Marcel Otto

@marcelotto
63 Followers
108 Following
40 Posts
@georges RTC should actually be a great fit for this! Grouping the triples extracted from a page into a compound and annotating it with provenance metadata (source URL, scraping timestamp, etc.) is pretty much the core use case of RTC.
Ontogen could work too if you want full version history of the dataset, but it still has some limitations. I'm working on a larger update that should address these, hopefully releasing this summer.

@billkiani Totally agree! That’s exactly why Magma is built on top of Obsidian: to leverage an environment where people already manage their knowledge, rather than forcing them into a new silo.

My current stack is pretty focused:
- Obsidian: As the hub for my knowledge base and prompt drafting.
- Magma CLI: For the "heavy lifting" of transclusion resolution and session management.
- Claude Code: As the primary agent for the actual execution.

🚀 Announcing Magma v0.3 – The Prompt IDE for the Coding Agent Era!

Tired of fighting with limited terminal interfaces or simple chat boxes when prompting coding agents? 😫

Magma has been completely redesigned to supercharge your workflow with agents like Claude Code. 🌋

Check it out:
👉 https://github.com/marcelotto/magma

#Magma #LLM #PromptEngineering #Obsidian #Elixir #ClaudeCode #CodingAgents #AI #Productivity #OpenSource

GitHub - marcelotto/magma: An IDE for prompts

An IDE for prompts. Contribute to marcelotto/magma development by creating an account on GitHub.

GitHub
@angelo Congratulations! That's great news! PodOS looks very interesting. I'm happy to see NLnet supporting exciting Linked Data projects this year again. I'm currently working on my RDF and Ontogen projects with NLnet funding as well.

🚀 New Releases! Along with the roadmap update, I'm happy to announce:

- JSON-LD.ex v1.0 with JSON-LD 1.1 support
- RDF.ex v2.1 with rdf:JSON literal support
- Grax v0.6 with a new JSON type and support for ordered lists based on rdf:List

These updates lay groundwork for the DID implementation and future developments. Please refer to the respective CHANGELOGs for a comprehensive list of the changes.

#rdf #jsonld #elixir #myelixirstatus

📣 Project Update: I've published a new article outlining the revised roadmap for Ontogen and related projects in 2025: https://ontogen.io/blog/2025/04/09/roadmap-update-1.html
Project Update: New Roadmap for 2025 and release announcements

The work on Mud, the extracted and enhanced version of Bog, had been progressing well towards solving the problem of static UUID graph names in Ontogen repositories. However, when it came to extending automatic URI generation and management to the resources of the dataset itself, it became increasingly clear that Bog’s strategy alone was insufficient. So, after going back to the drawing board, I’ve developed a new approach. While requiring significantly more effort than initially planned, it promises a comprehensive solution to the URI generation problem which, as described in the Bog article, I consider one of the fundamental challenges of the Semantic Web.

Ontogen

In the last few months, many people have called on the EU to start supporting open alternatives to US-owned social media sites. The thing is: The EU has been doing that for some time already. It has financed, among others, ActivityPub, Mastodon and PeerTube. For @netzpolitik_feed, we spoke to Michiel Leenaars, who professionally gives out buckets of EU cash to open source projects.

#eu #europeanunion #opensource #activitypub #mastodon

https://netzpolitik.org/2025/open-source-funding-on-the-brink-delivering-whats-needed-to-make-europe-sovereign/

Open source funding on the brink: „Delivering what’s needed to make Europe sovereign“

The EU is paying a fortune for broken commercial software, criticizes Michiel Leenaars of the Dutch NLnet Foundation in an interview with netzpolitik.org. Together with his team, he supports open-source, free solutions — funded by the EU. But that could soon come to an end.

netzpolitik.org

@rick_d_freeman As far as I know, no fundamental new solutions have been added to BEAM since then.

However, if you're working in an RDF-based project, the polymorphism solution developed for Grax (based on a pseudo-protocol) can be leveraged for RDFS subclass-based adapters. I'm already using this successfully in several places, for example with the store adapters in Ontogen.

📢 I'm happy to announce that @nlnet has approved follow-up funding for #Ontogen from the @NGIZero Commons Fund 🎉

🗺️ Read about the planned developments in the next year:
https://ontogen.io/blog/2024/08/22/roadmap.html

#semanticweb #linkeddata #rdf #sparql #ngi0

Roadmap for next NLnet funding

I’m pleased to announce that the NLnet foundation has approved follow-up funding for Ontogen from the NGI0 Commons Fund. I’m grateful for NLnet’s trust in this project and their general commitment to advancing open source technologies.

Ontogen
@georges Sorry for taking so long but I didn't have the chance to investigate this issue in more detail earlier. Indeed, there were problems with timeouts in situations like the one you described. These timeout issues have been fixed in the recently released Ontogen v0.1.3.
However, it's important to note that Ontogen is currently not suitable for versioning larger datasets. I've added a note about this limitation to the README.