Multiperspective Diagramming – A guide for system engineers | Hacker News Reader

Multiperspective Diagramming – A guide for system engineers

blog.ilograph.com

104 pointsVeuxdo14 comments

This is pretty cool. Reminds me a bit of C4 model [0]. The overall idea that I remember is like "google maps" for software. You can start with a high level view (like "North America") and then keep zooming in to see a neighbourhood view. This corresponds to "context" -> "component" views in C4. IIRC it's all autogenerated, where ilograph I guess is manual. Anyway very cool.

[0]: https://c4model.com/

Looking for a way to deploy a C4 modelling capability was the context in which I first bumped into this product. The last time I looked at C4 it was basically just a concept, I don't know if they have fleshed out a spec or not, but Ilograph can be used to build out something that meets it in spirit at a minimum.

I believe they had some tooling for Java; code annotations or something. Not a generalized solution.

This looks interesting and bears more investigation. One bit of quick feedback: the initial gif/movie is way too fast to follow if you don't really know what you are looking at.

especially with such a small typeface

It's like projections of a graph. Been looking for something like this for so long, wow!

This is very interesting. Do you have a version I can self-host?

Also do you have any plans of adding real-time collaborative editing support?

Hey, thank you. There is no self-hosted version as such, but there is Ilograph Desktop (https://www.ilograph.com/desktop/index.html), which is non-SaaS and works with local diagram files. You can then check your diagrams into e.g. GitHub.

A self-hosted version of what you see on web is probably a 2022 feature, after Confluence support. Real-time collaboration like Google docs is probably the same timeframe. Thanks for this feedback!

I'd be a regular user in a heartbeat if there's a self-hosted version.

I regularly use a self-hosted version of draw.io / diagrams.net for creating what you're calling "multiperspective" diagrams and walkthroughs by hand - basically drawing dozens of variants of a single diagram and embedding them into docs with some text around them. I know I'm not alone, there are probably 100s of potential users in my workplace alone.

This tool can save me days/months of time but the nature of my work is confidential and I cannot use any tools that are hosted elsewhere.

Thanks this, this is great to know. If I can ask, do you self-host as part of an organization, or using your own personal setup? I'm thinking about how authorization, data storage, backups etc. would work for your use-case.

I self-host as as part of an organization with a custom auth that integrates with the orgs SSO. Storage and backups aren't enabled.

I had to pass on it previously because it was hosted, the self-hosting + confluence support for next year will definitely put it in the running.

The other tool mentioned in the blog post is Structurizr, and has a self-hosted tool:

  https://structurizr.com/help/lite

EDIT: The one above is the free tier, they have an on-premises paid version:

  https://structurizr.com/products

Thanks for adding GCP iconset[0] Will wait examples of GCP infrastructure.

[0] https://www.ilograph.com/docs/iconlist.txt

Reply on news.ycombinator.com