since you’re directing this at me i think i should probably comment :). in general i think all of your observations here are mostly accurate. there was a strong vision about performance, modularity, language, and ui design in the beginning i would say. but also the dt project was always about building a community of people who enjoy working with the software. we always used to develop for our own needs. and i think this lively community is still strong and something to be proud about.
the downside of this concept is that pretty much all of the initial guidelines have been discarded over time (people have their own ideas about “direction”, they are grown-ups). but more than the occasional frustration that some features don’t quite work as i would have done it or are slower than i would wish (cf. community above, my opinion is also just one), what drove me to start vkdt is really that the times have changed. in 2009 i evaluated a pure gpu pipeline and concluded that the hardware/drivers weren’t ready. designing the very core for 2020 results in a few quite different decisions. same with the gui stack (xorg/cairo/gtk/8-bit/cpu). to a point that starting something else is easier.
(on a personal note i’m not so much into “strong” leaders, i will not agree that such a concept ever makes the world better)
Most importantly, thank you very much for taking my thoughts constructively.
I whole heartedly understand your note on strong leaders. However, bad strong leaders are in for the power while good strong leaders/visionaries have earned their right by the community. E.g., Matt Mullenweg from WordPress is not a (traditional) strong leader but sets the path somewhat gently/firmly. See 2 latest posts (aug 1 and sep 19): WordPress.org Project – Just another make.wordpress.org site
A clean slate?
‘With times changing’ it would be interesting to revive your vision/initial guidelines in a written form for vkdt, maybe stretch out and join forces with R&Dt - respectively not knowing anything about the perhaps ongoing, technical or personal rationality (no pun intended) - and not the least, make progress and peace. I think you have the right to do so keeping in mind that you can’t win them all.
Fair winds.
“dt, the best light manipulator on this pale blue dot”
AP is still contributing to regular darktable, and although he has demonstrated interesting in porting his work to vkdt, it seems like there’s still some work ahead that will land in darktable(both regular and his fork).
Nobody is at war here, so there is no need to achieve peace. People disagreed in the general direction the UI was going and created a fork, it happens all the time. Work can and still is being ported to and from both forks (Like paperdigits said up there). A good example is RT and ART as both diverged a while ago but now they are both benefiting from each other’s new features. Progress is still being made.
Maybe contributors will eventually move on to vkdt, but that still seems like it’s a few years away, specially when it comes to reaching some sort of UI parity with darktable.
His vision is concisely captured in the vkdt source code, both in what/how it implements and what it doesn’t. Mapping that to what end users will consider is down the road a bit, methinks…
If one developer doesn’t want to contribute any more, that up to him.
Maybe a crossroad - Yes.
If by “usual track” you mean: constantly trying to improve algorithms, adding features helping to develop raw images in the best way, fixing bugs, improve UI for interested people, try to support dt newbies, then I fully agree.
Yes. But we all work in a team; every single dev needs to be interested / have a desire to work in a cooperative way…
Hi! I’m reviving this thread just to get a feeling for the state of Aurelien’s fork which appears to be somewhat forgotten (last commit is from beginning of August). I did post in github’s discussion a short comment just as a way to short circuit AP’s comment but I had none.
Now, it is true that I am using R&Darktable as my only processing/organizational tool and don’t feel like anything’s missing but you know how it goes, you kind of feel left behind if there’s no one else around.
thanks Josema! this is good enough for me… I was probably expecting at least a ping from him on github just to know he’s alright but ok – he’s right that the current master works with no problem for me and I have now lost any appetite for new stuff that I need to learn from zero (I still have so much to learn on the old stuff!).
However, and this has nothing to do with the main topic, what is “matrix”??? another ‘twitter’? a mastodon clone (I don’t even know what mastodon is tbh)? I have checked out of so many social networks now, but I realize that sometimes there’s no email to contact people you want to talk to…
Only for a short while - on Linux they will soon be restricting it to one particular distribution (I don’t remember which, other than it’s not Fedora, which I use )
I think there are quite a few nice improvements in DT’s upcoming release and most don’t require any deep effort to learn… the new HLR is now compatible with filmic and sigmoid and works very nicely so no fighting with or much less the magenta cyans…
I suspect AP will dive into that other project so he may do tweaks for his own but he always seems quite focused and intense for the moment …also sound like he is working on a project that he could potentially monetize…
Well I would never speak for you or others … But the sigmoid alternative is a nice simple option to fighting v6 filmic. I actually had gone back to using v5 and found I could manipulate that in a way that had less artifact and met my expectations… But there are some nice new features. The new HLR works very well and so you likely wont’ see issues with the cyans and magentas and if for some reason you do you can tweak them and they are not nearly as cpu intensive in my experience as Laplacian. The filmic settings have been tweaked a bit and so I think they offer less chances of issues. There is a lot of work on the pipeline, wb issues, If you use styles you can preview them now by hovering… I am sure there is lots more but there are not AP modules that require several hour long videos to digest… IMO it would we worth it…
Tried R&Darktable - could not find the shortcut to export from the darkroom (used to be Ctrl+E). Now I have to go back to the lightroom and export from there. Not very convenient for my workflow.
It is very nice to see Sigmoid in the new version of Darktable. Simple orthogonal modules with fewer artifacts are always welcome. I guess I’ll stay with DT though these days ART gives me better results faster.
Yeah, I was thinking about sharing some images to compare DT and ART results. What is the best way to upload images - in this thread or in PlayRaw or something else? And how do I upload them?
The author of Sigmoid explained what he meant by orthogonality. Basically, it is when changing one module control does not require tweaking other controls to keep results good.
)