I had my Pentax modified to monochrome
Sometimes the reality is much simpler than anticipated. *facepalm*
At least for darktable it is sufficient to set the "Custom Image Style" with the four-way controller to "Monochrome".
To avoid debayering automatically it is sufficient to create a preset in the "demosaic" module where the "demosaicing method" is set to "passthrough (monochrome)". In the "store new preset" dialogue window activate the checkbox "auto apply this preset to matching images" and in the "model" text box replace the "%" with the camera model as is it detected in the "image information" module - in our cases "K-5 II".
But that still doesn't provide the answer to the whether there is a general way to provide a set of EXIF tags to "force" the image processing tools to treat the images as monochrome.
Edit: Last night I had a thought... What if not the value of a specific EXIF tag is the decisive thing, but the absence of a specific tag?
The good news are: I'm down to seven EXIF tags which might have an impact on the behaviour of the processing tools and I could enlist two supporters for the test images.
E.
Uploaded files:
exiftool has some limitations when it comes to change or delete some special EXIF tags so the reverse engineering approach seems to be a dead end at the moment.
Current state of the investigation: After reading through the DNG specification, I'm now digging through the source code of darktable. Today I found the code which checks some EXIF tags from the IFD1 part in the DNG file.
My first attempts to manipulate these values in the DNG file seemed to be successful, but darktable still treats that file with demosaicing. Probably that was not enough...
E.
Small update: because I got stuck in the source code of darktable and I didn't have the knowledge to understand fully what's happening there I decided to ask in a dedicated darktable forum. They responded quickly and suggested that I ask in the pixl.us forum. I'll do that tomorrow.
E.
Time for the final report.
EXIF tags are just information. There is no way via the EXIF tags to force a software to process an image in a predefined specific manner. The processing pipeline is defined completely within the software and only in the software. If eg. darktable interprets specific EXIF tags (as I found out: the tags of the embedded JPEG image in the SubIFD1 part), other software like digiKam or RawTherapee may use other tags to recognise whether an image is monochrome or not. Probably the "Monochrome2DNG" guys tried out a lot of processing software and found all the EXIF tags to "convince" all software. 😉
So now the converted Pentax K-5 II is set to monochrome style and I defined a preset in the demosaicing module in darktable, which applies the demosaicing method “passthrough (monochrome)” automatically if the camera model matches “K-5 II” and the image is monochrome. That worked fine. The countercheck with a chromatic Olympus E-M10II and an image in monochrome style showed the expected behaviour: embedded JPEG is displayed monochrome, the RAW data is correctly demosaiced and displayed in colour as it should be.
That investigation was very educational. 
E.
After a few months of experience with the K-5IIm it's time to do something like a report.
First of all: going fully monochrome is some kind of dedication, commitment, pledge, reduction, going for the very essential. You have to develop a habit to work exclusively with lights and shadows, for contrasts or for the absence of these. That is fairly challenging in the first time but after a while in some situations there is the thought "fortunately I don't have to cope with the colours here".
Working with coloured filters (green, yellow, orange, red) is somehow "ancient" and it decelerates the photography.
The K-5IIm itself shines most of the time with the higher resolution (I'd put the monochrome 16MP images of the camera in the same resolution class as the images of 24MP cameras with CFA) and a small file size for that resolution. The higher sensitivity expands the low light capabilities for at least 1EV and due to the "missing" CFA there is no chromatic noise at all.
So we have a lot of light and now it's time for the shadow: The dust removal system doesn't work any more, because the necessary piezo element has to be removed to convert the K-5II to a K-5IIm. So I have to work out a habit to tackle the "dust specks on the sensor" issue. @james-warner-b How do you cope with that?
Conclusion:
Was is the right decision to purchase a second hand K-5II and have it converted to a K-5IIm? Perhaps.
Would I go that way again? Perhaps, but probably not.
Would I change something? It depends. If I want to have a monochrome Pentax camera to do a lot in low light, then the K-5II is probably the best choice. If low light (like in my case) isn't that much, I'd purchase a K-3 or K-3II for converting: even higher resolution, manageable file sizes and a working dust removal system after the conversion.
Would I recommend the conversion? It depends. If you have the budget for a special camera and if you're doing a lot of monochrome photography (and if your're willing to work with colour filters), then a converted camera can be a revelation.
E.
There are news from the "post processing of monochrome imaging with open source software" department:
Currently the darktable module "hot pixels" isn't available for monochrome images but one of the developers put it on the list for darktable 4.6 which is expected for Christmas 2023.
For the record: ART (or RawTherapee itself) can remove hot pixels. That's why I'm switching to ART for monochrome post processing - at least currently.
E.
