Skip to content

Resources − technical stuff

Jump to bottom
Massimo Tisi edited this page Jan 17, 2025 · 3 revisions

Good practices for visual modeling

  • Use open-source tools or at least open formats
  • Use a well-known modeling language, or define precise semantics for your syntactic elements
  • Put diagram elements as close as possible to each others, to reduce empty space
  • Use consistent rules for capitalization (e.g., all attributes in a class diagram start lower-case, and subwords are separated with CamelCase)

Repositories for public models

Our own "zoo" repository with model examples (This is great for models in the Ecore modeling language): https://github.com/atlanmod/atlantic-zoo/tree/main/AtlantEcore

Some examples of repositories that contain models of different modeling languages:

If you look for public models, the best way is first to decide on one modeling language and then search for models for that modeling language. Other modeling languages that have a lot of public models are:

  • UML
  • Yakindu
  • Asmeta
  • Palladio
  • BPMN
  • SYSML
  • Capella

You can easily find models for these languages, e.g., on GitHub, with a query like: https://github.com/search?p=5&q=extension%3Abpmn+xmi&type=Code

Learning Prompt Engineering

These are some of the most interesting resources to learn about prompt engineering (PE) and a good path to follow.

  1. To start with PE:

    1. OpenAI guide for PE Although the OpenAI/ChatGPT is not the only LLM, it's indeed the most used one, and their tips should work for all other tools, too.
    2. Learn Prompting Course-like website that expands the previous guide in a more profound way, explaining in detail how to construct a prompt.

  2. PE techniques

    1. Lil's Log blog post This blog post from an OpenAI's engineer summarizes everything about PE techniques, including interesting paper references for each one.
    2. Prompting guide This guide includes extended explanations regarding each technique.

  3. Use cases and examples

    1. The Uber Engineering blog has an interesting post on their use of Prompt Engineering techniques
Clone this wiki locally