If we could have a solidly performing Linux mobile that has the capability of docking into a full desktop OS, that shit would be an absolute game changer for personal computers.
Give me something like the OG Moto Droid, or hell make it a tablet and I have to carry a bag, I don't care.
That would be my device for everything. I just remote into everything else anyway.
Its perfectly viable to run your support software on your own hardware (whether local or VPS).
I do this for myself, as well as for companies sized from 50-5000 (roughly). Larger ones deploy off my specs. The question to me is what is the plan around it. How will backups be handled? What if it goes offline due to a hardware failure? Do you have backups in place? A cold or hot spare? Multiple machines in an HA configuration? Do you need to go to that level if there is an outage?
I also prefer to make use of solutions with a support model that allows for locally hosted, but has a phone number that can be called. Part of this is because I don't want to field all these calls, part of it is for the comfort of the client that they have a number they can call (or a dedicated email, whatever, the point is a support contact not how they are contacted), and part of it is to support the project.
My wife has a (small) business, I have a small business, and I work for a consulting firm (design and engineering). All three make use of on-prem f/loss, all three pay support fees to those projects who do that (and random annual contributions where possible to those that don't).
So the short answer is: Figure out your requirements and your disaster recovery scenarios, then figure out what option works best for your needs from there. Cloud, VPS, or internally hosted are all viable, and all come with their own pluses and minuses.
Entirely possible, I know they have commented before on not wanting to do it.
Which I kind of get if you're parsing metadata, you'd have to create a whole new method for parse and view which can be a pain. Though I'd hardly call it evil myself since its how I sort too.
I'd agree on having to think and duplicates are possible, but that's possible in so many scenarios.
shrug
Either way metadata I think is important, but ive been a pain about metadata since the 90s (because burning CDs and having the actual track info was awesome).
I was wondering because all I do is /artist/album/, which is going to be the same logical appearance in navidrome as at would be from a dir structured view so couldn't understand the point.
The difference is the metadata, which I'm much more strict on so I haven't seen these sorts of problems.
Navidrome is what I would call the daily driver, but since its just a share that JF can also see, for TV playback its JF just like you.
But I'm also a stickler for metadata so I don't often run into issues. When I do see something come up thats not right, I correct manually. Which wasn't happening very often until lidarr had the metadata issue with musicbrainz, though the past few months its been a manual effort.
Thats basically an EV side-by-side, which can be made street legal in a lot of states BTW.
Edit: side note, EV conversions of older side-by-sides aren't too bad a process. I'm in a debate with myself of which route to go, but its mostly going to depend on what I can get my hands on. There are usually a good chunk of used kubotas by me, so I'll probably grab what I can used and working, and another thats not working/needs repait for the ev project.
A recent comment along similar lines was complaining about liberals saying primary them and then vote blue, which this is the exact case of.
So unless they want to clarify their position, I'm going to go ahead and take it as read with zero expectation of sarcasm being present. Too many are posting comments along the same lines with absolutely zero sarcasm for me to safely assume that to be the case here.