SciPiTie

Ahh g I don’t use paperless as an exclusive document storage but as a pure manager. It searches and tags but doesn’t have exclusivity over any files but it’s own indices!

It doesn’t provide more value than jellyfin in that regard - make it visible and accessible.

Worst case I have all my OCRed documents as raw files which I can migrate to whereever.

Files still exist. For my case encrypted as well. My backups roll on the data, not the container.

But I’m not trying to convince you, I tried answering the questions :)

And two answer your last question clearly: I survived before paperless, I’d get along without it. I find a new tool to mitigate my manual labor as good as possible - if that’s not possible then jo harm done. I know I’m flexible, I can learn new tools and I’m never vendor or tool locked-in. I have a high level of self confidence when it comes to my tool chain and how I’d adapt any part of it - from password manager to cloud storage and my mail flow.

To be honest I couldn’t self host anything if I’d had the fear of being lost if a tool is discontinued.

For me it was a few hours wrapping my head around how paperless ngx works and its setup. I had a folder structure as you described already on my Nextcloud so I just configured paperless to observe it for new files.

Where I spent more time then reasonable with was the tagging - you can automate it based on… Well everything.

Now I just let it suggest me tags based on my existing documents plus add a NEW tag to the ones I’ve never reviewed. That’s just a reminder for me though to review tags when searching, I don’t actively re tag new uploads.

If you have a docker environment I suggest just pulling a container up3, throwing all your documents in it and see if it would save you time or cost you time. Would be an hour well spent!personally the OCR alone is it worth it for me - my country still loves paper letters and being able to copy text out of that is awesome (IBAN, account numbers, etc - all the stuff that’s suspectible to typos).

The three letters OCR, tagging, fuzzy search and ease of use are the ones for me.

I never needed the date for a letter but quite often its context for example.

Your suggestion just digitalizes physical folders. If that’s enough for you ok - but you’re missing out.

Just curious is there any recent quantitative source to this? That statement was “common wisdom” already 20 years ago - 10 years ago I decided to just give it a try - and had issues three times in ten years, all three with missconfigured exchange servers.

And I’m not with a high profile provider either.

Just to make sure: I’m not claiming that you’re wrong, I’m simply curious on how lucky exactly I got!

Ohhh now that is awesome and makes sense! Thanks a lot for that find :)

But when I mount a shared /usr on a remote machine it will always have the mount point /usr/local as empty folder - and either have an empty folder or have a mount target that is dependent on a network resource - that’s why for me it’s so unintuitive.

But then again I started with network stuff way more than a decade after all this got created 🤣

This is really helpful, thank you!

I never understood why the shareable /usr is parent to the non shareable /usr/local. Wouldn’t a /usr/shared be way easier especially in the early network days?

If anyone has a link or some insights into this historical nitbit I’d highly appreciate it!

You are one of today’s lucky 10.000!

https://www.explainxkcd.com/wiki/index.php/1053:_Ten_Thousand

“muddy waters” is a saying, I don’t think you should take OP literally. The Rest you’ve written seems to agree with their sentiment btw.

There is literally not one singular(!) arr that does what you’re claiming, at least that I’m aware of. The indexing is done by a different thing than the tracking and the downloading.

That’s why you end up with 16 of them like OP after all…

The router is not directly involved in a dns query except, we’ll, the routing if it’s an non local IP. The DNS ip addresses is propagated either via dhcp together with the clients or directly configured in the client. That said: most routers serve as dhcp server at the same time. Perhaps your router is configured to always provide your ISPs DNS as primary.

How the client handles the decision which to query I honestly don’t know and I guess that’s why you and I made different experiences!

The client does a fallback if one dns doesn’t answer. That’s why dns ad blockers fail if 8.8.8.8 or some other dns is added as a secondary :)

I second openhab. Can’t speak for too many integrations but all I tried work without issues.

Especially the separation of abstraction layers is something that I came to appreciate highly. You have the physical object, it’s item representation and then the rules and interactions. On the downside might be the way that this abstraction makes the configuration a bit more complicated - but as you’re missing the yaml config you might enjoy the configuration files! I’d just give it a shot :)

HA has a sour taste for me since their broken promise about open sourcing their server side. It’s still a black box. Plus the whole dns debacle a while back. And I honestly don’t understand how HA is still the de facto standard for home automation - I tried recreating some of my more complicated rules in HA and it became such a mess very quickly (think of 3 or 4 non nested conditions and altering the states of multiple objects depending on virtual items).

Just curious: why?

I never tried proxmox that’s why I’m asking :)