Quality criteria for software

What are average numbers of bugs/errors found in commercial software? Not talking about sw for power plants, space/aircraft or medical devices or other critical infrastructure.
Looking for sw types
Word processor
Spreadsheet
DB server
Compiler
IDE

Are there established quality criteria for sw code?
If yes, how do dev teams measure quality?

Okay, let’s have a deeper look on Mysql. Around 6900 active Bugs. And over all since 2003 around 74000 Bugs.

Lethes dive in the history. I’ll pick an example out.

https://bugs.mysql.com/bug.php?id=16

The most of the Bugs having a bugfix which is pushed to the Sources. While MYSQL builds Updates relatively often, there is a fast fix possible.

But let’s look on a new one. MySQL Bugs: #363: SHOW SESSION VARIABLES LIKE 'x' != SELECT @@SESSION.x

Here we see a Bug which is a documentation Bug. Really hard to find. A long discussion. I think that shows how it should be.

I think the number of bugs is of less importance, than

  • First Response Time.
  • Priorization processes transparent and in place.
  • Average fix time per category (showstopper, business critical, error, nice-to-have, new requirement, documentation, etc.).
  • Quality and speed of fixes (how often do fixes break something else).

The above ones (of course not complete) can be measured. Then we have soft factors too:

  • Seeing feedback as a gift and a help, and not as an offense against the product.
  • Acknowledging that showstoppers for a rare group of users are still showstoppers for those people.
  • Clear transparency why something can just not be addressed right now, though still acknowledging that it is a problem.
  • is a company hiding bugs (closing them by miracle, or deleting them).

Meanwhile important to me: is there a LTS version of the product.

1 Like

I dont recall the specific ISO document or process # but I do know there is one for software. (@thorstenstueker do you know this one?)

Its more a well defined process than it is a specific set of metrics about “x many bugs per K lines of code”

As @Jeannot noted the process is what’s critical

There are standards like en ISO 62304 and also iso 9001 also iso12207 which All are also IEC standards.

For a very long time, I reached no bug at all.

Then, I started to see some and now is Xojo (since 1998-03 for me); the transition for that product started in the days of REALbasic 5.5.5 / REALbasic 2005; then the debacle with Carbon to Cocoa transition (so long time between the ANN of the transition and the Cocoa Product Release). I do not enter into the “How the hell I enter my license # in Xojo 2013 ?” (lack of information).

Some days ago, I saw a release add saying “removed bugs…” and I was asking to myself that… I never saw a bug in that software; just a strange behavior (versus other similar products), and I live with it / I do not understand why drag and drog does not works there, but…

This is what makes it hard for beginners with Xojo, who take significantly more time than an experienced Xojo dev to figure out why something is not working as expected and to identify if it is due to their own code or if it is a bug. Add wrong or missing information in the documentation to such a case…

Yes and that can drive crazy while you can’t find a solution. In that moment I have to decide: is that a strange behavior which I can accept or - ro pick up Emile’s example - is it stopping the show while I urgently need at that point Drag and Drop functionality. If not…okay. If…story ends at that point.

And therefore it is so that it is even if it will not really interest me: a Bug. And even if I say it is not from interest: it may stop the show for others or later for me.