Re: Stop making unneeded improvements

[Andrej Rode]

Hi all, 

to jump in from a point of someone who has contributed "improvements"
over the last couple of years.

Many of your points are vaild, I understand your frustation and pain of
continuouly having to adapt to new methodologies. 
Believe me when I say we are not these changes are not implemented just
for the fun of it. All of the changes you mentioned were mostly forced
with a gun on our chest to either implement a change or to simply not
have a usable GNU Radio for new Linux Operating System Releases. 

One of the reasons is that most of the GNU Radio development is
done by volunteers in their free time. Changes to GNU Radio reflecting
changes in dependencies which would have been useful to implement long
before said dependency is obsolete have been implemented in the last
possible momont, e.g Qt4,Python3. This lead to partially
untested/unmature code being pushed into a release. For at least Debian
and Gentoo GNU Radio has been the last package either on Qt4 or on
Python2 and patches have been backported to GNU Radio 3.7 by the OS
maintainers (Thanks!) to keep it in the operating system.

It's also quite hard to demand 100% backwards compatibility for
breaking changes and tools which provide full coverage for conversion
between breaking changes. I know the Python tools and I love them. But
development of these follow the Pareto principle. 80% of the tool is
written in 20% of the time. 80% or similar is what we are able to
provide and what gr_modtool provides in terms of conversion. For simple
cases conversion just works, but for complex setups you have to add
some additional changes by hand. 

TLDR: these changes are partly forced on GNU Radio by having a list of
dependencies. Core developers are doing their best to give users the
ability to convert between versions, but it's lacking and any help is
appreciated.

Cheers
Andrej