Jon Bates

@david_whitney @davidfowl @JamesNK The JVM world has the gradle build system whose definitions are written in Kotlin. It can indeed get a little mind bending

@davidfowl @JamesNK This, but really. Also I can only imagine the horrors I would see with my eyes...

Modernized to xml 🤓 #dotnet

@DalSoft I have spent the last few years building out a platform around Kubernetes, and don't recommend it for a lot of use-cases. I think you're close here: it is an amazing technology but a single project is rarely a good candidate because the required investment isn't trivial

Personally I will not use Kubernetes again, this is the second client project with frictions points because Kubernetes was selected. It adds too much of a complexity overhead for most projects, there are better SaaS, PaaS offerings that keep things simple.

@Dixie3Flatline OH MY CHEESE THIS IS ACTUALLY MAGIC! THANK YOU!!

@CFDevelop I noticed my timeline became really horrendous last week before I realised I accidentally opened "for you" instead of "your timeline". Just aweful

I created a Twitter account for my new company yesterday. The first things I saw were graphic real-life violence, nudity and a bunch of conspiracy theory BS. I've trained the algo well but this place is actually a toilet.

@ICooper @MartinDotNet It was as specific as I could write whilst making an emergency cheese sauce, but yeah, inter-bounded-context is probably links, intra-bc is /probably?/ the same span? This talk could provide valuable insight

@MartinDotNet @jonmbates Exactly why this needs to be done.

Me and @ICooper are planning a joint talk about Messaging, Events and Observability, what would you like to see? The general idea is the "how" but also the "why is it useful" and some of the decisions you'll have with different approaches like links vs parents and clean code

@MartinDotNet @ICooper I'm thinking mainly about useful patterns when hooking up async processes e.g. batches, and maybe even queue processing

@MartinDotNet @ICooper I think links are pretty underserved with respect to documentation. Their how and why would both be useful because most o11y practitioners don't seem to have a grasp on them

Hi @smarticu5 I'm still fretting over your Kubecon talk, so err... thanks? youtube.com/watch?v=f10WQl…

@ThisIsJoFrank Using the first random links as an example of each: Azure's IaC documentation e.g. learn.microsoft.com/en-us/azure/te… And Envoy documentation come to mind e.g. #envoy-v3-api-msg-extensions-filters-http-admission-control-v3-admissioncontrol" target="_blank" rel="nofollow noopener">envoyproxy.io/docs/envoy/lat… They are both pretty exhaustive and link to human-curated examples

Which software/tools that you use has great documentation? What makes the difference to you?

@MartinDotNet (Prometheus) Counters may only go up. If the exposed value goes down, it causes chaos with the metrics. I guess something that goes either way would be a gauge?

The fact that the UpDownCounter in #dotnet only has an "Add" function is something that REALLY annoys me... like.. I know I can add a negative number, but give me a Subtract?

@david_whitney Not sure about this. Standardising on k8s allows us to create common ci/cd for many apps, canary rollouts, alerting and other automations which work on-prem and on-cloud. Still use our cloud's bespoke features, but runtime hosting is less interesting.K8s=CLR; cloud=nuget packages

This is all true. However, *cloud agnosticism* will bind you to only the lowest common denominator features, and, with some irony, the k8s platform itself. An interface and standard without a runtime requirement is better. In that "running containers" > "using k8s".

Oh, it's explained in section 2 🙃

Hi @programmingart, thanks for your http/3 series! Why is the Connection Identifier unencrypted? I'm guessing it won't work well for sticky sessions / sharding because it can change, but I can't think of another use-case

@shanselman The problem arose when we turned them off and tried to turn them back on again...

My wife, a med-surg nurse (I don’t speak medical): we did a CBC, hA1C, and a D-Dimer but the hematocrit was still low Me, a programmer (she doesn’t speak technical): it’s probably DNS

@PR0GRAMMERHUM0R Mate, you need to see the video version; it's in a <marquee>

Hey guys how do I center this <clock>? reddit.com/r/programmerhu…

@sterenas @rbranson @kellabyte Its true that the amount of space taken by a larger varchar is the same, but there still may be costs. E.g. some dB engines will allocate memory when querying according to the max length of a column, regardless of the data actually stored

@rbranson @kellabyte Soooo you’re saying varchar(255) and varchar(5) take the same space? Maaan, TIL something new and am able to inceease lots of fields on my schemas 🤯 Thanks!!!

i asked GPT-4 to for help reducing the disk space of a table, based on it's schema. On #1: this is wrong. reducing below varchar(255) is a no-op. On #2: it recommended smaller int types even though it seemed to understand these were FKs (impressive!) in other answers

@housecor Pretty obvious TBH Network cables heat up as you pass more data; the first records are raw and dangerous, and everything after is too burned to be useful

I've been asked to fix a broken page, inherited from an outside dev team. Heavy use of "any". No tests. I just came across these 2 lines, and I'm scared to move forward. Why did the developer decide to keep only records 550 - 650 from the response?!

@DenisTRUFFAUT @mhevery Bloom filters aren't resizable, and you use multiple hash functions concurrently. You do some math stating how many total items you need to support, how many hash functions you're using and the rate of false positives you can tolerate to get the correct size.

@mhevery Modulo a hash will indeed reduce space but generate tons of collisions. Moreover, like any round robin on a static length, Modulo is terrible when, one day, you want to increase the size of the list. How do you reshard/remodulo 10 billion records ? Looks like a dead end.

Make your code faster with Bloom filters. They are like HashMaps but smaller and allow you to short-circuit more expensive operations. 🧵🪡🧶

@code_is_ @mhevery They both use hashing, but are quite different: * BF are fixed size * They are backed by a bit-array & don't store the input * One value can set multiple bits to true * Multiple values can set the same backing bits true (no collision avoidance) hence probabilistic behaviour

@mhevery checking if the index is non null and you're not traversing the collision DS to check for certain if the key is there?

@david_whitney @EE I use smarty.co.uk - they are actually 3 (not even an mvno), no price rises, no roaming fees and £10 for 30GB Only downside is no Wifi in the tube

Wow @EE's price increase is a touch eye watering this year.