Free, Libre, and Open Source Software Leftovers
-
The Strategist ☛ China’s use of foreign open-source software, and how to counter it
This poses a dilemma for the US, Australia and its partners. Since open-source software is shared freely and developed collaboratively, China’s efforts to develop local versions forces democracies to decide whether they should allow their own software engineers to contribute to Chinese projects that may end up modernising the country’s military, intelligence and political systems.
China’s pursuit of open-source software started in the 1990s when Gong Ming, the founder of Beijing Ningsi Software (aka Linx Software), transferred copies of the Linux operating system from Finland to China. For that action, Gong is now known as the father of China’s Linux and continues to develop software for the government. This includes software for the Ministry of State Security (MSS), which has been central in shaping Beijing’s policies to build its own open-source ecosystem that it can control.
-
Otávio C ☛ Copy Markdown URL
Steph Ango - Obsidian’s CEO - has created a useful bookmarklet that cleans up URLs for easier sharing. I’ve slightly modified his script to enable copying the link in Markdown format, making it more convenient for use in blog posts, Obsidian, or any other Markdown-compatible editor.
-
Adam Newbold ☛ Neatnik Notes · A Fediverse, if you can keep it
Before I go any further, let me be perfectly clear about this: Meta is an objectively awful company in every possible regard. I don’t like the company, and I am acutely aware of the terrible things that they have done and continue to do.
Many share those same views, and they see blocking/defederating Threads from their own Fediverse instances as a natural response to those views. The discourse around the topic is largely black-and-white, with heavy “you’re either with us or you’re against us” vibes. There’s a lot of passion around this, and much of it is tied to the idea of choices being made to protect the Fediverse. I shared my initial reaction to all of this back when the discussion first flared up. I stand by that post,0 but I have more thoughts now. Two, specifically:
1. The Fediverse wants to be open.
2. The Fediverse is inhabited by people, not companies.
-
Red Hat Official ☛ Preparing for OpenShift Service Mesh 3
There will be some changes coming to the service mesh control plane and deployment topologies. This article will aim to highlight some of these and if you are a current OpenShift Service Mesh user, point out things you can potentially do now to make moving later easier.
-
BSD
-
The Register UK ☛ FreeBSD Foundation gives Beacon gongs for safer software
The Beacon Awards is a fresh scheme from the FreeBSD Foundation, in partnership with the UK government's Digital Security by Design initiative, to reward efforts at safer software.
The Digital Security by Design initiative has been around for some six years now, and it funds multiple projects in the broader security R&D field. The Register reported on Arm jumping on board in early 2019. It worked: It was awarded £36 million ($45.43 million) at the gongs last week. Naturally, there were talks about much more money… but it's good to know that some real technological developments have come out of this.
-
-
SaaS/Back End/Databases
-
PostgreSQL ☛ Swiss PGDay 2024: Registration and Call for Sponsors
We are pleased to announce the call for sponsors and registration for the Swiss PGDay 2024, which will take place on Thursday, 27 June and Friday, 28 June 2024 at the University of Applied Sciences of Eastern Switzerland, Campus Rapperswil (near Zurich).
Registration is now open. We expect the schedule to be published by the end of April.
-
Cohere ☛ Cohere int8 & binary Embeddings - Scale Your Vector Database to Large Datasets
Most vector databases store embeddings and vector indices in memory. Each embedding dimension is typically stored as float32, so an embedding with 1024 dimensions requires 1024 x 4 bytes = 4096 bytes. For 250M embeddings, this results in 954 GB of memory without the ANN vector index.
The most common approach to reduce this huge memory requirement is dimensionality reduction, which performs poorly (see our research, where dimensionality reduction performs the worst).
Instead of reducing the number of dimensions, a better method is to train the model specifically to use fewer bytes per dimension. By using 1 byte per dimension, we reduce the memory 4x (954 GB → 238 GB) while keeping 99.99% of the original search quality. We can go even further, and use just 1 bit per dimension, which reduces the needed memory 32x (954 GB → 30 GB) while keeping 90-98% of the original search quality.
-
-
Education
-
Nate Graham ☛ Free software wisdom
It’s a collection of wisdom written from someone named Lars Wirzenius who started his software development career decades ago and has seen it all. While I don’t have 40 years of programming under my belt, I do have 16 years in programming, QA, release engineering, and management, and everything Lars wrote rings true to me. I’d encourage everyone to give it a read!
-
-
Open Access/Content
-
Society for Scholarly Publishing ☛ Guest Post – Making Sense of Open Access Business Models
Each time a new OA business model is announced, I dig into the description and break the model down, as far as possible, into component pieces. Back in 2020 when I first did this exercise, I found four broad categories of business models and three alternative categories. Thus far, nothing I’ve come across has broken out of those seven boxes, so figure 1 shows how the system breaks down models in use in the market as of February 2024.
-