More on STEP

From: Mike Pratt - MIDT (
Date: Fri Feb 13 1998 - 16:26:29 EET

Dear RP-ers,

Let me try to remove some of the mystery about STEP, by responding to
extracts from several recent messages on the subject. Jan Andrzewski
has already done this to some extent, but I will provide pointers to
further information and explain the current situation as far as RP is

> ...There is a new universal translator, STEP, that CAD vendors
> across the board will be promoting for interchanging 3D solid data
> with parametrics.

STEP is an International Standard (ISO 10303) rapidly growing in its
capabilities. The first parts of STEP were published by ISO in 1994,
and all the major CAD vendors either have translators on the market
already or are close to that situation. However, a parametric
capability is not yet available in STEP (I should know - I currently
lead the effort to provide it, and it is not a trivial extension).

> Could Autodesk users out there tell me when their software service
> providers intend to develop a STEP export option.

STEP translators for AutoCAD R14.01 and Mechanical Desktop R2.0 are
due 1st quarter 1998. The source of this information is the PDES Inc.
STEPnet Project:,
this page having last been updated on 3rd February '98.

Autodesk will provide translators with Conformance Classes 1,2,4 and
6, i.e. Configuration Management (versioning etc.), Surface and
Wireframe, Manifold Surfaces with Topology, and Advanced Boundary

> It would help us with one of our projects if ALL CAD users out their
> in RP-ML land could let me know when they intend to have STEP part
> 42 geometric and topological representation export capabilities.

This page also gives corresponding details of the current status of
STEP translators for all the major CAD vendors. See below for an
explanation of Part 42.

The parts of the STEP standard used for the transfer of data are
Application Protocols (APs). Many of these are being developed, for
different product classes and different product life-cycle phases.
So far three have reached the International Standard level:

   AP201 - Explicit Draughting, ISO 10303:201(1994)
   AP202 - Associative Draughting, ISO 10303:202 (1997)
   AP203 - Configuration Controlled Design, ISO 10303:203(1994)

The following AP is at a late stage of development but not yet
formally standardized.

   AP214 - Core Data for Automotive Mechanical Design Processes,
           ISO CD 10303:214.2. This is at its second Committee Draft
           stage, and if approved it will move on to become a Draft
           International Standard.

There is strong pressure from the automotive industry to push AP214
forward. Some implementations of AP214 already exist, based on the
current version; these generally cover areas (such as geometry and
topology) where there is overlap with the capabilities of AP203, the
most widely implemented of the STEP APs already standardized. These
capabilities of AP214 are stable because of the overlap.

The STEP standard has a complex structure. Part 42 is an Integrated
Resource that provides geometry and topology definitions for use in
any implementable part of the standard (AP). In principle all the
different APs are interoperable (i.e. can be used together, to cover
wider ranges of product class, for example) because they are all based
on the same set on Integrated Resources. In practice this is proving
less simple than anticipated, but work is in progress to overcome the

Interest is growing in the use of STEP for data transfer in RP (SFF,
LM or whatever). STEP AP203 already contains a faceted boundary
representation (Conformance Class 5) that includes the STL capability
as a special case, and does it more economically, with explicit
representation of topology that can make slicing more efficient.

At the next meeting of the STEP development community ISO TC184/SC4,
7 - 12 June '98 in Bad Aibling (near Munich), Germany, there will be
an initial meeting of an RP Interest Group that will probably decide
to include RP as a new product life-cycle phase within STEP, make
recommendations for the use of existing STEP APs for RP and develop
appropriate new resources for the future. The impetus for this is
coming from Canada, Japan and the USA.

> STEP has been some kind of a myth during the last years: it should
> be there soon and then solve all geometry exchange problems. I
> have never yet seen it work in practice, however one keeps hoping.

I hope the above will dispell the myth and convince everyone that STEP
is here already. However, the aim of solving `all geometry exchange
problems' is ambitious because of fundamental differences between CAD
systems. Nevertheless, success rates are increasing as experience is
gained in the use of STEP and translators, systems and work practices
are refined. The proportion of successful exchanges (needing no
rework) was well over 80% and still rising in February 1997, according
to the AutoSTEP project (see below).

> Now the above makes me a bit suspicious:
> what is STEP part 42? Does this mean that there are now at least 42
> versions, and if yes are they compatible? Anybody out there having
> experience whether or not it does indeed work ?

I have answered the Part 42 query above. Regarding experience, if you
look at you will find a list
of projects that have been testing STEP data exchange. Several of
these have now moved into production use. Some of the information
given is not very up to date: for example, the AeroSTEP/PowerSTEP
Project has for some time been used in production for the exchange of
geometric data in the critical engine/airframe area by Boeing and its
three engine manufacturers, General Electric, Pratt & Whitney and
Rolls-Royce. The systems concerned are CATIA, Computervision and
Unigraphics, and very substantial financial savings are reported from
the use of STEP technology.

Mike Pratt

| Dr Michael J. Pratt, |
| National Institute of Standards and Technology, |
| Manufacturing Systems Integration Division, |
| Building 220, Room A127, Tel. (301) 975-3951 |
| Gaithersburg, MD 20899-0001, Fax (301) 975-4482 |
| U.S.A. e-mail |
| URL |
| |

For more information about the rp-ml, see

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:44:56 EEST