Posted by testcrunch on 21st November 2007
Some bright spark has come up with the idea of doing future regression testing offshore in Asia.
Now this is a project which prides itself on doing agile and scrum and all that stuff, and it’s painfull enough for me to make headway in this kind of environment, but somehow we are (You call that progress? Ed).
I was introduced to a representative from the company, who were supposed to be doing regression testing in Asia, and who wanted to sit with us and pick up on how we do our work. Of course she hadn’t much idea of that, and she soon enough mentioned the lack of documentation, which she couldn’t believe. I confirmed that was the case, but even then she asked me that I must have some documentation that I could let her have. All we have is some acceptance criteria from the end users, and you most definitely can’t regression test from that at my desk, never mind 6,000 miles away. She was a quick leaner but this is just not going to work.
When I was testing some real down home functionality, I was having terrible trouble proving it worked. It just seemed to work back to front, if not front to back and even one side to the other. Talking to a few others on the team didn’t help much as their ideas on how it was to work were a little vague, if not completely contradictory.
The developer, who sits opposite me, decided this afternoon to have a look at a few issues I had mentioned to him. Now for this bit of functionality, for our benefit, we do have a very roughly worded spreadsheet of a few rules for the processing. The developer suddenly asks me can he have a copy of this spreadsheet. Eh! Where’s your copy then? Apparently he’d never seen this document before in his life. Well how about that! Lord knows how he wrote the code. Probably best not to even think about that.
He read a couple of the processing rules on the spreadsheet, then played with his own software for a minute to see if it processed those rules, before loudly exclaiming ‘this software is a piece of shit’. Made me laugh.
I think we’re all going to have to pull together on this one.
Posted in Agile - nothing written down, IT in the East - copy not create | No Comments »
Posted by testcrunch on 14th November 2007
There are now about 18 people at our Scrum meetings and that’s just too many. I thought the idea was to have about 7 or 8 people.
It’s so obvious that we are in pain due to the complete lack of documentation on our very Agile project. There are only two people who know how the system works and they both think in different flavors. Needless to say confusion abounds, but unbeliveably we are making progress. Though it’s not helped that the developers are running the show and they do exactly what they like. They have their development environment, the users have a UAT environment and it seems everybody, including that guy walking by the office right now, has access to our system test environment. What this means is that anybody can do anything to it, and do. We have background workflow processes being switched on and off like billy-oh, everybody uses everyone elses test data and mash that up with the assorted (mis)understandings on how the system is supposed to work and you have chaos.
Oddly enough a lot of them have never known any other way of working. They don’t mind some people attempting to put some order in this anarchy, but there is only so far they will go. I suggested dividing up the test data so we would all have our own batches of data, which would have been very easy to implement but they definitely didn’t like the idea of that. That was far too organised.
Apart from the test environment being an everyman for himself kinda place, the code is flakey beyond belief. What works one minute may not the next. The developers say let us know and we’ll have a look at it. I’ve tried that a couple of times and I get the predictable mumbo-jumbo about all the great reasons why that bit of functionality doesn’t work, right now, but will real soon. This has been going on for weeks.
I think I might start raising loads of defects on all of these ‘features’. No doubt they’ll be a quick fix to get the bugs closed and keep the statistics looking OK. What about if they clear the bug, I prove it works OK and let them know that it will be staying open until I haven’t seen the bug again for a month, just in case a rogue build reintroduces it. That should concentrate their minds a bit.
Am downloading the latest large patch from WoW, which is gonna take another couple of hours, hence this post. Have got to level 33.
Song title:
‘If the phone doesn’t ring, it’s me’ Jimmy Buffet
Posted in Testing software - watching bits drop off, Agile - nothing written down, WoW - not Vista | No Comments »