this is like month+ old info, at least mention that when posting. otherwise people assume it’s an update.

I deployed RocketChat on two different client installations (didn’t check the licensing you’re mentioning, I’ll have to look into that) and I run a Prosody instance (XMPP) on my own; tried Matrix for a short while and ran away from that mess as fast as I could. anyhow, although the messengers work without any significant issues or downtime, the amount of flak I get from non-tech normies about the client apps is staggering.

the apps just aren’t up to current UX standards. they’re used to Twitter and iMessage and Telegram quality UX, and getting used to these PoC-quality apps - both on mobile and desktop - makes them “feel icky”. I’ve had to intervene on a number of occasions when some of them transferred their business-related comms to other platforms because they just can’t/won’t get used to these apps.

check out dell latitude 5285/5290 2-in-1. they are Surface Pro lookalikes with detachable keyboards, but with way service-friendlier interior - easy to open and SSD, comms, battery can be easily replaced, whereas RAM is soldered. the screens (12" 1920x1200 IPS mutlitouch) are gorgeous and the hardware isn’t too shabby, kabylake (7xxxu) and kabylake-r (8xxxu), with standard UEFI BIOS so you can install Linux and have SecureBoot even. I can get them locally for $100-150, dependent on config and equipment (even less if they’re without battery and keyboard).

there’s no observable difference between a T420s, T480s and said T14 Gen1 AMD, they all lose battery while suspended. the battery drain is endemic to practically any modern laptop that’s not a Macbook Pro running macOS; even the MBP running Fedora has the same drain.

it’s easily solved though by implementing suspend-then-hibernate; the laptop sleeps normally and if not woken in, say, 60 minutes it hibernates to disk na shuts all power off - zero battery drain. once woken, it resumes from the disk super-fast, faster than cold-boot.

regarding the pricy enclosures, there are vastly cheaper eGPU solutions especially if you’re able to utilise the on-board M.2 or mini-PCI slot. if you don’t move the laptop around, it’s a viable option. this would be an example - not an endorsment. you’d need a $15 PSU to power the graphics and it works well in linux, with the hotpluggability being the primary issue; if you’re willing to shutdown before attaching the eGPU, close to no issues.

you can run it as graphics card (i.e. utilize its display outputs) or just use the laptop’s display with optionally switching between the onboard and discrete graphics.

I believe T480 is the last one that has two batteries; T49x and T14 def don’t have any, possibly some P-series models, can’t check now, but those are workstation-class machines with discrete graphics, so not in the same ballpark.

I don’t know your use case, but the dual-battery isn’t necessarily a good thing; apart from marginally improving battery autonomy (by like 20%), its primary use is you being able to hot-swap the second battery without shutting down the laptop. that (carrying around multiple fully charged battery packs) is a 20-year old paradigm that nobody uses nowadays.

my point being, going with a more modern platform nets you better battery autonomy, even with a single battery, and more computing power when running docked. case in point, T14s Gen6 (12-core Snapdragon EliteX) has but one battery but with 20+ hours of autonomy. that’s obviously a vastly more expensive device, but even T14 Gen3 with 6xxx Ryzens can sip power while on battery and go full blast when docked.

edit: can’t help you with the fingerprints as I routinely disable it on any laptop I have, but I seem to remember that all non-S versions mentioned ITT are good to go OOB.

all Thinkpad models (doesn’t extend to other Lenovo series) have exceptional Linux compatibility and support as long as you go a generation or two before the current one. there are some edge cases, like the fingerprint sensors on some less popular models, but if you stick to tried and tested models, you won’t have problems.

the S-suffix models are like the Macbook Air models, they’re thinner, have a single battery and have some more exotic materials used, like carbon, magnesium, etc. they also have one RAM bank soldered so that’s an issue if you get one with e.g. 4 GB soldered, you’re either maxed out at 8 GB if you want dual-channel performance or if you go with an additional 8 or 16 GB stick you’ll have some performance penalties.

I have a T480s (here’s how I got it) and even though it’s less serviceable and expandable than the non-S version, it’s light years ahead from the usual consumer grade models out there, everything is easily sourced and replaced, with detailed hardware maintenance manuals straight from the manufacturer.

anyhow, if you’re moving from some consumer-class model, you can safely ignore the additional thinness as even the standard T480 is portable enough in comparison to the usual drastic-plastic e-waste.

as to newer versions, I was merely pointing out that you don’t have to limit yourself to the exact model (e.g. T480) but look at other, similar models if you happen to stumble onto one. like, I got a T14 Gen1 AMD with a busted screen for about $100 and that’s a vastly better machine than the T480s. it’s hexa-core, way better graphics, connectivity, power efficiency, and since it’s younger the battery is in better shape. the soldered 16 GB limits it to 32 GB max, but that’s more than enough for my use cases.

to me the Ian Holm abomination was terrible, without any redeeming qualities. same thing as rogue one, exact same shittyness, zero technological progress, crude and unnecessarily distracting.

that’s about the crux of it. it’s a widely available platform that can be had for cheap, the replacements and upgrades are plentiful and cheap and cross-generational compatibility (both from earlier and later models) is exceptional. some notes:

• ignore the NVMe speeds; they’re perfectly congruent with the rest of the machine and there are virtually no bottlenecks for the use you mentioned
• no idea about the Guix thing, normal Linux distros (Fedora, Arch, Debian, etc) have OOB functionality for all comms; spotty fingerprint reader support though
• there’s a negligible performance difference between the top-of-the line i7-8650u and the entry-level i5-8350u, so feel free to ignore the CPU entirely. also be aware that earlier models came with the kabylake 7xxxu line, which you don’t want for any kind of money
• although panels can be upgraded cheaply and easily, make sure you don’t get a 768p panel. the factory 1080p stock panels are IPS and 6-bit, so they’re decent and power-efficient. if you’re into color-accurate use cases, you’ll need a 8-bit 400-nit upgrade at some point
• if at all possible, try to buy it without RAM and storage; they’re predominantly available with shitty 256 GB drives and one 8 GB stick and that makes a normally specced machine with 32 GB and 1TB+ storage unnecessarily pricier
• don’t shy away from newer models (T49x, T14 Gen1) if you happen to find them for cheap; T14 AMD for instance has one slot populated with 16 GB (1 free) and has only one battery, but it makes it up with a more modern, powerful and power-efficient platform
• if you plan on using it with external devices (monitors, sound, etc.) do yourself a favor and get a dock station. just like the T-series they’re widely available, cheap and work across generations; e.g. the 40AJ dock, you click the laptop into place and all your devices and PSU are connected instantly

using rsync:

why not btrfs send | btrfs receive? is there some advantage to rsync?

did you hotswap the drives after each btrfs replace or shutdown and then swap?

what’s your host OS and do the drives spin down if inactive?

thanks for the writeup!

I’ve set up a local mail archive with just dovecot + imap plugin. you don’t need the full mail server with postfix and whatnot, as it’s not intended to send anything anywhere, or even receive anything for that matter. it just sits there ready to be searched with thunderbird, no need for other complex solutions you’ve mentioned.

this right here. remove partition and apply and then start the installer. there is a way to map this partition in the installer, but that’s such a horrific UX mess that it’s a challenge even for experienced folks.

first off, I have serious doubts that any one dude - or even a group of those for that matter - can ascertain the security of such a complex system; a browser is essentially an operating system, with all the layers and complexities that entails.

even if you’re somewhat successful in such an endeavor, I don’t really care if it potentially is. chromium comes from those shitmakers and I’m not willingly using anything they had their nasty fingers in. they threw one shovel of shit too many on the heap and they are now forever on my ignore list. if that means that I don’t get to access certain domains, sites, and/or apps - so be it, I’ll make do without.

fucking edgelords… IA has trouble staying on its feet without this sorta crap.

this has “kicking puppies for palestine” energy - not sure where I’ve read this but it’s an apt analogy.

you can still run both tools, I do so on my main workstation. they stay in the status bar and don’t bother you none. when you’re done working and recline on your sofa, you can send e.g. youtube videos to the rig and they play fullscreen by way of mpv.

I’ve gone the other way - there is no interacting per se with the media PC; instead, it’s a dumb sink that plays back everything you send it, by way of macast and jellyfin-mpv-shim. you use android apps to send it stuff (e.g. newpipe share to allshare which connects to macast and jellyfin android app which connects to JMS) and to control playback (pause, skip, change subs, etc.). so, all media selection and playback control is done from the mobile device, no need to touch the media PC doing the playback.

not sure this will fit into your use case because of spotty internet, but that should prompt you to install jellyfin post-haste. then you have two options, the mentioned android app + JMS or just the jellyfin media player which can run in TV mode with a pared down controller (up/down/left/right/enter/back) - I’ve successfully repurposed an ancient Apple Remote that has just those six keys.

can you run it on multiple devices and have them all synced? how about chat export in a readable format?

hey E., tell Sloane I said “What up”

so I’m gonna pick this as a runner in Lutris, instead of e.g. wine-ge-8-26-x86_64 or whatever? sorry to reply to you, but I’ve read the initial github “explanation” what this does, reread it several times over the months, and now for the final time when it got official and I still don’t understand how this fits into my use case.

good messenger for what?

if you want a solution for you and a bunch of your henchmen to coordinate and discuss totally-not-crimes with ephemeral comms, practically any E2EE solution will work; once the not-crimen is done, burn your accounts and toss the devices for good measure and you’re scot free.

if you want a secure messenger that’s part of a widely used communication platform where you can also do normal people shit and also convert normal people to actually use it (think getting contact deets from cute boy/girl at a bar or giving yours to a business correspondent without an elaborate powerpoint presentation on how to use it) and you want to enjoy the fruits of 20+ years of continuous IM development, like having top-notch UX, battery efficiency, network resiliency, quality voice/video calls, etc., without being spied on then such a thing doesn’t exist.

how come? meredith baxter recently stated that it costs signal $50MM/yr to run their infra. that money has to come from somewhere. if there are no advertising dolts dumping cash on spying on your social graph and convos, the remaining avenues for financing are few and far between.

in closing, there aren’t any super awesome messengers you weren’t aware of, everything is shit.