digikata

I don't know, it's all probability in the dataset that makes one optimization strategy better over another. Git annex iirc does file level dedupe. That would take care of most of the problem if you're storing binaries that are compressed or encrypted. It's a lot of work to go beyond that, and probably one reason no one has bothered with git yet. But borg and restic both do chunked dedupe I think.

I would think that git would need a parallel storage scheme for binaries. Something that does binary chunking and deduplication between revisions, but keeps the same merkle referencing scheme as everything else.

(2016) https://news.samsung.com/global/samsung-mass-producing-indus...

The size is nothing new, but the packaging into some sort of slottable physical interface with good signal integrity is somewhat new.

You can use hurl in this way. I have projects with a test directory of hurl files, one hurl file per test case. The cases can run one or more http requests. The hurl file can reference external files, capture values from responses for subsequent requests, validate status and outputs. Hurl has various test runner modes and will optionally output overall test results in various parsable formats if you have a larger reporting framework that you would like to hook into.

The Video Game Museum is a smaller museum in Berlin https://www.computerspielemuseum.de/, but it has a lot of fun early history personal gaming equipment, with playable units. And including a lot of euro gaming and (at the time of my visit) running demoscene items I knew little to nothing about.

It really depends on the applicability of what is being trained and how good the training is. In the past, I've signed up companies for training for legal compliance, security, as well as tool/software stacks that everyone will use.

In the tool/sw case, you can do a rough estimation of time saved per attendee after receiving the training. One model of estimation is talking about a little bit of training to to knock off what would be consumed in self-learning time to get past initial stages of the tool. Something that would eat hours without the training, or be a barrier to uptake of something the org thinks it would really benefit from. Do you think that over the span of two years, the training will save say and additional two hours of time for each attendee, or raise usage by x%? Adjust numbers to suit.

In other cases, there is also estimate model that is "fast-forward" training to go very in depth & in specialized areas, but that's typically a payback calc with fewer people, but is weighed against something much longer to pickup on a self-learning arc. If its aligned with getting past specific technical barriers, this can be worth it too.

Could perhaps be fixed by adding a "R" for rust at the end: Faastar

China invested their low interest fiscal capacity into developing over decades via coordinated policy of their central bank / industrial strategy nudges to their businesses. China's factory ecosystem and the ability to build stuff meets my definition of a protective moat. Its very hard to replace and one can only contemplate it over relatively long time frames.

US businesses were free to do the same over the past low interest environment, but we did not have the same incentives, not inclinations in terms of prevailing business strategic appetite for factories. In contrast, for big tech, there was interest and appetite for it and significant capability with protective moats were built - but one could argue that software based tech moats may be faster to bypass.

Maybe, but 0% interest should have make it easy to invest in capital intensive endeavors that would have turned into great protective moats when the interest rates inevitably bumped up. Did that happen with factories and manufacturing?

I also think a significant influence on the Fed was a financialized business community demanding 0% interest.

I think large scale modeling and allocation for "more valuable" has been overly narrow - insufficiently diversified for uncertainty/unknowns, and subtly incorrect for western nations for decades now

Yes climbing the value chain was a necessity for nations like China. But in the US popularized in the 90's, was a business strategy trend that strongly discounted the value of long term capital investments - particularly for this discussion, investment in factories. They do require extra management attention and they do tie companies to strategies in longer time frames at lower margins - but they deliver long term value and long term synergistic growth benefits (in the vein of go slow to go fast). Many US business elected to chase short term growth, and short term and higher margins and minimize long term investments.

See a list of leading US companies that are off of being king of the hill - Boeing, GE, Intel, ... leading industrial US companies continually divested from manufacturing, or shorted long term investment, not because it wasn't profitable, but because it wasn't profitable enough in the moment. It took decades, and many dividends and stock growth was taken in the middle, but the shortfall manifests in time.

I think the mistake here is the model of low-skill/high-skill labor is not a useful distinction. Manufacturing is high skill period, however there are low-infrastructure and high-infrastructure products and factories. The labor wages themselves are a factor, but an increasingly minor factor in product costs. By bypassing investment in US manufacturing skills and infra, the US sat itself on the sidelines for the ability to build, staff, and supply modern low, medium and high infrastructure factories.

It's not impossible to build back, but it would require long term stable policies to favor it at more levels than just tariffs.

IIRC the SQL pipe google paper references PRQL. I think both make sense in their own context. SQL pipe is a smaller leap for users familiar with SQL and prql sort of takes on more change. I do wonder if the smaller leap might cause more deja-vu displacement initially than a bigger shift like PRQL. I don't know the answer other than look at users of both over time and see how they do with the different syntaxes.

We self-host on our own open-source product, but it helps to have a streaming data platform as a product. Not too difficult to write a connector-sink to S3. https://www.fluvio.io/

The caveat is that this is a general purpose data streaming product, so it isn't that opinionated or customized for logging. We still use logging custom frontends for browsing (e.g. grafana), but for core logging/monitoring/billing and critical alerts, it's just Fluvio.

Been looking at the https://coroot.com ebpf collectors to see if we can replace our sidecar sources recently.

Related and also a good read is "The Roald Amundsen Diaries : The South Pole Expedition 1910-1912". You can see the ship he used on the expedition, the Fram at the appropriately named, Fram Museum in Oslo. It's an incredible experience to see and contemplate the expeditions these explorers mounted, and what equipment and resources they assembled to do it at a very early time.

https://www.abebooks.com/products/isbn/9788282350105?cm_sp=b...

For time zone awareness, my go to is https://www.worldtimebuddy.com/

On git commands that accept pathspec, other git adjacent tools also accept path spec. e.g. I tend to use gitk or tig with a pathspec when looking at a bug in a particular file or directory.

Fluvio connectors can access websockets, stream events to a topic whose events can be consumed by multiple clients.

https://www.fluvio.io/connectors/inbound/http/

https://github.com/infinyon/fluvio

The websocket connection was recently added, so if you come up with any suggestions or other feedback on it feel free to share it with the project.

A feature that would have been useful from my IoT / embedded device development days would be the ability to fully turn power on/off of with usb port from software. It's nice feature for automated test and dev of embedded devices. Sort of adjacent to your kvm focus, but I suspect your h/w might not too far off from being able to support that use.

A full A/C plug pwr control is useful too, but a much bigger hardware delta.

One of the reasons I've switched some builds over to musl over glibc, is that I found that glibc linking is brittle if you're going to run a binary over multiple distros in various container environments. Particularly if you want one binary to work on linux across RH and Debian/Ubuntu derived distros or even different ages of distro.

An interesting related discussion to this WebDAV/SMB backed user space FS was the NFS backed user space FS that came up a few months ago on HN https://news.ycombinator.com/item?id=37573679

Deno can use the fluvio crate directly https://crates.io/crates/fluvio.

Eight is excessive. 2-4 is probably the sweeter spot. I was just thinking if you’re going to the trouble of laying out a pcie slot, lay out a full 16 lanes and an expander board will slot 4-8 drives optionally.

Well in general you probably want to be able to hook anywhere from 2-8 drives to to a mini NAS. There seems to be sufficient cpu and Ethernet, but it seems to have no SATA ports, nor does it have say 4-16 lanes of gen3 or 4 Pcie available for either a multiport sata or nvme expander.

PCIe gen2 x1 lane, aww I was hoping that there was an accessible low-power home NAS motherboard.

Many forms of hardware or low-level systems software development and integration. Work with regulatory or security type divides needing separate networks and computers. Sometimes where there can be major dedicated workstations and/or displays for single applications - CAD, Digital art/animation, industrial controls, etc.

I would recommend as a starter the Persistence section of the U Wisc Operating Systems open text book. https://pages.cs.wisc.edu/~remzi/OSTEP/. In particular the cited papers are foundations of major file system approaches of used many (probably most) systems today.

In the non-textbook sources, I would also recommend archives of the Usenix FAST conferences (File and Storage Technologies), where you often see newer and experimental approaches to file systems.

e.g. F2FS, https://www.usenix.org/conference/fast15/technical-sessions/...

In my mental map, yes, but in practice, they act a little differently than my intuition. Even on past non-colima docker usage, I came across surprises w/ `-v` vs `--mount` and so generally try both if I'm having problems.

For my configurations, enabling mounts with '--mount' is more verbose but works fine. https://docs.docker.com/storage/bind-mounts/

It doesn't support that out of the box, but you you can specify or capture values for variables like tokens in the hurl tests.

https://hurl.dev/docs/manual.html#variable

https://hurl.dev/docs/capturing-response.html