Re: [CAD_CAM_EDM_DRO] Re: Re: Re: HARDINGE HNC RETROFIT

On Friday 20 September 2002 10:51 am, you wrote:

> Hi Matt,
>
> Is NIST still working on EMC, I thought they kind of abandoned it.
> What is their official word on further development?
>
> Bill

Since I'm not an official representative of NIST, I can't really give you the
"official word" ;) , but I'll tell you what I think:

"No good deed goes unpunished." - Claire Booth Luce (1903-1987)

1. The EMC was originally developed as part of NIST's involvement with the
OMAC (Open Modular Architecture Controls) Users Group. See this document,
pages 19-20:

http://www.arcweb.com/omac/Techdocs/Open_at_GMPTG.pdf

2. As far as I know, NIST needs the EMC to do work in several current
projects such as STEP-NC and some specialized equipment for the Navy. Since
their use of the EMC in research is ongoing, I expect they will continue to
contribute new code and bug fixes generated as a result of this and other
future work.

3. NIST has a longstanding and rigidly enforced policy of not recommending or
endorsing any commercial product, even though in the ordinary course of their
projects they purchase and use many commercial products, if those products
meet their specifications. The work with OMAC required a cnc control that
came with all the source code. At the time (early 90's), such a thing didn't
exist as all controls of that era were at least partially proprietary. For
all I know, if TurboCNC had existed ten years ago, they would've bought it
and based their work on its source code!

4. NIST is also not in the business of competing with commercial enterprises.
Since their work is of the "breaking new ground" variety, and since they have
the means to purchase nearly anything that they require, it's faster and
makes more sense for them to use commercially available products in their
work whenever that's possible. It's also the cheapest way to accomplish their
goals because their effective labor rate is very high (this means that it
costs them a lot of money to reinvent the wheel, so they try never to do it).

5. All the intellectual work of public employees (that's not classified) is
free of copyright and in the public domain where it can be used by anyone for
any purpose, including development and improvement of commercial products.
I've retrofitted and sold a few machine tools using the EMC software, and
there are also quite a few hobbyists who've used the EMC to control their
machine tool projects. I know that several commercial cnc control programs
have incorporated sections of the code (particularly Tom Kramer's g-code
interpreter), thereby saving development time. I still hope that eventually a
major manufacturer of cnc machines will elect to use the EMC as their control
program. I had hoped it would be Bridgeport, but they decided to go out of
business instead...

6. A year or so ago, the main repository of code was moved to Sourceforge so
that the growing number of independent developers could collaborate in the
process of extending and improving the source code. New hardware support was
added, the installation procedure was greatly simplified, documentation was
updated, and the existing user interface was extended (and new GUIs were
added) by non NIST personnel. The bulk of this work was released under the
GPL and the GNU Free Documentation Licence.

7. Not long ago a small software company called NIST to complain that the EMC
was competing (apparently successfully) with their product. The folks at NIST
explained everything in items #1-#6 above, but they still objected to the EMC
web pages at NIST's site (http://www.isd.mel.nist.gov/projects/emc/). When
the company followed up their phone complaint with a letter, the pages had to
be removed. Some of the information on those pages was outdated, and most of
the really useful stuff is available elsewhere ( http://www.linuxcnc.org ).
My picture, taken I think in 1996 during a demo of the first Bridgeport mill
retrofit, was on the software page. I'm sorry it's gone, it was mortar in my
facade of respectability (perhaps more than I really deserved). If you've
never seen these pages (and are morbidly curious), you can just set the
wayback machine to the year 2001, like so:

http://web.archive.org/web/*/http://www.isd.mel.nist.gov/projects/emc/

8. I think the next steps to take are:

A. Write some really good documentation that explains how everything works
in the code. This is really needed to get new developers up to speed in a
reasonable time. I've done some research into software tools that help with
this task, but it's still a time consuming process. I don't really see any
alternative though, and I'm probably the one who'll have to do it...

B. While doing "A", we can also go through the code, make sure it's
formatted nicely, and see if it would benefit from some reorganization.
There've been changes and changes, and changes to changes, and it might be
time for a "spring cleaning". If so, the updated code might be able to be
GPLed if it's done outside of NIST.

C. While doing "B", we should probably try to make the IO system easier to
extend by incorporating something like:

http://mat.sourceforge.net

or

http://membres.lycos.fr/mavati/classicladder/

or both...

Thanks for asking,
Matt

"The lady protests too much, methinks." - WS (Hamlet, Act 3, Scene 2.)

> Well, Linux/EMC is still very rant-worthy: incredibly bloated, still buggy,
> long steep learning curve. Others will say what they will, but the best
> bang/buck ratio IMHO is Indexer-LPT from Ability Systems Corp. It's simple,
> powerful, easy to use. ASC is now offering a G-code interpreter which seems
> to be up to their usual standards, or a good alternative choice for running
> I-LPT would be our own Axis Wizard.
>
> Regards, Hoyt McKagen