Johan G wrote in Mon Jun 08, 2015 1:56 pm:@hooray: Umm. I think the intention was to keep the discussion on the developer mailing list...
Like I said, I don't have much to add beyond pointing out that this document is contradiciting statements made by a number of senior core developers, and that the repeatedly expressed concern about Qt5 being a major dependency, shouldn't be taken lightly - especially given the low number of active core developers, and the significant barrier to entry for those wanting to get involved - in fact, this could be compared to the migration from plib SG to OSG, which also wasn't too well embraced back in ~2006 - yet, you will see that adding Qt as a lib has been brough up by a number of contributors over the last decade, and there were a number of strong arguments made against doing so - for the record, I used to belong to the camp of people suggesting this, too - but I have to admit, that many arguments still hold true today. I have no intention to have this discussion here - as usual, "roadmaps" are more about documenting priorities of those involved, not of the overall effort/project.
And like hamzaalloush mentioned, phasing out PUI/PLIB is another long-standing goal of those involved (especially Zakalawe, see the devel list/forum for countless statements on PLIB/PUI) - so accepting a more modern, and actively maintained, dependency is a worthy thing to strive for.
To people familiar with core development, and those who have professionally used FlightGear, it is becoming pretty obvious that much of the existing SimGear funtionality could be easily made obsolete fairly quickly once Qt5 is accepted as an official, and required, build-time dependency (think I/O, networking, logging, property tree, UI etc).
This isn't necessarily a bad thing though - it is just a little surprising that a "google hangout" session among a closed circle of core developers was apparently sufficient to rule out all arguments previously made by other contributors against "requiring" Qt5, including people like Andy Ross (YASim/Nasal) and David Megginson (Property tree, SGSubsystem etc) back in the day - or more recently, Rebecca Palmer (again, refer to the wiki article linked to above, for the specific postings/statements).
Again, I am not affected by this, I do have a working build environment with Qt5 support available - but I guess that doesn't necessarily apply to many others, and given that Qt5 isn't exactly a trivial dependency, I would have suggested to tread more carefully than has been the case here apparently.
Either way, I am convinced that should this remain the "consensus", that FG will significantly improve over the next 5-10 years due to this decision, not unlike the migration to OSG has been a technology enabler, too. It may just be more inconvenient than many/most of you realize currently....