is a waste of my time and talent. I'm not great at it and I don't love it.
So why is bad automation still considered better than the best human testing in the eyes of so many?
I guess you can say I am out of the blogging closet because I can no longer fight this battle, that feels like a losing battle, alone.
Bravo Lanette! Well said.
Reply to this
I've left my testing days well behind me, but I quite agree that testers can never be meaningfully replaced with automation. People wishing to do so are simply seeking to save money and don't understand.
Automation's place in the testing world is in the exact same place as for the rest of the technology world... we should automate those things that CAN be easily and affordably automated.
Nobody really wants to run the hundreds of hours of test-scripts to check that every single error or warning message appears as expected... if we could use some robot to check those, GREAT! When I was a tester for a telecom company, we originally had to press every single keystroke combination to verify that the voice prompts that came out were the ones we expected to hear, and we didn't get the "message deleted" voice prompt when we pressed the "forward message" button. After a year, we developed automation to handle that set of tests.
No robot is going to realistically catch those esoteric bugs that fall between the cracks of ambiguous documentation and ambiguous specification.
So we need to use automation to let the humans focus their attention in ways that will challenge them and be of highest benefit.
Automation is a tool for humans, certainly at this early stage before simulated intelligence. It is not a replacement for a skilled professional.
It may, of course, reduce the headcount of departments who use a lot of 'unskilled' testing labour. (tire-kickers).
Interesting post though. I hope you have a lot more to say about automation.
Reply to this
I use automation to find all the bugs it can. I use humans to find the rest of them. (:
Automation is good at regression testing, so that known bugs don't reappear, and unit testing, so that average and boundary conditions can be checked every compile. Those are long, repetitive and boring tasks that computers are better than humans at, and that no one wants to do by hand (or really can afford to spend the time to do).
Humans are good at asking, "What if...?" Many of the unit or regression tests wouldn't exist if a human hadn't sat down and asked, "What happens when I do this?" Once they try it, and something breaks, they add it to the list and move on to new things.
Maintenance of the automated tests is firmly in the hands of the developers. Those tests are there to prevent the developer from introducing new bugs, and should be set up to automatically run after every successful compile. If tests fail, then the developer must look at them, and determine if they either broke something (in which case they have to fix their code), or changed the success criteria (in which case they need to adjust the test case, possibly working with a tester if it isn't a straight-forward fix).
In any case, automation is simply part of the infrastructure, put there by testers to trap known bugs without having the tester constantly running around trying to corral all the bugs that are already in the corral, as it were. Instead, the tester's job is to hunt the "wild" bugs, and shove them into the corral.
At least, that's my opinion. (:
Reply to this
U crazy freak..unsubscribe me from ur highly useless blogs.
Reply to this
I do not subscribe or unsubscribe anyone. If you subscribed yourself figure out how to use your own software and remove yourself. You are the only person responsible for your learning.
Reply to this