MY TEAM

Gkam84 wrote:

Its a pretty basic model behind the scenes, there are only a couple of things that I can do wrong...

I'm a little ignorant about these things... Well, granted, about a lot of things... I know there's a lot of things I don't understand... And this is one of them... I know jack about database engines, coding, queries, 'undo' mechanisms, data structures, SQL... I just know Excel, and just a little bit... I like its colors... But, this is why I'm confused...

.
This what I think happens:

1. You add a rider to the database
2. You ask the database to assign a value to the rider based on an algorithm that takes into account previous scores in previous stages and sets that value in relation to the highest scoring rider for the competition (valued at 40.0) and a rider who hasn't scored at all previously (3.0)...

So I don't understand why it would be that the rider valuation process could not be fixed immediately... This is why... I imagine that by just running the valuation process immediately, right after the 'mistake' run, ignoring the previous additions, you could restore the previous rider evaluation results...

Granted, I don't know what happens in the database when you run the rider valuation process... Are the rider valuations 'written over' irrevocably?... I have to imagine that if you can write over them once, you can write over them once again.... I also imagine that rider scores are not discarded, so I have to imagine that the process can be run again...

I just think Dave is a guru with a great sense of database design and, at least in my mind, he can probably fix anything (relatively?) easily... I think he knows that much...

stumps wrote:

I cant see the problem in this.

chiv, you obviously know what your talking about but its only a small part of the bigger game and i'd like to think that this sort of testing will create a better game for everyone.

Stumps as I've said I don't have an issue with the testing in fact I agree with it , but having worked on development and seen the carnage that one incorrect line of code can create it's the whole testing on a live system that is the disconcerting part to me especially after confirmation that there isn't a ringfenced system to work on prior to go live tests .

Anyway my concern has been voiced ,Gkam is gonna revert at a quiet time and everything is still working as it should so life is good

Gkam84 wrote:

We can only test things in a live environment, because that is where people are complaining that things are not working.

We could set up another server and test just bits and bobs, but if that didn't fit in with the real game, what would be the point. Things will change through the night and my team will go back to what it was before.

Look I'm not having a pop , as I said I appreciate testing is required but having worked in IT it's not the norm to test in a live system until you need to . There is usually an offline or ringfenced system used that exactly replicates the live system that is used for development/testing hence I said it's disconcerting .