Files
Why exact duplicate detection starts with size and ends with a full hash
How ExactDupe narrows candidates by size, hashes full files, and keeps deletion outside the automated result.
First-party technical writing
Field Notes from building the tools.
These are not generic “best tool” articles. Each note documents a concrete design decision, failure mode or verification step from a published WEBBE-B utility. The aim is to explain the behavior behind the interface so a user can decide when the output is trustworthy and when another method is needed.
15 notes · published September 9, 2026. Topics include full-file hashing, folder audits, rename collisions, screenshot seams, image metadata, print dimensions, packing heuristics, schedule cycles, citation extraction and browser puzzle performance.
Files
A folder size map is a prioritization tool, not a delete button
How to read FolderHeat as a prioritization report and combine it with inventory and duplicate verification before cleanup.
Files
Bulk renaming is safe only when the destination names are proven first
How RenameDeck treats preview, collision detection, ordering and rollback as the core of a safe batch rename workflow.
Images
Screenshot stitching fails at the seams, not at the export button
Why long screenshot quality depends on overlap and seam verification, with practical browser-memory and sticky-UI limits.
Images
A sharpness score can rank images without knowing which photograph is better
How sharpness metrics help shortlist near-duplicate photos while leaving focus placement, noise and artistic intent to human review.
Images & privacy
Removing photo metadata changes the file, not the history of where the photo has been
What metadata cleanup can remove from a sharing copy and what remains outside the file, including prior uploads and visible clues.
Images & print
Aspect ratio, pixel dimensions and physical print size solve three different problems
A practical separation of crop ratio, pixel sampling and exact physical print size, including a worked 5 × 7 example.
Workshop
Frame cut calculations are only as good as the measurement chain around them
How frame dimensions interact with rebate, mitre reference points and kerf, and why a dry fit remains essential.
Craft
A fabric nesting result is a good candidate layout, not a proof of global optimum
Why ScrapFit searches multiple heuristic layouts, reports precision and leaves material-specific constraints to the user.
Planning
Repeating shift schedules are modular; real calendars are not
How Overlap Days maps anchored repeating cycles and why real-world exceptions must be verified separately.
Documents
Citation extraction can save search time without becoming legal research
How CiteSweep extracts common case citations from searchable text while leaving authority, context and good-law checking to legal research.
Creative
A 1,000-piece browser puzzle is mostly a rendering and interaction problem
How Puzzle Parcel keeps high-piece-count puzzles practical by treating geometry, memory and interaction as one system.
Images
Local image repair works best when the missing region has a believable neighborhood
Where lightweight browser-local image repair works well, where it breaks down, and how mask size and context affect results.
Games
Crafting calculators are dependency graphs disguised as shopping lists
How PalStack expands nested recipes, combines shared dependencies and subtracts owned inventory while remaining sensitive to game-version changes.
Files
An inventory is the cheapest rollback evidence you can create before reorganising files
Why a simple folder inventory is useful before renaming, moving or deleting, and what it can and cannot prove afterward.
Editorial method
Who, how and why.
WEBBE-B identifies the author, separates first-party implementation notes from external claims, documents AI assistance, and publishes corrections when a tool or explanation changes.
Who
WEBBE-B is created and maintained by Jisung Kim. The About and Contact pages identify the operator rather than hiding behind an anonymous “team.”
How
Notes are based on the behavior, implementation constraints and test cases of published tools. AI may assist drafting or test ideation, but the public text is reviewed before release.
Why
The notes exist because a tool result is more useful when the user understands its assumptions and failure modes. They are written for the work around the tool, not to manufacture search pages.