Hacker News Reader

Comment by codingdave | Hacker News Reader

Honestly, I think the answer is no. But the sheer amount of software in our world is increasing. So if software quality lands on a bell curve, you'll see the lower end of that curve more often. Which actually argues that software quality is increasing - when so much of it "just works", the ones that don't stand out.

View on HN

I lean to yes, because even things that formerly worked, fail without good reason. If feels like quality is not a top priority anymore, and I think that trend started with network patches, suddenly you could sell unfinished software while you still work at it.

For example, the most recent update to iBooks from Apple glitches if you rotate the screen, and stays that way even if you rotate it back. iBooks is an app that recently got a tiny facelift, yet no one tested what happens if you flip the phone.

I agree that we use more software now, so we should experience more buggy software, but if that was the case, we should have experienced the same proportion of buggy software all our life. I still think easy patching facilitated bad quality practices, although those bad practices probably make commercial sense.

Te most recent update fixed this issue. While the easy workaround was loking rotation inside iBooks preferences, I much prefer for it to follow the orientation.

That was fast, which is the bright side of network patches.

Yep. https://en.wikipedia.org/wiki/Sturgeon%27s_law

Indeed and the issue with the web is that 90%, instead of being buried in the pile of broken software, it is only a click away.

For software quality to have bell curve distribution, you have to define negative software quality first and then proceed to define inifinite qualities also. Otherwise it is either one-sided distribution (when there is no negative quality) or bounded distribution. Both cases have no guarantee to have sums of independent variables to converge to a bell curve. Log-normal distribution certainly does not converge to normal distribution however hard you try [1].

[1] https://stats.stackexchange.com/questions/238529/the-sum-of-...

Which brings up the question: what is negative software quality? Can we define that?

Reply on news.ycombinator.com