· 2030 words · 9 min read · 7 likes on WordPress AI-generated★
Index note huge AI taxonomy (Three.js, VTK.js, GPU prices, AI labs)
Mostly AI-generated: the bulk of this post is AI output that the author prompted, pasted and kept. The author's own words are typically the short prompts and the closing remarks.
Introduction
This document provides a comprehensive, hierarchical categorization of scripting languages, markup systems, CSS frameworks, and 3D visualization technologies, ordered by their relevance, implementation frequency, and production suitability in modern software development. Technologies within each category are sorted from most to least impactful.
1. Scripting Languages (Logic, Application Control)
Rank
Language
Description
1
TypeScript
Strongly typed JavaScript; default for scalable, maintainable frontend logic
2
JavaScript
Core language of the web; still widely used directly
3
Python
Dominant in backend, ML, and scientific computing
4
Lua
Embedded scripting in games and plugins
5
PHP
Used in server-side templating, CMS (e.g., WordPress)
6
Ruby
Clean syntax; niche use in Rails web apps
7
Bash
Used in DevOps, scripts, CI pipelines
8
Perl
Legacy systems and text manipulation
9
R
Used in data visualization and statistical apps
10
Go (Golang)
Efficient backend scripting, real-time APIs
11
Rust (WASM)
High-performance WASM scripting potential
2. Markup, Templating, and Server-Side Rendering
Rank
Language/Tool
Description
1
HTML5
Standard markup, universal
2
JSX / TSX
Abstracted HTML in component frameworks like React
3
Pug (Jade)
Minimalist indentation-based HTML templating
4
Jinja2
Python-based HTML templating (Flask, Django)
5
EJS / Handlebars
Templating in Node/Express apps
6
PHP
Used in legacy and CMS-based templating
7
Razor
ASP.NET’s templating syntax
3. CSS Frameworks, Utilities, and Design Systems
Rank
Framework/System
Description
1
Tailwind CSS
Utility-first, responsive, modern, scalable for design systems
2
UnoCSS
Atomic, on-demand generation, ultra-fast and extensible
Accessible, aesthetic component system based on Tailwind + Radix
6
Bootstrap 5
Classic grid + component system, still popular in admin dashboards
7
Bulma
Modern CSS framework using Flexbox
8
Styled Components
CSS-in-JS, component scoped, dynamic theming
9
Emotion
CSS-in-JS with style props, animations, and theming
10
CSS Modules
Scoped styles with .module.css
11
SCSS / SASS
Traditional preprocessor for structured, nested styling
12
Vanilla CSS
Hand-written styles, fine for static or small apps
4. 3D Visualization Engines and Libraries
Rank
Technology
Description
1
Three.js
Core WebGL engine; used in all serious web-based 3D rendering
2
React Three Fiber
Declarative React binding for Three.js
3
Babylon.js
Full 3D engine; supports physics, PBR, VR/AR
4
Unity WebGL Export
Powerful but heavy; used for games/simulations in the browser
5
A-Frame
Declarative HTML-based 3D/VR; easy for beginners
6
VTK.js
Scientific and medical imaging visualizations
7
XTK
Medical visualization for DICOM, MRI, etc.
8
ShaderToy / GLSL
Custom shaders and raymarching for advanced visual effects
9
CesiumJS
Geospatial 3D maps, terrain, satellite visualization
10
WebGPU
Future replacement of WebGL; currently experimental
5. Volumetric 3D and Scientific Visualization
Rank
Technology
Description
1
VTK.js
Web version of the Visualization Toolkit for volume rendering, CT/MRI
2
XTK.js
Lightweight framework for DICOM, NRRD, medical scan rendering
3
AMI.js
Abstraction layer for medical imaging using Three.js
4
ITK.js
Web adaptation of Insight Segmentation and Registration Toolkit
5
Volume.js
Volume rendering helper for scientific WebGL projects
6
Kitware Paraview Web
Server-backed volumetric visualization toolkit for large scientific data
7
Plotly.js (3D)
High-level 3D volume plotting for data science
8
Raw GLSL Raymarching
Manual implementation of volumetric SDFs and isosurfaces
9
WebXR Layers
Experimental browser-based volumetric rendering for AR/VR
Conclusion
Modern software development involves selecting the right toolchain across multiple domains:
Logic: TypeScript + Python dominate
Markup: JSX + HTML5 are standard
Styling: Tailwind CSS leads, but design systems like Fluent 2 and Material 3 are critical in corporate-grade UIs
3D Rendering: Three.js + R3F for control; Babylon.js or Unity for complexity
Volumetric 3D: VTK.js and ITK.js dominate in medical and scientific domains
Matching the stack to the project’s complexity, performance needs, and target audience is critical for delivering effective, maintainable software in 2025.
—
then which libraries are for visualization, or plotting purposes, CSS itself is styling, but for scientific and imagery data in various forms which are the libraries or packages for that
— Libraries for Visualization and Plotting (Scientific, Image, and Data Forms)
CSS is for styling UI, not data visualization. For real visual rendering of scientific or image-based data, you need specialized JavaScript, Python, or WebGL-based libraries.
Below is a categorized list of core libraries/packages for data visualization, scientific plotting, image processing, and multimodal rendering.
I. General Data Plotting (Charts, Graphs, Dashboards)
Library
Language
Description
Plotly.js / Plotly.py
JS / Python
Interactive line, bar, surface, 3D plots; based on D3 and WebGL
D3.js
JS
Low-level control over SVG and canvas graphics
Chart.js
JS
Simple charts (bar, pie, line); canvas-based
Highcharts
JS
Commercial-grade dashboards, finance charts
Apache ECharts
JS
Advanced interactive charts, maps, graphs
LightningChart
JS / .NET
WebGL-accelerated real-time charts for massive data
Vega / Vega-Lite
JS
Declarative grammar for interactive visualizations
II. Scientific and Statistical Plotting
Library
Language
Description
Matplotlib
Python
Foundation of 2D plotting in Python
Seaborn
Python
Statistical plotting on top of Matplotlib
Bokeh
Python
Interactive plots in browser from Python
Altair
Python
Declarative statistical plotting using Vega-Lite
ggplot2
R
Grammar-of-graphics-based statistical plotting
ROOT
C++ / Py
Particle physics and statistical histogramming from CERN
III. 3D and Volumetric Visualization
Library
Language
Description
Three.js
JS
3D rendering engine for meshes, point clouds, images, and surfaces
Babylon.js
JS
Game-level 3D engine with physics and animation
VTK.js
JS
Volume rendering, scientific visualization, medical imaging
XTK.js
JS
Lightweight toolkit for CT, MRI, DICOM, etc.
ITK.js
JS
Web version of Insight Segmentation Toolkit
AMCharts 5 (3D)
JS
Offers 3D pie/bar charts, not full 3D spatial rendering
Unity WebGL
C#
General 3D game/simulation rendering (when exported to WebGL)
As of 2025, the training of large-scale AI models—especially frontier foundation models (LLMs, vision, robotics, biology, multimodal, agents)—is dominated by a consolidated group of entities across five sectors:
1. Big Tech Corporations (US, China, EU)
These control the infrastructure, capital, and data pipelines necessary to scale gigamodels.
Corporation
Key Models / Domains
Hardware/Infra
Origin
OpenAI
GPT-4o, Codex, Whisper, DALL·E, Sora
Microsoft (Azure)
US
Google DeepMind
Gemini, AlphaFold, AlphaCode, WaveNet, Flamingo
TPU-based (JAX)
UK/US
Anthropic
Claude 3, Constitutional AI
AWS (TPU/A100)
US
Meta AI
LLaMA 2–3, ImageBind, DINO, EGO4D, SAM
FAISS, PyTorch
US
Microsoft Research
Orca, Kosmos, DeBERTa, integration with OpenAI
Azure + GPUs
US
Amazon (AWS AI)
Titan, Bedrock, CodeWhisperer
Trainium/Inferentia
US
NVIDIA Research
Megatron, NeMo, BioNeMo, L4T, diffusion models
GPU Labs (H100)
US
Apple
On-device ML (Siri), Ferret vision-language model
M-series NPUs
US
xAI (Elon Musk)
Grok, plans for Optimus robot agent stack
X servers
US
Baidu
ERNIE, Wenxin/Yi models
Kunlun AI chips
CN
Alibaba DAMO
Tongyi Qianwen, M6, multimodal commerce models
Heterogeneous
CN
Tencent AI Lab
Hunyuan, games + vision + RL
CN GPU farms
CN
Huawei Noah’s Ark
PanGu series, NLP, IoT AI
Ascend AI chips
CN
2. Elite Research Institutions / Labs
Academic-industrial hybrids with access to compute grants, academic talent, and large-scale funding.
Institution
Focus Areas / Models
Location
Stanford CRFM
Alpaca, HELM, RLAIF, language agents
US
MIT CSAIL
Classical ML, robotics, vision, systems
US
UC Berkeley BAIR
VLMs, robotics (RT-X), foundation agents
US
CMU
Embodied AI, language + planning agents
US
University of Toronto
Deep learning theory, vision transformers
Canada
ETH Zurich
Simulation, perception, autonomy
Switzerland
University of Oxford
NLP, RL, causality, alignment
UK
EPFL
Decentralized ML, federated learning
Switzerland
Tsinghua / Peking
Vision, bilingual LLMs, chip-algorithm codesign
China
HKUST
Multimodal agents, game AI
Hong Kong
Sorbonne / INRIA
NLP, theoretical ML
France
Aalto University
Maritime AI, healthcare ML, speech
Finland
3. Private AI Startups (Frontier + Applied)
Often backed by VC or major cloud providers, producing highly competitive frontier or specialized models.
Company
Known For
Origin
Mistral AI
Open-weight LLMs (Mixtral), dense/sparse training
France
Cohere
Command R, retrieval-augmented models
Canada
Adept AI
ACT agents, tool-using foundation models
US
Inflection AI
Pi (emotionally aligned personal assistant)
US
Runway
Gen-2 (text-to-video), diffusion, production tools
Training frontier AI models now requires extreme compute, capital, and data control. Only a dozen or so players can train trillion-parameter scale models. Everyone else uses, distills, or fine-tunes these giants.