Preservation Notes
Overview
The 1990 Web Archive operates as a digital archaeology initiative focused on the systematic capture, verification, and long-term preservation of early World Wide Web content (1990–1999). This document outlines our methodology, technical standards, ethical guidelines, and restoration protocols.
All preservation work adheres to the Authenticity First principle: we never modernize, sanitize, or alter original markup, assets, or behavior unless strictly necessary for safe rendering. Our goal is historical fidelity, not usability optimization.
Core Preservation Principles
01. Structural Integrity
DOM trees, frame configurations, and table layouts are preserved exactly as served. We do not convert to semantic HTML5 or inject modern CSS resets.
02. Asset Fidelity
All images, MIDI files, scripts, and linked resources are captured in their original format. Corrupted or missing assets are marked with structured placeholders.
03. Behavioral Emulation
Deprecated tags (<blink>, <marquee>), client-side includes, and early JavaScript are sandboxed and emulated without altering original intent.
04. Ethical Transparency
We document every intervention. Original headers, timestamps, and server responses are logged. Content warnings are applied only at the viewer layer, never the archive layer.
Technical Methodology
Our preservation pipeline operates in five sequential stages, each verified through cryptographic checksums and metadata logging.
1. Discovery & Seeding
We maintain a curated seed list of 120,000+ verified URLs from 1990–1999, supplemented by researcher submissions, academic datasets, and gap-filling crawls of the Wayback Machine.
2. Protocol-Aware Crawling
Our crawler uses a stack of vintage HTTP/1.0 and HTTP/1.1 agents to accurately negotiate early server behaviors. It respects robots.txt archives from the era, not modern equivalents.
3. Snapshot & Manifest Generation
Each capture produces an immutable tarball containing:
- Original HTTP response headers
- HTML/CSS/JS assets (unmodified)
- Embedded media & linked resources
- Cryptographic manifest (SHA-256 + timestamp)
4. Verification & Indexing
Manifests are cross-verified against known reference dumps. Metadata is extracted and indexed for era-specific search. Broken links are logged but never auto-fixed.
5. Long-Term Storage
Data is replicated across 3 geographically distributed cold-storage facilities using append-only WORM architecture. Refresh cycles run annually to prevent bit-rot.
Supported Formats & Emulation
We preserve and safely render the following legacy formats. Modern browsers are used as viewers, but emulation layers handle deprecated behaviors.
| Category | Formats | Handling |
|---|---|---|
| Markup | HTML 2.0, 3.2, 4.01, Frameset, Tables | Native |
| Styling | CSS Level 1, inline attributes, Netscape extensions | Native |
| Images | GIF87a, GIF89a, JPEG, early PNG | Native |
| Audio/Video | MIDI, RM, early Flash (.swf) | Emulated |
| Scripts | JavaScript 1.0–1.5, VBScript | Sandboxed |
| Plugins | QuickTime, RealPlayer, Shockwave | Stubbed |
Restoration & Verification Protocol
When original content is partially corrupted or inaccessible, we follow a strict restoration hierarchy:
- Primary: Retrieve from our own historical snapshots or verified academic mirrors.
- Secondary: Reconstruct using archived HTTP headers and partial DOM recovery tools.
- Tertiary: Generate structured placeholder with
<!-- RECOVERED: [format] -->comment and visual indicator.
We never replace missing assets with modern alternatives, auto-correct HTML syntax errors, or strip deprecated tags. All interventions are logged in the manifest and visible in the viewer's dev console.
Archive Journal
Selected entries from our preservation logs. Full journal available via API.
Contribute & Report
Researchers, historians, and former webmasters can submit URLs, verify captured pages, or report discrepancies through our open portal.
- Submit a URL: /contribute
- Report an Issue: /issues
- API Documentation: /api
- Contact: preservation@1990webarchive.org
All submissions are reviewed under our ethical guidelines. We prioritize historically significant, culturally representative, and structurally unique content.