Discussion about this post

User's avatar
WhysEnd's avatar

We definitely and urgently need new modeling tools not diagramming tools. The "diagrams" and their elements are to models like letters and words and sentences and paragraphs and chapters are to stories. The actual content manipulated must be a model with the abstractions one actually thinks about. With the most popular tools the only abstractions are boxes and arrows themselves. Not activities, not events, not entities, not control flows, not data flows. The specialised tools e.g. for BPMN and UML and EPC and UPN for activity flows offer notations that were good as starting points in the 90ies but absolutely inacceptable as the end of all efforts. We can and must do better. Everyone is waffling about AI and yes LLMs do help a lot. But a totally overlooked space are high-usability visual graph editors of all sorts: true visual modeling tools. They, too, had their Winter (CASE tools) but the opportunities are great because they offer a path to higher human intelligence and understanding combined with better machine understanding. They offer rather formal specifications while still feeling intuitive. But the tooling is rotten these days and the standard notations actually block innovation. I prefer to do my tax declaration over using Sparx EA and that says a lot. So by all means go on building modeling tools based on boxes and arrows: graphs. There are some folks who aren't visual thinkers but for those that are they can operate and understand way faster and better. My personal approach is #deepmodeling as I see the IT world as a quasi-fractal with depth and a tool like Google Maps/Earth but for editing is the way to go. For all critics of boxes and arrows: go try to manage life without maps of the Earth.

Expand full comment

No posts