oh just reading this now. let me try and answer some things.
first of all, glad you got it working. it requires a fairly recent vulkan/nvidia driver (i think one from january will do but that’s apparently not a given in most distros). apparently now it also works with wayland and amd cards, but i haven’t tested myself.
yeah, lack of pain: means that i can now develop tools that would usually be too expensive to do. also always processing the full buffer means way less pain developing modules. not sure i care about ignorant persons and drugs.
and yeah it’s possible to slow it down if you input multiple 50MP raw images into the same processing graph or things like that.
re: orientation: this is read from exif tag if you compiled with exiv2 support (see the file bin/config.mk and the .defaults version of it). the crop module takes care of it, and it has a rotate slider (which is not present in the default darkroom.ui file that determines which sliders will appear in the favourites tab). probably you only looked at the favourites tab. the crop module allows perspective correction, cropping, and rotation. the ui is terrible/autogenerated but works for me ™.
re: draw module: there is a button insert block.. in the pipeline config tab. it has a draw.cfg that connects all necessary modules to setup some basic stroking. i don’t really use it so i hope it still works.
re: sharpness at 100%: there is no sharpening whatsoever applied by default.
re: thumbnails: here it is much faster than dt. it’s limited by disk io, so that probably only shows when creating the images from an appropriate ssd (it’s a lot slower from my external usb drive).
re: crash: i’d be interested in a stack trace or at least a short description of what you did to cause this. i’m occasionally fixing crashes, but often don’t even publish on github (because it seems not many people run recent drivers or have appropriate gpus at their disposal nowadays).
similar for the dependency list. what were you lacking? in fact i think the readme is outdated, rawspeed is optional now (in case you’re not interested in raw photography).
re: artists: i’m using it as a testbed for interesting processing. it has temporal image alignment, it supports temporal feedback in general in the graph, it can read magic lantern .mlv raw video files and do a bit of 3d rendering. it supports key frames so it can be used to process timelapse footage etc. i can experiment with a lot of interesting methods that would have been a pain to develop both in terms of code and speed in vanilla dt. this is where i currently see the value of it. we’ll hopefully get an all new toolchain/processing pipeline out of this.
developing software that is useful to end users definitely needs more emphasis on user interface. at this point i’m just more interested in the core processing part.
so any input/preferences you may have are interesting to hear for me at this point.
If needed I can can also borrow and try other cool input devices like Magic Trackpad 2. I’ll probably get that for myself too after the release of GNOME 40. It seems like gestures will be a thing in Linux and it might be helpful to have it. It’s a great device for doing video and scrubbing trough footage too.
:shaking head: I think if I were a developer I would rather develop tools that can be used right now like for real art, but… there is no doubt that people with a “vision” are also necessary for progress.