momothereal

Writing to Kafka allowed them to continue their current ingestion process into MariaDB at the same time as ClickHouse. Kafka consumer groups allow the data to be consumed twice by different consumer pools that have different throughput without introducing bottlenecks.

From experience the Kafka tables in ClickHouse are not stable at a high volumes, and harder to debug when things go sideways. It is also easier to mutate your data before ingestion using Vector's VRL scripting language vs. ClickHouse table views (SQL) when dealing with complex data that needs to be denormalized into a flat table.

The head of the US Copyright Office has since been fired: https://www.cbsnews.com/news/trump-fires-director-of-u-s-cop...

If you go on the page for Manitoba Hydro, it says there is 1 customer in Quebec: "South Cypress". Although South Cypress is a municipality in Manitoba.

Seems like a data issue where the municipality was double-counted and one in Quebec for some reason.

https://poweroutage.com/ca/utility/1408

What is awful about the Helm provider for Terraform?

What happens when the tables and columns have cryptic names/acronyms? Do you need to inject documentation?

They actually have an in-house solution for this: https://analytics.usa.gov/ (though based on Google Analytics)

Details here: https://digital.gov/guides/dap/common-questions-about-dap/

Looks like this is being done under contract/partnership with 18F (GSA): https://18f.gsa.gov/

> The fundamental problem that we’ve observed is that weather.gov reflects its organizational silos (Conway's Law) more than its users’ needs. A lack of overall strategy, feedback/monitoring, and tools have perpetuated this problem.

Wow, my org would not have the guts to write that down in a public README!

I'm working on a tool that does the opposite: to split layers into smaller, deterministic deltas.

Something tells me it will get shortened to "Apple unveils M3" soon

The "Current events" portal is my main source of international news. I love the density of information, link to varied sources, and direct links to Wikipedia articles for background information.

https://en.m.wikipedia.org/wiki/Portal:Current_events

It looks like it just got taken offline

I feel like I used to see a lot more Launch HN, Y Combinator batches, etc. Or maybe I started tuning them out

I was going to ask the same. That is a really important feature to have to replace traditional indexes and usually poorly implemented in vector search libraries.

For example, filtering by arbitrary time range.

No but I'm sure many big orgs to this! Much easier than updating a home page with 1000+ links

My company intranet has a shortcut/URN website that we all configure as "go", so "go home", "go paystubs", etc. Anyone can create URNs. No more stale bookmarks when the HR system keeps changing. Very useful!

The new logo and palette reminds me a lot of the recent ClickHouse facelift, https://clickhouse.com/

> ETL stands for Extract, Load, and Transform of data.

ChatGPT can pretty easily generate documentation from code though :p

And from real-world experience:

- Feature flags should be easy to create or even automatic. Otherwise your team will bypass or forget to create them, or start reusing the same ones until the flag doesn't make sense anymore

- Don't let devs name the flags, or agree and enforce a convention. They should also make sense without context (NOT "test-fix-error")

- Remove flags over time or they will lose all meaning and context

- Categorize your flags to automatically grant them to certain roles/groups

You can have devices in different locations linked together (by opening ports on the WG interface) without opening holes in your network

in the terminal, doing the following freezes the page:

    touch abc
    cat abc

I'm thinking some non-targeted honeypot, given he reached out to the undercover agent directly...

It's up for me, but very slow.

The amount of time spent reading and understanding something should be proportional to how controversial it seems at face value. Half the comments here clearly did not spend more than 30 seconds reading or understanding the changelog.

I hope GH fixes the notification link to PR updates. Right now it brings you to the latest commit, and if you comment on that page, it creates a commit comment instead of a PR review.

Because the right arm of the 'case' keyword is not actually a statement being executed, but its own syntax element to represent a pattern. It is not expecting two sets of brackets there.

I was going to mention government jobs as well. I'm not sure how good the US/state governments are with this, but in Canada the feds are extremely accommodating for any mental and physical illnesses.

That seems weird... is it too hard to feed everything into an Elasticsearch index? Even with default settings and some decent hardware you'd get good stemming and relevance ranking pretty easily.

Legally speaking, the federal government can drop all enforcement across the country. Striking those laws would take a parliament vote however.

The reason provinces need to agree is because healthcare is under their jurisdiction. Parliament and Health Canada will not impose a drastic shift in drug/addiction management on the provinces without working with their health units.

For instance, only the BC government and the Toronto Health Unit have petitioned thus far.

*Only in British Columbia, which is less than 15% of the population.

Also important to note that this has to happen in agreement with both the provincial and federal governments, which is far from guaranteed to succeed in other provinces. See: Alberta.

IMO this is a more detailed article: https://www.cbc.ca/news/politics/opioid-crisis-bc-canada-1.6...

Not sure how widespread this is, but I took taxis in my city recently for work and I was surprised at how much it's gotten better on the UX side (for scheduling):

- When I called, a robot answered and prompted me for my current/origin address (the voice-to-text seemed to work very well)

- It sent me a text message with a link to a live GPS map & time estimate

Still, there are some inconveniences which is why I still prefer Uber for personal travel:

- Even if Uber is more expensive, at least I know the cost upfront, vs. only knowing at the end of the taxi ride

- I get a rough time estimate for the closest Uber before I order one

- No need to handle payment/tipping with the driver, which is the worst because IME taxi drivers always try to make you feel bad about whatever amount you end up tipping them. At least with Uber, I can tip exactly what I feel is 'right' without the awkwardness or over-spending.

- Some taxi drivers are still opposed to using a GPS for directions (??), so when I need to get somewhere that's not a peak tourist attraction, I have to guide them myself.