Artist Website Feature Request List

Or, why I've been using Substack as a portfolio site for four years.

The Moon, at which I shoot.

When I first started building websites for my art in the late 1990s, it was possible to keep the entire LAMP stack in your head, and have room left over for enough CSS and Javascript to get the job done. Those days disappeared in 2010, but I pressed on anyway, crafting an XML-based site driven by CherryPy around 2014. It can do some interesting things. It has ontology tables that encode knowledge such as <uses-medium><such-and-such-painting/><oil-paint/></uses-medium>.

That was great until a data loss wiped out the scripts I used to maintain the site, and I never rewrote them. I can tell you when that happened because it corresponds to a low period of my life and the last time I updated the image gallery, which was 2021. The experience exhausted what was left of my patience for hand-rolling sites. Artblog.net, too, was an XML/CherryPy project. It did everything it was supposed to do perfectly. I was proud of that, because far better programmers than me were defeated by their content management systems. The WordPress blog of Eric S. Raymond does nothing but cough 503 errors at you.¹

But there was no user interface, unless emacs and bash is a kind of user interface. When readers emailed comments, I would retype them. This, like my art site, was unsustainable and causing my audience to dwindle. I needed a brand refresh anyway, so I moved to Substack. Substack supports tags, and soon Dissident Muse Journal became my portfolio site.

Smart programmers are really smart and I don’t pretend to be one of them. But I find the related problems interesting and I have a weird idea of recreation. Last year I quietly completed Hoon School, just in time for Urbit to have some kind of implosion that I still don’t understand and seems to have derailed the project. You may in fact have to be Curtis Yarvin to be productive in a language in which “Hello, World!” is this madness:

|= [=bowl ^] ^- (quip card _this) ~& "Hello, World!"

For that kind of trouble you can just learn Rust, which I did. Don’t ask me to explain the macro system to you, but I built some small, useful programs with it, like the Asynchronous Studio Book Club calendar generator.

I’m now in a tough spot. I’m sufficiently technical to know what’s possible, but not sufficiently technical to execute it. In fact, no one has the entire stack in his head any longer. A modern web service supplier typically has a whole technical team, each, for front end, back end, user experience, and design. The sensible choice would be to migrate to one of the available platforms for artists.

There’s just one problem: they all kind of suck. They don’t suck relative to what most artists are asking for. They suck relative to what they could be asking for. Remember when I solved arts calendaring? That was another Rust project. While I won’t say that nobody cared, I was working on a problem that no one was asking to be solved because arts calendaring works well enough in its janky way to make the investment in change unattractive. Most artists are just asking for a site that looks professional and flatters their work and doesn’t require inordinate technical knowledge (cough). And that is what the providers provide.

I don’t care intensely about the state of arts calendaring, I just needed a project for learning some Rust. My art site is a different story. Putting images up on Substack is convenient, and something that is convenient that exists is preferable to something ideal that doesn’t. But it’s far short of what’s necessary to turn viewers into collectors in a consistent way, which is not to say that it never happens. And compared to what’s possible, it’s nothing. What’s possible is glorious. How glorious? Let me count the ways.

I’ve been working on a feature request list for an ideal site for a few months. It is now quite extensive. Many artists are having perfectly good careers putting their images on attractive-enough display websites and posting their email for contact, and in the meantime, I will do likewise. But I’m hoping that in a year or two I can ask an AI entity to build me a site with the following features, and it will.²

Core Platform Features

Database and Schema

• PostgreSQL database with Schema.org-compliant structure

• JSONB columns for flexible Schema.org property storage

• Integration of the Getty Art & Architecture Thesaurus

• Recursive queries for ontology traversal

• Full-text search capabilities across all content

Content Management

• Support for artworks of any type: plastic, digital, and/or time-based

• Person/artist management with biographical data

• Art criticism article system with rich text editing and HTML source view

• Collection and exhibition management

• Location tracking for artwork provenance

• Flexible tagging and categorization systems

API Layer

• RESTful API built with the Rocket framework

• JSON-LD output following Schema.org standards

• Authentication and authorization system with 2FA

• Rate limiting and API security measures

• Error handling and logging

Image Management Features

International Image Interoperability Framework Integration

• IIIF Image API compliance for high-resolution artwork images

• Dynamic IIIF manifest generation from database

• Support for image regions and annotations

• Deep zoom functionality for detailed artwork examination

• Image comparison tools

• Thumbnail generation with multiple size variants

Image Processing

• Automated tiled image generation

• Batch processing capabilities for existing collections

• Image metadata CRUD

• Color palette analysis and storage

• Automatic thumbnail and preview generation

3D Visualization Features

3D Model Support

• Support for multiple 3D file formats (OBJ, PLY, STL input) and conversion capabilities between them

• Automated conversion to web-optimized glTF format

• 3D model compression and optimization pipeline

• Thumbnail generation for 3D models

• Model validation and quality checking

3D Viewer Capabilities

• Interactive 3D viewer using Three.js

• Orbit controls for 360-degree examination

• Zoom and pan functionality for detailed inspection

• Lighting controls for better model visualization

• Measurement tools for dimensional analysis

• Viewpoint saving and sharing

3D Annotations

• Region-based annotations on 3D models

• Connectivity of 3D regions to ontology concepts

• Scholarly commentary on specific model areas

• Annotation display and management system

AWS Infrastructure Features

Storage and Delivery

• S3 integration for image and 3D model storage

• CloudFront CDN for content delivery

• Automated backup and versioning for all assets

• CORS configuration for web viewer compatibility

Processing Pipeline

• Lambda functions for automated image processing

• Lambda functions for 3D model optimization

• SQS queues for asynchronous processing jobs

• AWS Batch for heavy computational tasks

• Progress tracking and error handling for all processing

Monitoring and Operations

• CloudWatch monitoring

• Cost tracking and optimization recommendations

• Security monitoring and compliance checks

• Performance metrics and optimization insights

• Automated scaling based on demand

User Interface Features

Frontend Build

• Responsive design for desktop, tablet, and mobile

• Modern templating³

Frontend Experience

• Unified interface for 2D and 3D content

• Advanced search with ontology-aware filtering

• Faceted search by material, technique, period, etc.

• User-friendly content management interface

• LLM content management capability

Accessibility and Performance

• WCAG 2.1 accessibility compliance

• Progressive loading for large images and 3D models

• Offline capability for previously viewed content

• Keyboard navigation support for all features

Data Integration

• Schema.org structured data for SEO

• Open Graph Protocol and X Card metadata

Analytics

User Analytics

• Visitor behavior tracking and analysis

• Popular content identification

• Search query analysis and optimization

• Performance bottleneck identification

Content Analytics

• Artwork viewing patterns and preferences

• 3D model interaction heatmaps

• Annotation usage statistics

Customer Relations Management

• Featureful interface for tracking and developing customer base

• Whitelisted, segmented email functionality for announcements and newsletters

• Management of physical mail addresses with export capabilities

Dissident Muse Journal is the blog of Dissident Muse, a publishing and exhibition project by Franklin Einspruch. Content at DMJ is free, but paid subscribers keep it coming. Please consider becoming one yourself, and thank you for reading.

Our next title in the Asynchronous Studio Book Club is Art in America 1945-1970: Writings from the Age of Abstract Expressionism, Pop Art, and Minimalism, edited by Jed Perl. For more information, see the ASBC homepage.

The current exhibition in the Dissident Museum is David Curcio: The Point of the Needle.

1. I have said in the past that hacking WordPress is like organizing Cthulu’s sock drawer. If ESR can’t get his WordPress installation back up, there is something profoundly, abjectly wrong with WordPress.

2. When vibe coding became a thing, I thought, great, when do we get vibe chainsawing? Because if we’re going to fool around with projects in which failure has life-altering consequences, we might as well risk literal amputation. But the consistent pattern of AI is that is starts off comically bad at a given task, and in less time than anyone expects it’s performing better than any human on earth.

3. I may be the only artist on the planet who likes XSLT, but it’s time to let go.