WIML: Writing Improvement Markup Language

A Simple Way to Mark Writing Issues

What is WIML?

WIML (Writing Improvement Markup Language) is like SSML for writing style. Instead of just finding typos, it helps identify and fix bigger writing problems like wordiness, unclear passages, and story inconsistencies.

Think of it as a way to track changes for prose quality. You mark the problems, and AI tools fix them.

---

Why WIML?

The Problem

AI writing is so obvious you will likely know this is the first sentence fully authored by a human. Even with sophisticated fine-tuning and prompt engineering, creative writing from an AI can only go so far, it needs a human editor to fix stuff like this:

• “Show don’t tell” issues
• Wordy, confusing sentences
• Character inconsistencies
• Plot holes and logic problems
• Awkward transitions

The Solution

WIML provides a standardized way to mark these issues. It follows the same pattern as heading tags in html.

• Writers can quickly identify problem areas
• AI tools can learn to fix specific types of writing problems
• Editors can provide consistent, actionable feedback

---

How It Works

1. Simple Severity System

Like heading levels in Markdown (#, ##, ###), WIML uses severity levels:

r1 = Critical (breaks the story)
r2 = Major (hurts readability) 
r3 = Moderate (worth fixing)
r4 = Minor (style preference)
r5 = Nitpick (optional)

2. Mark Problem Text

Wrap problematic text in severity tags:

The night was dark and stormy. <r2>Albert was very scared and afraid because he heard weird noises coming from downstairs and it made him feel uncomfortable.</r2> Lightning flashed.

3. AI Processes and Improves

The AI removes tags and improves the marked text:

The night was dark and stormy. Dread coiled in Albert's stomach as distant sounds echoed from below. Lightning flashed.

---

WIML Severity Guide

<r1> Critical Issues

When to use: Problems that break the story
Examples:

• Plot holes
• Character contradictions
• POV violations
• Timeline errors

<r1>Albert's stepmother called him Hunter again, even though his name is Albert.</r1>

<r2> Major Issues

When to use: Problems that hurt readability
Examples:

• Wordy, confusing sentences
• “Telling” instead of “showing”
• Unclear passages
• Repetitive phrasing

<r2>Albert was very scared and afraid because he heard weird noises.</r2>

<r3> Moderate Issues

When to use: Flow and transition problems
Examples:

• Awkward transitions
• Pacing issues
• Repetitive sentence structure

<r3>Then suddenly everything changed completely.</r3>

<r4> Minor Issues

When to use: Style preferences and small improvements
Examples:

• Word choice improvements
• Minor awkwardness
• Dialogue tag refinements

<r4>"I don't know," he said in a confused voice.</r4>

<r5> Nitpicks

When to use: Optional improvements
Examples:

• Could be better but not necessary
• Personal style preferences

<r5>The old, weathered, ancient door creaked.</r5>

---

Optional Type Classification

You can optionally specify what type of problem it is:

<r2 type="style">wordy sentence</r2>
<r1 type="logic">plot contradiction</r1>  
<r3 type="flow">awkward transition</r3>
<r4 type="voice">dialogue issue</r4>

Standard types:

• style – Prose craft (wordiness, weak verbs, clichés)
• flow – Narrative flow (pacing, transitions, clarity)
• voice – Character/dialogue (consistency, authenticity)
• logic – Story logic (plot holes, contradictions)
• ai – AI-specific annotations or artifacts

---

Real-World Example

Before (with WIML markup):

<r1>Hunter</r1> walked into the basement. <r2>The basement was a very scary place that Albert did not like going to because it was dark and spooky and made him feel bad.</r2> <r3>Then he saw the jar.</r3> <r4>"I found it," he said loudly.</r4>

After (AI-processed):

Albert walked into the basement. The narrow stairs descended into impenetrable darkness, each step creaking ominously beneath his feet. The olive jar sat waiting in the corner, its contents shifting with malevolent purpose. "I found it," he whispered.

---

Benefits for Different Users

For Writers

• Quick identification of problem areas
• Consistent feedback across different tools
• Learning tool to improve writing skills

For Developers

• Standardized format for training AI models
• Clear severity levels for prioritization
• Simple syntax easy to parse and process

For Editors

• Actionable feedback with specific issue types
• Priority levels for efficient editing
• Tool compatibility across platforms

---

Getting Started

1. Manual Tagging

Start by manually marking obvious problems in your text:

• Wordy sentences → <r2>
• Plot holes → <r1>
• Awkward transitions → <r3>

2. Tool Integration

Use WIML-compatible tools that can:

• Auto-detect common issues
• Process tagged text with AI
• Generate improvement suggestions

3. Learning and Improvement

As you use WIML:

• Learn to spot patterns in your writing
• Build better writing habits
• Train AI tools on your specific needs

---

Next Steps

1. Get in touch – Github

---

Remember: WIML is designed to be conservative. Only mark text that actually needs improvement. Good writing should be left unmarked.

The goal is better writing through focused, actionable feedback – not overwhelming annotation of every sentence.