QGIS is not your run of the mill project, though. This is specialized software and a rare breed. It's in their right to decide either way, and this is the right call for this niche, IMHO.

▲

mpalczewski 4 hours ago | parent [-]

they own the project of course it is their right.

if someone disagrees they can fork.

I don't understand why it would be the right call. not my field.

haven't seen a good explanation in the comments. (yet?)

▲

bayindirh 4 hours ago | parent [-]

GIS is basically "serious mapping software", and when it's used for city planning or other large GIS-relevant activities, accuracy and precision are non-negotiable requirements.

IOW, your city might be using QGIS to plan the city and verify everything being done on it, down to cm level accuracy. So, any subtle bug is a big deal for these kinds of software packages.

While doing construction work (regardless of infrastructure or buildings), the plans are laid and verified to that level. The haphazard spray paints on the ground are indeed big deal, and has a digital copy in the GIS software.

It's not only used by municipalities and/or governments either. Research teams use GIS packages extensively as well. I briefly worked with a research team working at Antartica. These GIS software and related ecosystem is their both bread and butter and foundation to anything they do.

A patch generated by an AI model breaking these software packages have cascading and important effects. So, the software needs to stay dependable and reliable.

Data portal for South Korean team: https://kpdc.kopri.re.kr/

They're not the team I was working with, though.

▲

mpalczewski 4 hours ago | parent [-]

ok it needs to be reliable. do we believe people are more reliable than ai?

▲

bayindirh 3 hours ago | parent [-]

Short answer: Yes.

Long answer: These software packages generally have a couple (or more) graybeards on the team knowing the intricacies of the domain and the software pretty well, so they can read a patch, see the effects, make an informed decision and give required feedback, or sometimes get the patch, mangle it to correct form and merge.

LLMs tend to write overly verbose code with many file edits at the same time, accompanied by walls of text to justify itself. This overwhelms these kinds of communities. So, they prefer targeted, smaller patches arriving at a slower rate.

...and maybe that big patch can be written as a one-liner somewhere else, just because AI failed to find where the problem originates and just patches where the problems happen, making thing way more complicated than before?

These are complex software packages and are sometimes things looking correct are wrong, and things looking wrong are correct, esp. on the math side.

I developed software using boundary element method to evaluate things, and it was complex to put it mildly. Verifying everything down to 32 significant digits after comma, and tearing the code down when the tests deviated for no apparent reason is something I did regularly.

GIS is no different. You're modeling the world, down to the earth's curvature, and putting things on it with cm level accuracy, and a couple of people knowing what they are doing is way more effective than a statistical model which emits code with no bounds and understanding of the domain.

	▲	mpalczewski 3 hours ago \| parent [-]
		oh. so basically, no low effort patches. Sounds like llm's have made is so easy for anyone without a shred of know how to submit anything that people prompt one or two times and consider the job done, not knowing any better becausue they don't have to.