Skip to content

Name Normalization

ixhlbxi edited this page Apr 13, 2026 · 1 revision

Name Normalization

The core of Folder Comparing is its filename normalization engine. Engineering files often include project-specific prefixes and site-specific suffixes that differ between copies of the same deliverable set.

The Problem

The same file might appear as:

Folder Filename
Client copy 25012218-EXC3D_Anna.dwg
Internal copy EXC3D.dwg
Archive copy 25012218-EXC3D_SiteB.dwg

A naive comparison sees three different files. Folder Comparing sees one file (EXC3D.dwg) present in all three folders.

How Normalization Works

Step 1: Prefix Stripping

The tool auto-detects common prefixes across files in each folder. Patterns like project numbers (25012218-) are identified and stripped.

Detection method: If a majority of files in a folder share a leading pattern (digits, hyphens, underscores), it's treated as a project prefix.

Step 2: Suffix Stripping

Site-specific suffixes (e.g., _Anna, _SiteB) appended before the extension are detected and stripped.

Detection method: Trailing segments after the last underscore that vary across folders but attach to otherwise-matching base names are treated as site suffixes.

Step 3: Fuzzy Matching

After normalization, files that still don't match exactly are compared using sequence similarity:

  • Threshold: Configurable (default 80%)
  • Algorithm: Character-level sequence matching
  • Result: Files above the threshold are reported as "Fuzzy Match" with the similarity score

Examples

Original Normalized
25012218-EXC3D_Anna.dwg EXC3D.dwg
PRJ-2024-BOUNDARY.dwg BOUNDARY.dwg
TOPO_SurveyArea1.dwg TOPO.dwg
UTIL-WTR.dwg UTIL-WTR.dwg (no prefix/suffix detected)

Tuning

  • Lower threshold (60–70%): Catches more potential matches but increases false positives
  • Higher threshold (90–100%): Only near-identical names match; good for well-organized projects
  • Default (80%): Good balance for typical engineering naming conventions

Clone this wiki locally