Re: [rp-ml] STL files

From: Wesley Brooks <>
Date: Wed Dec 06 2006 - 13:50:48 EET

Good morning (here anyhow!) all,

Here's my two pence worth, developed from using RP machines and coding
software for bespoke applications relating to RP machines.

The structure of the STL file is not great, for example there is not a list
of shared points and a list of IDs linking the points to make triangles.

instead of this in STL files each triangle is separate and further
opperations (and asumptions) have to be made to figure out which points and
edges are shared. In addition to single points being listed many times there
is redundant data in that the order of the points is enough to tell the
orientation of the triangle, but normals are also needlessly included in the


Wesley Brooks

On 06/12/06, Delft Spline Systems <> wrote:
> Hi to all,
> As a countermovement to all STL bashing on this list:
> the STL format is the best CAD data exchange format I have ever worked
> with (and I am active in the field for over 20 years).
> Positive points:
> - all 3D CAD programs can export STL
> - no incompatible variations do exist (I do not use color). Compare this
> to DXF, where the specification is changed with every new Autocad release,
> and to IGES, where every CAD system has its own flavour, often incompatible.
> - there is no choice in entities (like a sphere being described OR as a
> center + radius OR as or NURBS surface OR as polygon data OR ...). All
> geometry is described using the same trangle entity. I think this is in fact
> the best point of STL.
> - the specification is so very simple that there is not much room for
> error (it is also inefficient, however large files are not a large problem).
> - it allways works, at least for me: our STL processing software does not
> care about cracks, gaps or orphan surfaces ( I would
> say that refusing to build is a problem of the RP software, not of the STL
> data format.
> The only limitation is that STL can be used for downstream processes only,
> like RP and rendering. Modelling polygon data is not a good option. Since I
> am in model building this does not bother me.
> Best Regards,
> Lex Lennings.
> At 20:11 05-12-2006, Adrian Bowyer wrote:
> >Quoting Todd Pederzani <>:
> >>steve wrote:
> >>>I strongly urge all software vendors
> >>>to support (at a minimum) PLY and COLLADA and to try to wean
> >>>their customers off of STL as fast as they can run away!
> >>You and I may agree, as programmers, that the STL format has
> >>shortcomings, but from a user's perspective it "works."
> >No it doesn't. I've lost count of the number of people coming to me with
> STL files that simply won't build because they do not obey the Euler formula
> for a solid.
> >
> >> What are the
> >>"killer app" features of PLY or COLLADA that would motivate vendors to
> >>implement support for those formats?
> >That you can check it against the Euler-Poincaré formula - simple (and
> fundamental) as that.
> ----------------------------------------------------------------------
> Delft Spline Systems, The Netherlands.
> We offer DeskProto: affordable Rapid Prototyping using CNC milling.
> --- website:
Received on Wed Dec 06 12:42:00 2006

This archive was generated by hypermail 2.1.8 : Tue Jul 21 2009 - 10:27:52 EEST