3 ms·
The software I work on has about 50,000 daily users. Every time a change is made, there are extensive automated tests that test all of the software features. Th
by bonestamp2 8y ago
The software I work on has about 50,000 daily users. Every time a change is made, there are extensive automated tests that test all of the software features. Then the nightly build is automatically connected to every compatible piece of hardware made in the past 20 years and tests that change against all of that hardware (this step takes about 10 hours to complete). Then, and only if it passes all of those automated tests, there is a team of people who continually test any new changes, and try to recreate any problems that have been reported by beta users and field users. Before a release, they spend a month testing it from end to end.
If a critical bug ever makes it to the field, which almost never happens anymore, not only do we revert to an old version immediately (which I suppose Apple could make available if they wanted), we spend as much time as it takes to understand how it got past our tests in the first place. Then, not only do we write a specific test to catch it next time but we also look at how we can change our process to fill the hole that allowed it to get past in the first place -- so other problems can't get through that same gap later.
If that's our process, and we're a small outfit compared to Apple, I can't imagine how rigorous Apple's process is.
- josteink 8y ago> If that's our process, and we're a small outfit compared to Apple, I can't imagine how rigorous Apple's process is. Why would you blindly assume they have a more rigorous process just because they are bigger?
- natecavanaugh 8y agoMy guess would be because they have more at stake with each release?
- manicdee 8y ago1. larger customer base means more angry customers to deal with 2. they have more money to throw at the problem 3. any problem that makes it into the wild gets blown out of proportion, so pre-release testing is part of their PR budget