usernam

I still think notmuch is a great backend, but a poor client. Despite using mu4e, I still find Gnus superior in terms of pure reading efficiency (I still use Gnus regularly with gmane). Scoring and thread handling is just superior.

There's a lot to gain from adaptive scoring alone.

In fact, I also switched over mu4e after trying a search backend for Gnus, and I've been oscillating from mutt/gnus/mu4e a lot over more than a decade.

Tagging is powerful, but I find it too constrained as a single tool to manage everything.

If you're using IMAP, deletion is a thing. I often get people sending me huge stuff I have no intention to keep or sync. I need to be able to sync the deletion flag. This is a one line change I did myself and used for a long time, but it doesn't get merged because as messages are currently indexed by message-id, tagging a message for deletion might involve more than one actual underlying message.

I also use folders as a mean to exclude useless stuff from syncing. Spam is the most glamorous example: I do not sync spam. I was able to get spam to be moved in/out according to a tag, but only with afew, and again with other issues involved. Doable, but again, ...

Also, notmuch in emacs is very much thread oriented. But with email, I found the behavior very annoying. People will mail you in response to a 1yr old message, and notmuch will happily bring and show the thread from something which is completely unrelated. I found myself switching in/out of tree mode too often, and then just gave up.

I actually enjoyed (and still enjoy) Gnus much more. It's just more efficient for large message volumes. It shows you what's _new_ by default, which notmuch doesn't seem to get right.

Hopefully this is actually a MUA, and not a search interface with message-mode sprinkled on top?

I tried using notmuch on emacs as my primary mail client for several months, but it falls short on so many things (just to name one: cannot delete messages, cannot have copies in multiple folders, cannot apply a tag to an individual message without passing by the threaded view first, no notion of folders, and so on...). You can work around most of those, but it's just a work-around.

It's right that it's "not much" of a client. Notmuch combined with afew is incredibly powerful as a backend. But it's not a particularly good mail client.

You can use notmuch efficiently with neomutt, which is a great combination (excellent MUA with fast backend).

But for something closer to a real client on emacs itself, mu4e is much better. It supports identities out of the box, and has no silly limitations imposed by blind message-id indexing. Unfortunately, mu (the backend) has no concept of batch tag processing, which made certain operations very convenient [one thing I did was killing threads I didn't care about - currently impossible to do with mu alone].

The OP was right, though. Just making the data accessible and documented is a huge undertaking. It's truly a laudable effort.

From the other side, I cannot imagine using the data without intimate knowledge of the entire setup. There are very few people that can look meaningfully at such kind of data, which is why publishing the dataset (and all the work it involves) is often neglected.

Not to defend the behavior in any way, but remind yourself: in the current research model, if you can't make a publication with it, it's not worth your time. For a dataset, I've seen people factor-in the effort to make it public VS the potential citations you could get out of it.

disclaimer: I work in research

I find Discourse absolutely horrendous. Discourse mailing-list mode is a tack-on at best, and the formatting you get in mail notifications basically require you to be online to visit the link. And the gamification is also a huge put-off.

Web forums made us forgot how far we came with NNTP years ago. Even the shittiest NNTP could do subthread splitting, hiding and scoring. Handling hundreds of messages per day was easy. The convenience of doing-it-all under one powerful client was massive.

You can still do that with mailing lists, as long as it's not as badly managed as forums-with-a-wanna-be-mailing-list-mode addon as Discourse is.

You have to note: are you paying with the ATM circuit or with the credit card?

The ATM circuit often doesn't require an additional signature (although in some countries like IT it's still required)

In EU you have a pin and you also have to sign the receipt. In supermarkets, cards (debit or credit) queue up the checkout line.

Cash is way faster, hands down. I bothered to compare the difference once, since cards here are not the norm. We're talking between 5-10 seconds (cash register) vs 30-45 seconds, most of which are dead times (waiting for confirmation).

Indeed, the number seems disproportionate. But it can be possible depending on the definition of "medication".

I'd expect the number of mental health issues with an underlying [known] physiological nature (say, bipolar, schizophrenia), and thus mediated with the proper treatment to be _very_ low.

Mental health issues resulting from workspace, work, stress and in general lifestyle issues can still be medicated to help recovery, though the proper solution would obviously be a change in lifestyle.

I've witnessed many people with depression. I can only speak for relatives for which I know the full story. When faced with the issue, at least here, you have two choices. The psychologist will talk you into a lifestyle change whenever possible.

Going to a psychologist is still considered medication here, and it's covered by national health. When going to a psychiatrist instead, the medication (chemical) is almost always guaranteed.

It's helpful to say though that no amount of medication fixes mental health issues related to workplace issues if the underlying cause is not removed. This is a very dramatic issue, as for some people with limited working experience (and thus unable to have a leverage), it's often insurmountable. I known relatives in this situation for 20+ years.

Couldn't agree more. Gitolite is enough for most use cases, and for technical users which often use other tools there's no need for a web interface (or anything beyond gitweb anyway).

We use it in combination with BugsEverywhere for issue tracking, with the added advantage that bugs are tracked per-branch. You can work offline without issues.

The major complaint with this system is the disconnect for users to report issues. BE includes a web frontend, but it's not great.

We use mantis as a general tracker, but still keep detailed bug tracking in the VCS with BE.

Absolutely. BASIC on a Sharp 6000 pocket pc[1] for example was limited to (32k or) 64k lines of 256 characters each. Like in many other basic versions, there was no automatic renumbering. Discovered it the hard way.

[1] https://duckduckgo.com/?q=sharp+pocket+pc&t=ffab&iar=images&...

A local company I know of still runs under dos with a custom accounting software, still running on a memorex telex 286 pc, with a whopping 10mb hard-drive connected to a parallel port dot matrix printer, green crt and a mouse that would make anybody with RSI cringe just by looking at it.

The same setup has been used since the first eighties. Data entry is so quick you can barely see the forms before the secretary presses enter and stuff gets printed (the slowest process by a huge margin).

Would put to shame most software I've seen nowdays. AFAIK the system only uses 6mb of the 10 available and includes all accounting history of the company.

I have commented on #225 about this.

ripgrep doesn't yet support compressed files (z/gz), while grep/ack/ag do. On *nix it's extremely common to have sparsely compressed directories, from logfiles to non-changing documentation. I really wished ripgrep would just stream to a fast coprocess and support any stream compressor transparently instead of trying to use a built-in rust library.

The additional problem is that the search results are also incrementally updated. On an open dialog, I had in three different occasions click on a result just to automatically select an unwanted [new] result.

Also, the path bar is long gone. It used to support tab completion, which was actually faster than selecting files and/or searching for them. I don't know what they're thinking.

The list of gripes I have with GTK 3 is increasing at each release.

I don't expect anything anymore from GTK given it's current track record. The latest issue I discovered was the 'letter' to filename shortcut you used to have in file dialogs. It now triggers recursive search instead. Incredibly irritating, especially when used with large and/or FUSE directories, even considering there's a one line patch to allow customization though dconf which has been ignored since more than one year.

GTK 3 has been downhill usability-wise for me. One of the primary reasons I've stopped as a developer to use it entirely.

Does anybody know what's the _current_ status of window decorations in wayland? Client-side decorations was an incredibly poor choice in the original design. From the screenshot I cannot tell if it's just a custom theme or the "wm" is drawing them.

In a similar vein[1], using the python client:

    $ dl-cli some/files
    https://dl.example.com/t=<hash>

additionally, the file is automatically cleared when it goes unused for a while. There is also FEX[2], which is geared more towards power-users and includes bi-directional transfers. Both are fully self-hosted and trivial to set-up.

[1] https://www.thregr.org/~wavexx/software/dl/ [2] https://fex.rus.uni-stuttgart.de/

You're confusing the binary format with the tamper protection, or file "sealing". The two features are really independent.

systemd is not perfect. Debugging systemd is still very hard. There is no way to simulate what actions would be performed to reach a target without actually executing them. There are plenty of issues related to boot and shutdown which are _hard_ to track down because of this.

The journal is nice, especially on desktop systems, but it's way too slow to be used in any moderately busy system. The disk space utilization is also _prohibitive_ for archival.

We still rely on remote syslog because of this, but it's not without issues: the journal->syslog translation is lossy, and it's compounded with some "quirky" upstream decisions (purely cosmetic) to write unit descriptions in the log as opposed to the real unit name which make tracking down units harder if you don't have the actual journal file.

Journal files also get corrupted easily with unclean shutdown. In case of issues (from software to hardware), the local journal quickly becomes completely useless.

All in all, systemd is an improvement, but it lacks the debugging tools, performance and overall polish you would expect from a central piece of software such as init.

The internal APIs are mostly undocumented or self-referential. The DBus interface is opaque. Things gets muddy very fast as soon as you touch non-central components such as systemd-resolved or networkd.

I'm not a detractor, but there are many things systemd can still improve, but it feels we're kind of stuck. I'm quite happy if we have some competition here.

DNT is flawed. As an user, I have no way to verify that DNT is honored. As such, I assume nobody respects DNT and proceed accordingly by taking my own tracking countermeasures.

You can also assume DNT is pretty much ignored. For instance, if I set DNT and I visit a website adhering with those EU cookie regulations, while I'm still being shown that cookies are being used to track me and that by using the website I agree (never mind that 3rd party cookies are already being sent)? I already stated that I do not want to be tracked.

Oh, but maybe you would assume setting the DNT preference in the browser does something meaningful, such as disabling 3rd party cookies, disabling beacons, ServiceWorkers and cache lifetime?

Nope. There is no point in honoring DNT: you are either tracking or not tracking your users depending on which resources you're including on your website. If you don't want to track your users, do not include 3rd party resources. If you do include 3rd party resources, then it's it's up to the 3rd-party to honor DNT.

If you want to maximize for network effects, it's GitHub. I'm not keen on GitHub issues, and the lack of a mailing list is disappointing as well (which would funnel _actual_ social collaboration), but in terms of popularity is still your best bet.

By experience, GitLab is not worse and the performance and UI is ok despite the critiques here in HN. I do not find it significantly better to justify the switch though. Issues are similarly watered down, and there's no option for a mailing list.

SourceForce's UI is bad, their reputation is tainted, but the functionality is all there.

I'm unsure about launchpad. Their project page is confusing for newcomers, but their issue management is much better than github/gitlab.

I never contributed to bitbucket enough to comment, and do not host any project there.

I know and cannot deny the enormous effort of OCE. I tried to develop a simpler parametric 2d cad once for almost two years, but couldn't keep up with the complexity before getting myself /something/ I could use.

This is why I keep trying. FreeCAD is the closest package we got to a decent CAD in an OSS package, and I cannot wait to avoid using closed software in my designs.

But I'm just stating facts here: I couldn't follow the tutorial without incurring in crashes. I similarly tried CadQuery, but incurred with mostly the same issues.

But like I said, the issues are also on the UI, not just on the OCE engine.

UNDO management is often completely broken. Undoing line properties for instance used to only partially revert line properties, but revert previous geometry changes as a side effect. I got crashes just by toggling snapping states (even before hovering over the geometry that would cause some actual computation to be done). I thought I could help on those issues, but the build process involving OCE is time consuming and requires some non-trivial time investment before you can become an effective contributor (as for most large projects - cannot blame anyone here).

I occasionally use FreeCAD also as a simple 2d (non-parametric) CAD package. Still very frustrating. You would think we had more options for 2d drafting, but not really. I do not consider qcad/librecad to be real CAD packages. They're like toys in my eyes. I used inkscape (yes...) with greater results.

I keep trying FreeCAD every 3-6 months, but it's so full of little bugs and crashes it's not really usable in my mind. I'm using the version from Debian unstable.

The UI is unstable (very easy to crash just by fiddling around), and the geometric kernel has a ton of little issues that you have to work around. At every new project I have the liberty of giving it a try without many constraints (no import/export required, minimal assemblies of elementary geometry). I really wonder how can you actually go through the relatively complex projects I've seen as a tutorial, really. I would always give up after a dozen crashes.

I've used SolveSpace in at least 4 projects, and I absolutely love it. The UI is different than FreeCAD, but in a very positive way (after learning the basics, I genuinely prefer SolveSpace's a dozen times!). It's a bit sad that SolveSpace's kernel is definitely inferior in terms of capability to FreeCAD (notably: no offsets and no chamfers!), but what is supported _is_ stable. I cannot say the same in FreeCAD, where chamfers would sometimes cause degenerate geometry or downright crash. The only minor complaint in SolveSpace is that currently it's not always obvious to see the construction tree, as it's represented linearly.

I bought a copy of VariCAD that I use on linux. But while it's certainly a capable CAD, I really don't like the UI, and the parametric modeling is clunky at best.

I don't think there's much more available in terms of parametric modeling on linux if we exclude prohibitive stuff like NX. I'd like to be proved wrong here.

I still think the best modeler I've ever used is still Rhino, a joy to work with. But while it supported parametric surfaces, it didn't really allow parametric modeling when I last used it.

We used BeeGFS since it was called FhGFS on a cluster of 32 nodes for a total capacity ranging from 256tb-1Pb. BeeGFS is very easy to configure and maintain, outperforming Lustre for every task we threw it at. In fact, we never transitioned lustre into production. Lustre was often slower than plain NFS for the cases where all storage could be mounted by a single node(!).

Unfortunately, it's prior licensing scheme prevented larger initial adoption. A good chunk of large scale setups I've seen have adopted Ceph, but I cannot comment on performance.

I'm only using linux, and any company-provided laptop I've been using after 2010, at least a different one every year, was working without any driver issue. This included various models of HP, Dell and Lenovo. I'm currently on a brand new Lenovo Yoga X1, again with absolutely no issue whatsoever.

It sounds like linux breaks, but you know what happens when "stuff on windows breaks"? Well, it's the same: you'll have to fix it.

Ubuntu LTS is really hard to break unless you're constantly screwing around. The same happens on windows.

Don't complain about community-supported distros. Buy a commercially supported distribution with long term releases, and you're set.

The actual game canvas. It seems to switch between bilinear filtering or nearest-neighbor scaling. FF 52 on linux here.

Amazing feat for the PICO-8 :)

I wonder why sometimes the view is blurred, and sometimes it's sharp (looks like some post-processing on the image)?

Any idea on what material the thick support wires mounted on the PCB are made of?

Lou's page is one of the reposts that I'm always happy to look at. And it has been reposted frequently.