Skip to content
View nazeeh111's full-sized avatar

Block or report nazeeh111

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
nazeeh111/README.md

A mesh, a pixel, a math step

I'm Nazeeh. I like to open an answer back up. Shrink a mesh failure, follow pixels through preserved code, or find the math step that changed the result.

Open a working example: Inspect the missing-body case · Open the native image workbench · Check a math step

Three separate projects: Solidrift's original two-body input and recorded one-body FLOAT output; a pixel crop from Capsulith's prepared image with RGBA 126, 135, 250, 255 at 64,48; and three inked equations from the current Syllune studio. The Solidrift image does not show the later reduced witness. Syllune's hosted checks use typed steps, not handwriting recognition.

Three separate projects. Syllune ink is pictured; its hosted step checker uses typed text. Image sources.

The work behind the objects

  • Solidrift: a Blender Boolean dropped one body. Inspect its recorded reduction from 132 triangles to a 12-triangle witness, then explore the recorded repair and undo. The image above shows the original input and FLOAT output; a new native run needs local Blender.
  • Capsulith: run a prepared ARM64 QOI component in your browser, compare returned RGBA pixels, then restore its native image allocation in a fresh worker. The pictured crop is from its prepared source image.
  • Syllune: write in a private notebook, check supported typed steps, and practice from the first error. The pictured ink is from the current studio; handwriting recognition requires the optional local Mac companion.

Elsewhere, Caustryn exposes a synthetic radio-fit check that accepts a wrong point; Certloom checks stored linear equations independently of its solver; and SolidFrame makes editable browser CAD that exports STEP and STL. Browse the full project index.

Work accepted elsewhere

My Pattern Projector viewport fix was merged across zoom, pan, rotation, reflection, and calibration; projection hardware was not tested. GitHub credited me as Analyst on an advisory correction to Django's package and affected-version mapping (GHSA-wvqv-fj8w-qmhm).

Pinned Loading

  1. Solidrift Solidrift Public

    Shrink broken mesh Booleans into valid reproducers, trace their geometry, and test reversible repairs.

    Python

  2. Syllune Syllune Public

    A private math studio with ink, quiet step suggestions, exact checks, local tutoring and targeted practice.

    JavaScript

  3. SolidFrame SolidFrame Public

    Design plates and brackets in browser CAD; save locally and export STEP/STL. Built on Chili3D.

    TypeScript

  4. Capsulith Capsulith Public

    Run preserved ARM64 code in Node or a browser; inspect decoded image pixels and initialized memory.

    JavaScript

  5. Caustryn Caustryn Public

    Test simulated radio locations against withheld positions; reveal a fit 3.488 m from synthetic truth that still passes its check.

    Python

  6. Certloom Certloom Public

    Check exact error bounds for dense and sparse linear systems in Python or a browser.

    Python