[Interest] Semi-OT: Was Nokia net good or bad for Qt?

[André Pönitz]

On Fri, Sep 27, 2013 at 11:20:24AM -0700, Thiago Macieira wrote:
> Every year in the Qt Developer Days plenary sessions, the audience asked for 
> more bugfixing, fewer new features, and definitely no regressions. We listened. 
> So instead of breaking QtWidgets by refactoring it, we kept it as-is, we're 
> fixing bugs, and we're introducing a new solution, step by step, so we can 
> achieve the "code once" goal again.
> 
> Tell me that was wrong.

Implementing the features that have been asked for while keeping the
stability of the existing stack might indeed have required a second stack,
which might very likely have had a different API. However, the scale of
additional differences on the frontend side of the presented solution is
way beyond what is necessary and not what has been asked for.

People did not ask for replacing a well-known standardized language with
established development and deployment processes by some ad-hoc domain
specific language without similar provisions. People did not ask to shift
their compile-time effort onto their user's startup and run times. People
did not ask to depend on technologies that are easily, and on some
platforms commonly, blocked by downstream distribution channels. Etc.

Listening to the audience was not wrong. The presented solution partially
is, as it bundles compulsory dependencies which are technically not needed,
and often enough counterproductive.

Andre'