The first release candidate of FreeCAD 1.0 is out

Original Article ↗ Hacker News Discussion ↗

Highlighted 1.0 RC features

  • Topological naming mitigations seen as the headline change; models are reported to be much more robust under edits, though not perfect.
  • New integrated Assembly workbench with a new solver, replacing the previous fragmented third‑party assembly ecosystem.
  • Major Sketcher upgrades: auto-dimension tool, polar arrays, better array controls, curved slots, offset/scale, automatic midpoint constraints.
  • Part Design gains multi‑solid bodies and the ability to apply features (pad/revolve/pocket) to selected sketch sub-shapes.
  • CAM/Path workbench (renamed CAM in 1.x) reportedly improved and ships with multiple post-processors.
  • UI tweaks: dark themes, quick transparency toggle, optional tabbed workbench bar; useful but considered secondary to core modeling improvements.

User experience, stability, and topology fixes

  • Opinions diverge sharply: some find current dev builds and 0.21.x “usable” and “night and day” better; others still “rage quit” over crashes, strange errors, and brittle models.
  • The topological naming fix is widely described as a game changer for parametric stability, especially for people who previously built large libraries of parts.
  • Persistent pain points: confusing Part vs Part Design workflows, awkward sketch orientation/attachment tools, lack of intuitive drag handles for features, finicky selection (especially on HiDPI), and inconsistent behavior when switching workbenches.

Comparisons with commercial CAD

  • Fusion 360 is praised for polish, CAM, assemblies, tutorials, and robust surfacing/fillets, but criticized for cloud dependence, feature downgrades in the free tier, and flaky export “translation services.”
  • Solidworks, Catia, Creo, and others are seen as more stable and capable for complex and organic geometry, but priced out of reach for many hobbyists; cheaper “maker” or hobby tiers have licensing and cloud caveats.
  • Onshape is repeatedly praised as intuitive and reliable, though concerns are raised about relying on a cloud-only free tier that could change.
  • Some professionals say FreeCAD (or forks like RealThunder’s) is viable for simple to moderate mechanical work but still wastes too much time compared with commercial tools, especially around complex fillets, NURBS, and assemblies (blamed partly on the OCCT kernel).

Code-based vs GUI CAD tools

  • OpenSCAD is popular for parametric, text-based modeling and simple 3D-printed parts, but criticized for:
    • Weak support for complex models (heavy trigonometry, no constraints/STEP, CSG-only).
    • Slow rendering in stable builds (nightlies with Manifold are said to be much faster).
  • Alternatives like CadQuery, Build123D, Replicad, JSCAD, and Python-enabled OpenSCAD variants are discussed:
    • They build on b-rep kernels (often the same as FreeCAD), allowing operations on faces/edges/vertices and STEP export.
    • Trade-offs include heavier dependencies and less “standard” status compared to OpenSCAD.

Learning curve and resources

  • Many find FreeCAD’s interface “opaque” or “GIMP-like”: powerful but hard to approach, especially with many overlapping workbenches.
  • Recommended learning paths:
    • Start in Part Design (sketch + pad/pocket workflow) rather than bouncing between workbenches.
    • Use recent YouTube tutorial series and a free PDF course from a maker magazine.
  • Third‑party variants (e.g., Ondsel, RealThunder’s fork) and UI add-ons are mentioned as significantly improving usability and assembly workflows, though they fragment the ecosystem.

Licensing, pricing, and ecosystem outlook

  • Strong desire for a “KiCad of mechanical CAD”: free, offline, stable, and good enough to choose over proprietary tools.
  • Thread notes that KiCad and Blender only became widely loved after coordinated UX and architectural investments; some see Ondsel’s involvement as a similar turning point for FreeCAD.
  • Many expect FreeCAD to remain behind top commercial tools for complex professional work in the near term, but see clear progress and potential—especially now that topological naming is addressed and a default assembly workbench exists.