Thanks for this opening Stuart. Let me take a few minutes of your time
and go through just a bit of the history of EMC project leadership.
Over the years we have discussed at length the need for a definite
structure for the EMC project. Back in The NIST days, Fred, Will, and
Tom provided that structure because it was their project.
As The NIST began to release the project to us by moving the source
repository from Will's engineering computer "spinnerfen" to the
relatively new service at SourceForge we struggled with leadership and
the easiest solution was to lay most all of it at Matt's door.
In a recent post Paul was upset by the fact that the sourceforge project
admins were the ones who assigned privileges to project members, not the
board. Over time the heavy load of source commits moved from Fred and
Will to Paul with targeted commits by several other developers. For
several years, Paul provided much of the guidance or steering of the
software development part of the project.
This changed during a week long "FEST" at the NIST campus in Gathersburg
MD USA. It changed because the assembled group, including all of the
members of the board, decided to move the EMC user space structure over
the top of JohnK's Hardware Abstraction Layer and abandon the older ways
of implementing real-time motion processes. I believe that this
decision, made with the support of the NIST folk was what caused Paul to
fork the EMC project and create his own Linux based motion control
system. Throughout that week Paul repeatedly insisted that no formal
board action was needed. Project leadership changes and project splits
happen quite often in open source projects. And that's okay.
It took nearly a year to complete the HAL transition. We had to abandon
one of the tasks set at that NIST FEST but by the first year of FEST at
Roland's place in Gaithersburg Illinois USA I was comfortable with
saying that we had a working system. John, Alex, SWPadnos and others
were convinced well before I was. As Stewart is quick to point out,
these guys give enormous time to the project and a week of total effort
and their own cash to the FEST. Thanks.
What almost no one knows is that John and I quietly spent a week at
Roland's working on the Mazak project after that first FEST. I
shouldn't say almost no one knew because we were on IRC and Alex,
Steven, and others helped us a lot by making software connections from
their distant locations. I think that it was during that week that both
John and I were convinced that we had the essential parts of a great
system and that we could apply it without hesitation in industrial
settings. We knocked a lot of the edges off a potentially explosive
relationship during that week and saw a lot of the future needs for the
EMC project.
It was about that time a couple of guys from Nebraska USA started
raising a fuss. Chris and Jeff have taken a lot of leadership in the
software and in the help with our customers/consumers. Thanks to you
guys as well.
And so the leadership both in the software development and in consumer
relations changes over time. I think what Todd was trying to say that
in spite of the personnel changes, we must all keep our eyes on the goal
of a system that is as professional as possible and as safe as we can
make it. Notice that I'm trying to avoid the hobby v industrial
distinction. Any software we make must be professional and to be
professional it must accept that there is a profession that it is a part
of.
It is not enough to say let the buyer beware. They can build a safe or
an unsafe machine. I know that buyer beware is not enough because of
the many hours of discussion we have had here, on IRC, and in person
about the nature of estop. Jerry's recent post about the mini graphical
interface starting up with estop released gives weight to this also.
Estop is a whole set of national/international standards. If you place
a machine in the world of production you'd better make damn sure it
complies with the standard where it is.
I'm using Estop here as an example. The early EMC authors had little
idea about the difficult nature of Estop so when someone said, "How are
you handling estop for the GM test?" they simply wrote a software button
to handle it for their half of the experiment. Software writers know
how to write hooks into a running system. Application of the hook was
up to someone else.
The fact that there was a signal out of EMC named estop troubled some
folk because an estop is not supposed to be passed through a silicon
junction. Most estop chains are "anded" rather than "orred" and must
fail open while silicon junctions in low-voltage low-power circuits can
fail closed.
That confusion over the name estop is why iocontrol uses
iocontrol.0.emc-enable-in
iocontrol.0.user-request-enable
iocontrol.0.user-enable-out
Rather than estop-in, estop-out, estop-request. It's also why Jerry
prefers to start a machine by some act of human will rather than by a
software routine saying, "okay I'm alive so go."
IMO this kind of safety in our software is what we must strive for.
That is why I objected to the idea that EMC2 should be able to shut an
open loop machine down and have it remember where it is for the next
start up. What happens when the Sherline lathe is down and junior
twirls a knob or you bump one. The chuck barrier is not where you
thought it was or center of rotation has shifted. Professional machines
do this of course but they do it using absolute position encoders or
always on encoder counters and a fail condition.
I think sometimes that the enthusiasm to add some feature just because
that another CNC program has it pushes safety considerations into the
background.
Thanks for putting up with this long post.
Ray
On Sun, 2007-07-29 at 22:42 -0500, Stuart Stevenson wrote:
> Gentlemen,
> Regarding having a president to guide the corporate development.
> The president of a corporation answers to the board of directors. The
> president implements the policies of the board of directors. The
> president is authorized to spend money to implement the policies. The
> president is authorized to hire and fire employees as necessary to
> implement the policies of the board of directors. That is basically
> the total job of the president of a corporation.
> Is EMC incorporated?
> In what state is the bank account for EMC? 'state' can have more
> than one meaning.
> How much does EMC pay to the developers?
> The preceding two comments are meant to be facetious.
> I attended the workshop in Galesburg this last June. It was very
> educational. I met the five developers in attendance. I encourage
> everyone interested in EMC to attend. The developers were all very
> patient and helpful with anyone and everyone. They acted like the only
> dumb question was the one you didn't ask.
> In talking to them I learned they each paid their own way there.
> They each paid their own expenses for food and lodging. None of them
> were paid for their time during the workshop. They either used
> personal time off or vacation time. When they were finished they
> published their work for anyone to use FREE OF CHARGE. They
> continually spend time, again free of charge, on the internet
> answering questions and offering suggestions to anyone and everyone
> with an interest in EMC. If the development of EMC has a
> personal(hobby) rather than industrial emphasis it is because the
> developers are GIVING their time and energy to the project. They, as
> everyone, will work on the areas most interesting to them. I can't
> fault them for that as I am that way also. I have employees, they have
> a tendency to work most on their area of interest. The result of the
> developers work is a very well written, well thought out product that
> works. I appreciate their effort and work.
> I have one machine using EMC with more coming shortly. This is for
> a production environment. As I progress with these machines I am
> prepared to pay for the development of features not in EMC. Then and
> only then will I be able to say "I want EMC to do 'this'". I talked to
> some of the developers at the workshop about my hiring someone outside
> for development. Their only concern was "would I release it under
> GPL?". If I pay for development I fully intend to release it under
> GPL. They implied they would help me with my project even if I didn't
> release it under GPL but they would much prefer me to release my
> project under GPL. I understand and respect that position.
> I have asked for features to be included. So far the developers
> have included almost everything I have asked for. Sometimes it takes a
> little convincing. The dialog during the convincing is always fun.
> They can ask very pointed questions. Sometimes, it takes a while to
> reach a mutual understanding of the situation or of the proposed
> feature.
> Anyone wishing to contribute would find help from the developers.
> I, too, have a more industrial shop focus instead of personal shop
> focus. I use the term personal instead of hobby. I don't see a
> disconnect or a divergence from industrial and personal. The more
> features in each environment make for a more rounded useful control.
> I suggested to one developer that EMC could stand for Enhanced
> Motion Controller or Enhanced Multi Controller. It could also be
> Everything Multi Controller. I can see this controlling your home,
> shop, car and any industrial or personal process. I may have it
> controlling an anodizing line before too long. Let's see how much
> better and more flexible it can be.
> Maybe I misunderstood the proposed role of a president. My
> comments are my opinion and only my opinion. If I am wrong, talk to
> me.
> In conclusion, an unpaid board of directors and unpaid developers
> will not be very well served by an unpaid president. This is a wide
> open project. Anyone can download it. Anyone can use it. Anyone can
> modify it. Anyone can contribute.
> thanks for listening
> Stuart
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems? Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
