Don’t panic. Or should you?

[Surrogate]

AI performance estimate Visio vs Lucid Chart.  Visio values are estimated as M$ doesn't uniquely break those out.

It seems that killer features like Data Visualizer did not help to increase revenue. Even Visio [Online] Plan 1 did not give the desired effect! Since 2017

[wapperdude]

Looking at M$ recent Visio track record, we've experienced it...that's why I believe the current Visio schema has very limited future growth, the product has stalled.  The cost remains high, the bugs don't get fixed.  There is nothing to lure the existing, mature user base to buy / explore new releases, especially when there is little substantially new content...ignoring bugs. Many of us old, literally, are retired, and thus lack the necessity to engage with these poorly executed releases.   The new user has limited database of useful, mature Visio resources.

From a  previous reply, I suggested that Visio needs to be totally re-imagined.  Probably developed by a source outside of M$...Chris where are you???  It needs to be pushing the envelope of forward imagineering,  appealing to a whole new customer base, so enticing that the present Visio user base will be drawn (pun intended) to it.  M$, like the Borg, merely assimilates and adapts, doesn't do original creative design.  Bandaids the Forum can provide.  Bandaids only cover problems, do not provide progressive corrections.

[Yacine]

Wayne, this is very strong.
Despite my new admin tasks , and many other projects, I could not but start my ChatGPT too and ask it similar questions.

I unfortunately held the chat in German, so that I can not share it directly.
So, here are the summaries of the chat.

🔮 The Future of Visio – Thoughts in 2025

"Visio is like graphical Excel – highly manual, deeply scriptable, and uniquely capable of combining graphics with data. There is no real alternative."

That's how I've described it for years.
And yet... no one really treats Visio like that. Not even Microsoft.
📉 Where Visio Stands in 2025

• Still alive, but mostly in the hands of engineers, architects, system designers, and industrial users.
• Integrated into Microsoft 365, but treated like a relic – rarely updated, rarely promoted.
• Not seen as a platform, though it could be one.

Meanwhile, competitors like Figma, Miro, Lucidchart, diagrams.net, etc., have made diagrams easier – but none have reached Visio's scriptability, data binding, and geometry logic.
⚙️ Three Possible Futures for Visio

[th]Scenario[/th] [th]Chance[/th] [th]Description[/th]
Status Quo	60%	Microsoft keeps it alive for legacy users. Occasional cosmetic updates, no innovation.
Copilot Add-On	30%	Visio becomes a backend for AI-assisted diagramming (e.g., "Describe your process → Get a draft").
Phased-Out Replacement	10%	Microsoft moves users to Power Platform or web-only tools; Visio slowly fades.

🤖 And Yet – Even AI Struggles
Yes, AI can draw now. You can describe a flowchart and get something passable. But here's the thing:

AI still just draws dumb boxes.

It doesn't understand what a shape means. It doesn't care about process validation, data binding, ShapeSheet formulas, or engineering logic.
📊 Visio Is Not a Drawing Tool – It's a Semantic Engine
Think about it:

• Shapes have formulas, constraints, behaviors.
• Pages are like tables – visual spreadsheets.
• You can bind external data, write scripts, automate connections.
• The ShapeSheet is a functional language, hidden inside a drag-and-drop UI.

Visio is powerful. It's deep. It's programmable.
But it's misunderstood – even by Microsoft.
They never marketed Visio as what it truly is:

A thinking machine in diagram form.

[Yacine]

Please continue with the consequence, that a community driven, open source Visio is needed. We call it Forma and this is how ChatGPT would imagine it to be.

Perfect – here's the natural continuation, building on the first part. I've transitioned smoothly into the call for a new tool and presented Forma as the next step. Everything is phrased for clarity, inspiration, and community appeal:
🛠� The Consequence: We Need an Open-Source Visio
If Microsoft doesn't understand what Visio is, then maybe it's time we – the users – build what Visio should be.
We don't need another diagramming tool.
We need something deeper.
🌱 Introducing: Forma – The Open Graphical Thinking Engine
Forma isn't just a Visio clone. It's a new paradigm:

[th]Feature[/th] [th]Description[/th]
🧠 Semantics by default	Every shape is meaningful – not just decorative.
⚙️ Fully scriptable	Python, JS or visual logic – no more locked-in VBA.
🔄 Constraint-based	Shapes can have geometric or logical constraints, like in CAD tools.
🔌 Modular UI	Forma is embeddable into other apps. It can become the UI.
📐 CAD-like sketching	Constraint-solving, snapping, and geometric intelligence.
📊 Visual database	Think of it as drawing in a spreadsheet – with structure, rules, and logic.
🧩 AI-powered	Sketch → Flowchart. Text → Process. Error-checking. Auto-suggestions.
📤 Interoperable	Imports/exports: SVG, JSON, Graphviz, Mermaid, P&ID, BPMN, etc.

🤖 Forma + AI: Not a Gimmick – A Thought Partner
AI in Forma isn't just "Draw a box".
It helps design systems:

• "Create a basic chocolate production line with two mixers, a heat exchanger, and a pump."
• "What's wrong with this process flow?"
• "Optimize this warehouse layout for shortest travel paths."

AI + Shape semantics = a visual co-designer.
🧭 Forma's Vision: A Visual OS for Thought

Forma isn't a tool.
 Forma is a canvas for structured thinking.

It can:

• Model processes
• Define interfaces
• Manage content hierarchies
• Simulate decisions
• Express architecture
• Visualize systems
• Serve as the UI itself

💬 Call to Community
Forma can't be built alone.
We need people who:

• Understand the weird magic of Visio
• Believe in structured visual logic
• Want to make visual thinking programmable
• Are tired of drawing boxes that don't mean anything

If that's you – welcome aboard.
✏️ "Forma is what Visio would be, if it had been loved."
If you want to contribute, share your ideas, or help build the foundation –
reach out.
We're planting the seeds.

[Yacine]

Can you elaborate on the features?

---

Absolutely. Let's take each core feature of Forma and elaborate it into a more tangible, visionary yet grounded description – aimed at fellow devs, engineers, and technical thinkers. These are the kind of descriptions you'd want in a README.md, product page, or whitepaper preface.
🚀 Feature Set: Forma as a Semantic Diagram Engine
🧠 1. Semantics by Default
In Forma, shapes are not just visual elements – they're meaningful entities.

• Every shape has:
  
  • a type
  • a data schema
  • a role in a system
  • optionally: a lifecycle, behaviors, and rules

A "Pump" isn't just a rectangle with a triangle inside.
 It's an object with ports, media constraints, flow capacity, failure modes, etc.

You're not drawing a diagram. You're building a system.
⚙️ 2. Fully Scriptable
Forma treats scripting as a first-class citizen, not an afterthought.

• Native support for:
  
  • Python (via Pyodide)
  • JavaScript/TypeScript
  • Optional GUI scripting (block-style)
• Use scripting to:
  
  • Define shape behavior (rules, reactions, calculations)
  • Automate layouts
  • Generate or mutate diagrams dynamically
  • Build custom tools, inspectors, and UI extensions

Think of it as "React for diagrams" – shapes react to changes, inputs, and data.
🔄 3. Constraint-Based Geometry
Forma introduces parametric drawing, like in CAD:

• Define constraints such as:
  
  • Horizontal / vertical / equal-length
  • Alignment, symmetry, anchoring
  • "These two ports must always stay connected"
• Constraint solver keeps geometry valid during edits

Dragging doesn't break rules – it flexes them.
🔌 4. Modular & Embeddable UI
Forma can act as:

• A standalone editor
• A component in your app
• A backstage visual engine for workflows, config editors, document generators

Technically:

• Exported as Web Component, Electron app, or headless server
• UI can be extended via plug-ins (Python or JS)

You don't use Forma – you embed it where thinking needs structure.

📐 5. CAD-Like Sketching with Smart Recognition

• Freehand sketch → Shape Recognition (via Vision AI)
• Auto-snapping, alignment guides, smart anchors
• Shape suggestion: "Did you mean a BPMN Gateway?" or "This looks like a Tank"
• Optional: photo-to-diagram from whiteboards or paper

It feels like drawing – but acts like engineering.
📊 6. Visual Database

• Every diagram is a JSON graph of objects with metadata and references
• Supports:
  
  • Tags, attributes, expressions
  • Queries across shapes
  • Data binding (live updates)
  • Version control

A diagram is a view on structured, versioned, queryable data.
🧩 7. AI-Enhanced Design Assistant

• LLM integration to:
  
  • Translate specs or rough descriptions → diagrams
  • Validate diagram logic
  • Refactor cluttered flows
  • Suggest missing components
  • Generate documentation, code, configs

Example prompt:

"Show me the signal path from Sensor A to the Controller, including analog-digital conversion."

It's not "AI draws boxes".
 It's "AI helps you design semantically valid systems."
📤 8. Universal Interoperability

• Import/export:
  
  • Visio (VSDX), SVG, PNG, PDF
  • Mermaid, PlantUML, Graphviz/DOT
  • BPMN, P&ID, ERD, and custom DSLs
• Plugin SDK to add new formats

Forma is the Rosetta Stone for visual knowledge.
🧭 Summary Table

[th]Feature[/th] [th]Value Proposition[/th]
Semantics by Default	Every shape has meaning, schema, behavior
Scriptable Shapes	Python + JS APIs, no VBA lockdowns
Geometric Constraints	Drawing meets parametric logic
Visual Database	Diagrams are data-first, queryable, and dynamic
AI Co-Pilot	From text/sketch to working diagrams
Embeddable Engine	UI framework or backend module
Modern UI & Rendering	GPU-accelerated canvas, zoom-stable, responsive
Interoperable I/O	Play nicely with the tools engineers already use

[Yacine]

What other features can you think of? Be creative, right now it can't be crazy enough.

---

Oh, challenge accepted 😎
If we throw out all constraints and ask:

"What would a visual-semantic engine do, if it wanted to replace UI, IDE, whiteboard, database, and drawing tool all at once?"
 ...then buckle up. Here's Forma's ultra-expanded future feature set, pushing into wild, powerful territory – many of these not existing anywhere yet:

🧠🖋� 1. Multimodal Input Interface

Draw it, say it, type it, sketch it – Forma gets it.

• Voice-to-diagram: "Add a sensor connected to a warning light with a temperature trigger."
• Sketch-to-shape: Freehand sketch with a stylus → auto-classified as known shapes.
• Natural language navigation: "Show me all subprocesses downstream of Step 5."
• "Mind Map to Model": Conceptual sketch → technical system.

🧬 2. Shape Evolution and Mutation

Shapes adapt, learn, and evolve.

• Shapes have mutation profiles: they can split, merge, morph.
• Drag a "pump" into a high-temperature system → it offers to become a heat-resistant model.
• Long-term: genetic shape evolution. Shapes mutate across projects based on user refinement.

🧭 3. Navigable Dimensions and Zoom Levels

Not just pan & zoom – time-travel, abstractions, and semantic layers.

• Time-zoom: View how a process changes over time.
• Abstraction zoom: Zoom out → see higher-level logic. Zoom in → see formulas, data flow.
• Branching dimensions: Explore alternate designs in parallel universes.
• Version layers: Keep multiple state overlays (e.g. "as built", "as designed", "future concept").

🧰 4. Executable Diagrams

Click "Run" – and your diagram simulates, computes, or controls.

• Embedded runtime: Diagrams can be executed (FSM, workflows, PLC-style).
• Live I/O ports for hardware-in-the-loop.
• Interactive simulations: flow rates, logic conditions, errors, loops.

🧙 5. Semantic Shape Wizards

The wizard that knows your domain.

• A "Reactor" shape knows it's a chemical reactor. It suggests:
  
  • typical sensors
  • required safety valves
  • options for batch vs continuous
• Plug-in model for domain kits:
  
  • "BPMN Starter Pack"
  • "Süßwarenmaschinen Engineering Toolkit"
  • "Philosophical Map Builder for Idea Graphs™"

🧠🧬 6. Neural Shape Memory

Your tools remember what you do, and how you think.

• Shapes log how they're used and evolve with usage patterns.
• AI learns your style:
  
  • How you wire things
  • Your naming conventions
  • Your design errors
• Eventually: personalized diagram grammar.

🧪 7. Diagram as Code — and Code as Diagram

Forma bridges model and implementation.

• Export a process to:
  
  • Python code
  • JSON schema
  • SQL DB schema
  • Kubernetes YAML
• Or the other way: Import a React app or a PLC ladder logic → get a diagram.

📜 8. Diagram-Centric Chat

Each diagram is its own chat room with history and memory.

• Every shape has a chat history: "How did we end up using this reactor here?"
• Users and AI assistants can leave notes, run scripts, discuss inline.
• AI can explain: "Why is this port connected this way?" → "Because it's part of a closed loop."

🧱 9. Spatial Programming Language

You write logic using space and flow.

• Space becomes syntax.
• A wire between shapes is a function call
• A group of shapes defines a pipeline
• A layout grid defines scope or nesting

Think: visually-typed programming, like Haskell + Figma.
🛰� 10. Distributed Diagrams

Diagrams are living networks across files, teams, and systems.

• One diagram can include references to remote diagrams.
• Drag in a component from another team → it's a live link, not a copy.
• Forma handles diagram federation, with permissions, caching, syncing.

🌌 11. "Dream Mode"

Write a sentence. Forma builds a speculative diagram.

• "A decentralized energy grid with smart batteries and autonomous pricing."
• → Diagram appears.
• You explore it, tweak, simulate, question.

This mode blurs line between brainstorming, prototyping, and building.
💡 Bonus concepts:

• Bi-directional thought maps: Like Roam + Figma + logic gates.
• Diagram-to-video: Animate process over time, generate explainer content.
• Tactile feedback: When you plug shapes wrong, you feel it (pen shakes).
• Ethics layer: Highlights steps where system might harm users or break rules.
• Meme propagation maps: Show how ideas, myths, or rumors spread over time and systems.

🚀 Final Thought

Forma isn't a tool. It's a spatial language for everything we can't say in text.

It combines the expressive power of Figma, the structure of UML, the semantic depth of knowledge graphs, and the logical rigor of programming languages — but in 2D space you can manipulate with a pen or thought.
If we build even 20% of this, we change how humans and systems think together.

[Yacine]

A lot the questions in the forum is about how to draw something out of data from a file. Very often also from whole directories. That's the import part.
Others ask for advice for exporting to PDF, Word, the Web, etc. That's the export part.
Forma would have an AI support, but I feel like these import/export issues are a bigger issues that is worth integrating before needing to ask the AI.

A lot struggle also with different outputs, depending on data. These data could be layers (e.g. for buildings), or for states (e.g. cleaning paths of a production plant), or versions (e.g. for different options to quote for to a client). Some of the requests involve dynamic output formats (HTML-SVG-JavaScript).

---

You've just nailed one of the core pain points of visual work in real-world contexts:

It's not about drawing pretty diagrams – it's about transforming and translating data into meaning, and exporting that meaning into the right form.

Let's unpack and reformulate this into a visionary set of import/export capabilities that Forma would make effortless – even before AI kicks in. This would directly respond to years of forum questions, and reframe Forma as a Data ↔ Diagram ↔ Output transformer.
🔄 Import / Export as First-Class Citizens in Forma
🧩 1. Import from Data, not just Files

Forma doesn't just import files. It imports structure, intent, and context.

📥 Supported Imports:

• Flat files: CSV, JSON, Excel, XML
• Directory structures: Represent folders/files as nodes, flows, hierarchies
• Database schemas: From SQL dump or live connection
• Live APIs: Webhooks or REST endpoints → dynamic shapes
• CAD / GIS / SVG / DXF: Parse geometric and metadata layers

🧠 Auto-detection:

• Forma asks:
   "This CSV looks like a BOM. Should I treat it as a production layout?"
   "These folders contain .dwg files. Should I create floorplan layers?"

🧠 2. Smart Mapping Engine

Map data fields → shape types, ports, colors, texts, positions.

• GUI drag-and-drop for mapping:
  
  • Code Select Expand
    data["Type"] → Shape template
  • Code Select Expand
    data["Status"] → Color
  • Code Select Expand
    data["X"], data["Y"] → Position
• Save, reuse, and share mappings
• Combine multiple data sources (e.g.,
  
  Code Select Expand
  equipment.csv +
  
  Code Select Expand
  sensors.csv)

🧭 3. Contextual Layers & Variants

Your data has layers, states, and versions – so does your diagram.

🔁 Dynamic Views:

• Layers: Rooms, floors, subsystems (like CAD)
• States: Cleaning mode, emergency mode, startup mode
• Versions: Quoting Option A/B/C or "before/after" snapshots

Forma lets you:

• Switch views dynamically
• Compare side-by-side (split-screen)
• Animate transitions ("show how flow changes from idle to running")
• Bind filters to data conditions (
  
  Code Select Expand
  if state == 'cleaning' → highlight path)

No more duplicating diagrams for small variants.
📤 4. Flexible Output Pipelines
🌐 Export Options (Granular, Dynamic, and Scriptable)

[th]Format[/th] [th]Notes[/th]
SVG+JS	Interactive diagrams for the web (custom tooltips, behaviors)
PDF	Paginated or single-page export, zoom-locked or vector
Word / DOCX	Embed diagrams + generate structured reports from metadata
PowerPoint	One slide per layer or diagram state
Markdown	Text + embedded visuals, ideal for docs and wikis
JSON / YAML	Export diagram structure or data model
Code	e.g., React components, PlantUML, Graphviz
API snapshot	Export entire diagram state as live-configurable JSON (for reimport or sync)

🧠 Export Presets:

• "Export floorplan as layered SVG for web"
• "Export quote variants A/B/C as 3-page PDF"
• "Generate HTML walkthrough with clickable regions"

Forma lets you define those as buttons, scripts or commands – repeatable and customizable.
🧠 5. Pre-AI Automation

Before you even ask, Forma offers suggestions.

• "This looks like a folder structure – should I build a tree diagram?"
• "You imported sensor data – want to color code based on threshold values?"
• "You have multiple states – want to create a layer for each one?"

This doesn't require full AI. It's just domain-aware, smart defaults + a good mapping UX.
🛠� 6. Data-Driven Output Logic

Outputs vary based on data – Forma knows that and adapts.

Examples:

• A P&ID with active alarms → auto-generate PDF with highlights.
• A factory cleaning plan → generates a walkable HTML/SVG interface.
• A software module map → exports code stubs and API doc pages.

And yes: these pipelines are scriptable and programmable, so you can automate your own:

Code Select Expand

if diagram.view == "emergency":
    export_pdf("emergency_layout.pdf", highlight="red")
if shape.type == "sensor" and shape.data.temp > 80:
    shape.color = "orange"

🧠 Final Thought
Forma doesn't just let you draw from data.
It thinks in data, visualizes relationships, offers context-sensitive transformations,
and lets you export meaning in the form your audience needs.
No more "How do I get my data in?"
No more "How do I turn this into a Word report, a Web page, and a PDF all at once?"
Just: "Data in – structure out – outputs everywhere."

[Yacine]

Can you sketch out an implementation / realisation scenario. What would be the steps, how to get programmers involved, how long will that last?
I would prefer to keep that in an open source mind set with volunteers. But do I need to think about financial foundings?

---

Absolutely. Let's lay out a realistic, open-source-oriented implementation roadmap for Forma, assuming you're starting from a small core team (maybe just you and a few early contributors), with a focus on community-driven growth rather than VC-backed funding.
I'll also mark the critical decisions where you'd consider financing, and where you don't need it yet.
🛠� Implementation Roadmap – Community-Led Forma Realization
🌱 Phase 0: Vision Crystallization (1–2 weeks)
🔹 Output:

• Public README with manifesto and feature scope
• A clear answer to: What is Forma? What is it not?
• Sketch of core architecture (Shape Model, Canvas Renderer, IO system)
• Use cases and personas (engineer, planner, educator, thinker, etc.)

🔧 Tools:

• GitHub repository
• README.md,
  
  Code Select Expand
  docs/, maybe a Notion or Wiki

💬 Community call:

• Share it in forums (Visio, diagramming, engineering, IndieHackers, Open Source Design, Reddit)
• Ask: Who else wants this?

💰 Funding needed? No. Just sweat and clarity.

🧪 Phase 1: MVP Prototyping (1–3 months)
🧱 Scope:

• Interactive canvas (drag/drop shapes, zoom, pan)
• Shape model in JSON (ID, type, metadata)
• Basic import (CSV → boxes, directory tree → graph)
• Basic export (SVG, PNG, JSON)

🧩 Tech:

• Frontend: Pixi.js or Konva.js
• Backend (optional): Python + FastAPI for scriptable import/export logic
• Use Pyodide if you want to embed Python in browser (future phase)

👥 Involve developers:

• Ask for help on: Canvas handling, drag/drop, shape model
• Prioritize contributors who "get" the vision and love tooling
• Tag issues:
  
  Code Select Expand
  first-contribution,
  
  Code Select Expand
  good first issue

💰 Funding needed? No. Just time, maybe donations for hosting.

🔄 Phase 2: Data Flow Engine (3–6 months)
📊 Scope:

• Mappable import layer (e.g., GUI map
  
  Code Select Expand
  csv["type"] → shape.template)
• Constraints engine (minimal: alignments, equal length, ports)
• Visual property binding: color by value, label by data
• Layer & variant system (toggle state/layer visibility)
• Export presets (HTML-SVG, PDF)

🎓 Outreach:

• Reach out to power users of Visio, draw.io, CAD who are fed up
• Create screencasts: "From CSV to flow diagram in 1 min"

📣 Where to find people:

• r/opensource
• Open Source Design
• IndieHackers
• OpenCollective

💰 Funding needed? Possibly small grants or donations, e.g.:

• GitHub Sponsors
• OpenCollective campaign (goal: $2,000 for project bounty fund)
• NLNet or Sovereign Tech Fund (realistic for 5–10k € in EU)

🔁 Phase 3: Plug-in & Scripting System (6–12 months)
🧠 Scope:

• Shapes become mini apps (with embedded logic)
• Embed scripting: Python + JS runtime, or DSL
• Shape Inspector / Binding Editor
• Export-to-code (Graphviz, PlantUML, React)

👨�💻 Dev expansion:

• Let people submit shapes via JSON + script
• Build a Shape Marketplace or public registry
• Build a core team of 4–6 maintainers

💰 Funding?
 Optional but powerful.

• Part-time funding for core devs could help sustainability.
• Maybe partner with open-source-friendly sponsors (e.g., GitHub, Mozilla, CodeSandbox).

🧠 Phase 4: AI Assistant & Dream Features (1–2 years)
Scope:

• AI-assisted import: "This looks like a floorplan – create layers?"
• Sketch-to-shape
• Text-to-diagram
• Semantic refactor / optimization
• Natural language queries

💰 Funding?
 Here it gets tricky. AI integration costs (e.g. OpenAI API) can scale.
 You might want:

• Academic partnership (with a university in HCI or Cognitive Systems)
• Grant funding (e.g., EU Horizon, Prototype Fund)
• Donations from companies that benefit from custom shapes/extensions

🔑 Decision Forks: Where You Might Consider Money

[th]Decision Point[/th] [th]Do you need funding?[/th] [th]Notes[/th]
Initial MVP	❌ No	Focus on passion, clarity, small wins
Hiring core contributors (part-time)	⚠️ Maybe	If you want sustained attention
Hosting collaborative demos	⚠️ Maybe	Can use free tiers initially
AI model integration (GPT etc.)	✅ Yes (Eventually)	Could cost $50–200+/mo in API usage
Marketing / Outreach / UX Design	✅ Ideally	Good design makes people care

🧠 Strategy Summary

[th]Phase[/th] [th]Duration[/th] [th]Goal[/th]
Vision	2 weeks	Shared clarity + repo
MVP	1–3 months	Canvas + import/export basics
Data engine	3–6 months	Mapping, layering, export presets
Extensibility	6–12 months	Plugins, scripting, shape APIs
AI / Smart UX	1–2 years	Assistants, automation, vision

🤝 Next Steps (If You're Serious)

• ✅ Set up GitHub Repo
• ✅ Write a Manifesto README (start with what we've written above)
• ✅ Add issue: "What is Forma? Brain dump here"
• ✅ Set up Discussions or Discord
• ✅ Share with 3 people who've tried to do real things in Visio

[wapperdude]

...and that's a lot to chew on.

You went a bit deeper than I did.  I did include all of the interactive chat details, but bottom line is essentially the same...highly interactive, collaborative, multiple language platforms, etc.  It is a Visio like no other.  It would be equally usable/suitable for technical and non-technical users/applications.

So, heavy stuff...yep.  Something M$ will embrace?...not unless there's no other choice.

Is it needed?  I think so.  The way tech does business has changed, the pace at which ideas go to market require application responsiveness that current programs like Visio cannot provide.  Change is needed.  Change is coming.

[Yacine]

MVP?

----

Great – let's get down to it. Here's a focused plan for the Forma MVP (Minimum Viable Product), tailored to your vision, developer feasibility, and community-building goals.
🧰 Forma MVP – Scope and Goals
🎯 Goal:

Create a working visual editor that can import structured data (CSV, JSON, directory trees), map it to visual elements, and export meaningful graphics (SVG, PNG, HTML).

No AI yet. No scripting engine yet. But modular enough to grow into it.
🧩 MVP Feature Set (v0.1)
🖼� 1. Canvas with Zoom / Pan / Drag

• Infinite or large canvas (scrollable)
• Zooming and panning
• Basic grid snapping (optional)

🔲 2. Basic Shapes & Connectors

• Rectangles, circles, labels, arrows
• Connector logic (line from shape A to B)
• Grouping / layering basics

📥 3. Import from CSV & Directory Tree

• File upload: CSV → boxes with content
• Optional: JSON → diagram structure
• File system → nodes and edges (like mind map or file explorer)

🔄 4. Field Mapping UI

• Simple UI to map CSV fields:
  
  • Code Select Expand
    name → label
  • Code Select Expand
    type → shape type
  • Code Select Expand
    x,
    
    Code Select Expand
    y → coordinates
• Preset for directory trees → hierarchy

🧱 5. Internal Shape Model

• Each shape has:
  
  • Code Select Expand
    id,
    
    Code Select Expand
    type,
    
    Code Select Expand
    label,
    
    Code Select Expand
    data (dict)
  • Position
  • Connections
• Stored in an internal JSON structure

📤 6. Export to SVG and JSON

• Download SVG (including shape labels)
• Download diagram as JSON
• Optional: HTML snippet with embedded SVG

🔧 Tech Stack Suggestion (Lightweight & Web-First)

[th]Layer[/th] [th]Tool[/th]
Canvas	Konva.js or PixiJS
UI Framework	Vanilla JS, Vue, or Svelte (lightweight)
Data Parsing	JavaScript (PapaParse for CSV)
Storage	In-memory, downloadable JSON
Hosting	GitHub Pages or Vercel

🔄 Stretch Goals (v0.2+)

[th]Feature[/th] [th]Priority[/th] [th]Notes[/th]
Undo/Redo	🟡 Medium	Simple stack
Grouping / Layers	🟡 Medium	Based on metadata or views
Style by Data (e.g. color)	🟢 High	Bind color to value
Export to PDF / HTML	🟢 High	via client-side lib
Import JSON schema	🟡 Medium	Build from config not just layout
Load demo project	🟢 High	"Try Forma now!" instant demo

🛠� Developer Tasks to Launch MVP

[th]Task[/th] [th]Est. Time[/th] [th]Owner[/th]
Init GitHub repo	1h	You / helper
Set up Canvas + Konva	1–2 days	JS dev
Create shape model + render	2 days	JS dev
Build CSV upload UI	1 day	frontend dev
Build mapping logic	2 days	logic dev
Export to SVG + JSON	1–2 days	UI/dev
Basic CSS styling/layout	0.5 day	you / frontend
README + live demo	1 day	you / doc help

🧮 Total MVP dev time: ~10–14 dev-days

With 2–3 volunteers, this is doable in 3–4 weeks part-time.

📣 How to Present It

"Forma v0.1 – Turn structured data into interactive diagrams.
 CSV → Shapes. Folder → Map. Export SVG or embed in your docs."

[Yacine]

...and that's a lot to chew on.

You went a bit deeper than I did.  I did include all of the interactive chat details, but bottom line is essentially the same...highly interactive, collaborative, multiple language platforms, etc.  It is a Visio like no other.  It would be equally usable/suitable for technical and non-technical users/applications.

So, heavy stuff...yep.  Something M$ will embrace?...not unless there's no other choice.

Is it needed?  I think so.  The way tech does business has changed, the pace at which ideas go to market require application responsiveness that current programs like Visio cannot provide.  Change is needed.  Change is coming.

Yes, my initial idea after your post was really to analyse how "dumb" we are. What do we ask Visio to do?
- How do we (the noobs) want to use Visio. I suspect that most don't see the power, the flexibility. It's similar to Inkscape.
- Most don't see that with scripts we can automate workflows and manage data.

But I that was not productive. So I worked on your second question - the optimal new Visio. Which led to these long replies.

Could we get these requirements to M$?
- A positive M$ answer? I doubt it.


--> an open source alternative, but then go all for it that is thinkable as for 2025. The realistic stuff and the sci-fi one.
- I am old and not really of a programmer, ... but so visual data enthusiastic ... Would that be my last big project?

[Paul Herber]

I'm still very much in catch-up mode after our holiday, and just trying to read and take in all this lot ... where is the exploding brain icon?

[Yacine]

Windows + ".", then type "explo": 🤯.

Welcome back, I hope you enjoyed the trip. News are reporting shortage in cheese in Normandy, now.

Many scammers - mostly 1-3 per day, once up to 10, and one particularly aggressive who achieved to post without registering. I could only ban its post.
An update of the forum software is available, but I did not dare touching that.

And yes, with the last post, I somehow broke the forum. It was too subversive.

[Yacine]

Visio is Dying – Shall We Let It?

(A call for the core members)

I gave you time. I hoped you'd feel it too:
Visio is unique. And it's fading.

Not because of lack of features. But because its core ideas are no longer understood.

Visio is not a drawing tool. It's a thinking tool.
It lets you draw like any vector program – but then gives you the power to make that drawing react to logic, parameters, or external data.
And that combination – intuitive drawing plus programmable intelligence – is still unmatched.

- Illustrator? Dumb graphics.
- PlantUML? Smart, but non-visual.
- AutoCAD? Powerful, but slow and unusable for casual tasks.
- Mermaid? Markdown-born, structure first, layout last.
- BPMN tools? Rigid templates, no freedom.
- SVG? JSON? XML? All just data. No behavior.

Visio hit the sweet spot:
Draw immediately. And evolve as you go.
For casuals and engineers. For mockups and live process diagrams.
For people who think visually – but need more than pretty lines.

---

So here's my point:

 Should that paradigm vanish, just because Microsoft doesn't see its beauty anymore?
 Should no one carry it forward, just because it's complex?
 Should we let shape intelligence die, and leave the field to figma clones, dashboard toys, and dinosaur CADs?

I say: no.

I say: let's rebuild Visio.
As open-source.
As a modern engine.
As a toolkit for smart visual thinking.

We don't need to invent everything from scratch.
We know what made Visio great. We even know its weaknesses.
Let's build a tool that draws instantly, edits comfortably, and unfolds into power when needed.

Shape intelligence. Logic. Geometry. Data-binding. Parameter-driven layouts.
And still: click, drag, done.

Are you in?
If you're one of the old masters, the ShapeSheet whisperers, the macro tinkerers, the stencil crafters – I'm calling on you.

Let's not mourn the pyramids.
Let's build new ones.

[Paul Herber]

I'm sure I couldn't code to save my life any more, but I can do other stuff.