Originally posted by cgarvieuk
but like i say they have the shit job of discovering that if yiu change X then blow me, why the hell did Y break
Which is exactly why these things should be tested by actual users before going live. People who don't use the site are never going to be in great position to spot the unintended consequences or even have a nose for where they might occur.
|
Originally posted by chriso Originally posted by cgarvieuk but like i say they have the shit job of discovering that if yiu change X then blow me, why the hell did Y break Which is exactly why these things should be tested by actual users before going live. People who don't use the site are never going to be in great position to spot the unintended consequences or even have a nose for where they might occur. Getting them tested by people whose job it is to sit there and rigorously test them still fails to find many errors and triples or quadruples the time to fix a bug. At least now when things go wrong, fixes come within hours or days and not 10 days, weeks or months. We are far better off now. Getting real users to rigorously test anything is like herding cats. We have tried such thing before. Implementing this idea would likely make fixes that take hours, take weeks or months instead. I know people are unhappy, but the cold hard facts are that we are much better off now than we were a month ago. And the process difference and the 100 code changes related to a myriad of improvements in the last month are related to the bold elimination of QA. The most recent performance issue was unrelated to a code change of mine. It resulted from a code push with several pieces of code that got into this release. Hang in there! Things are far better now than they were last year, last month or a decade ago. They will get better!
|
I can still rate beers. I'm good, I'm hanging in
|
|
Originally posted by DuffMan
I can still rate beers. I'm good, I'm hanging in
+1
|
I'm still a bit upset that we no longer have jmm635 as default *4AD (aka RateBeer Spotlight), though...
|
Originally posted by joet
Originally posted by chriso Originally posted by cgarvieuk but like i say they have the shit job of discovering that if yiu change X then blow me, why the hell did Y break Which is exactly why these things should be tested by actual users before going live. People who don't use the site are never going to be in great position to spot the unintended consequences or even have a nose for where they might occur. Getting them tested by people whose job it is to sit there and rigorously test them still fails to find many errors and triples or quadruples the time to fix a bug. At least now when things go wrong, fixes come within hours or days and not 10 days, weeks or months. We are far better off now. Getting real users to rigorously test anything is like herding cats. We have tried such thing before. Implementing this idea would likely make fixes that take hours, take weeks or months instead.
Fair points, but how close are we to having a platform where changes can be implemented without breaking all sorts of things all round the site in a an unpredictable manner? Will we ever get to that stage? Right now it often seems that changes break things that are far more damaging to site usability than whatever they were intended to fix or implement in the first place.
|
Originally posted by Benzai
Originally posted by DuffMan
I can still rate beers. I'm good, I'm hanging in
+1
|
|
Originally posted by chriso
Originally posted by joet
Originally posted by chriso Originally posted by cgarvieuk but like i say they have the shit job of discovering that if yiu change X then blow me, why the hell did Y break Which is exactly why these things should be tested by actual users before going live. People who don't use the site are never going to be in great position to spot the unintended consequences or even have a nose for where they might occur. Getting them tested by people whose job it is to sit there and rigorously test them still fails to find many errors and triples or quadruples the time to fix a bug. At least now when things go wrong, fixes come within hours or days and not 10 days, weeks or months. We are far better off now. Getting real users to rigorously test anything is like herding cats. We have tried such thing before. Implementing this idea would likely make fixes that take hours, take weeks or months instead.
Fair points, but how close are we to having a platform where changes can be implemented without breaking all sorts of things all round the site in a an unpredictable manner? Will we ever get to that stage? Right now it often seems that changes break things that are far more damaging to site usability than whatever they were intended to fix or implement in the first place.
Also fair points. We have been more or less stable in our releases over the years with instability positively correlated with the size and scope of changes, familiarity with code and a need for speed. We are tuning a new process now and have made improvements. The latest process change is that I was not signing off on all new web code, and now I am.
|
Originally posted by ebone1988
Originally posted by Benzai
Originally posted by DuffMan
I can still rate beers. I'm good, I'm hanging in
+1
+1
even though I have lost 27 ratings overnight.
The count yesterday was 50778 and today after having a beer, counter is on 50752
|
Originally posted by fonefan
Originally posted by ebone1988
Originally posted by Benzai
Originally posted by DuffMan
I can still rate beers. I'm good, I'm hanging in
+1
+1
even though I have lost 27 ratings overnight.
The count yesterday was 50778 and today after having a beer, counter is on 50752
Same here, I am constantly having to rerate beers I have already rated weeks/months ago. I want to continue to rate but this is getting very frustrating.
|