Popular in Education

Education

The Role of Education in Reducing Racial Inequality

University of Southern California Pedro A. Noguera

Education

The Birmingham Book

University of Birmingham Colin Diamond

Education

Asian Americans and affirmative action: Addressing the color of privilege

University of California Los Angeles Mitchell Chang

Education

Enchanted animism: A matter of care

Murdoch University Jane Merewether

Education

Reader, Come Home

University of California Los Angeles Maryanne Wolf

Education

Teacher Expectations: Impact on Student Academic Outcomes

University of Auckland Christine Rubie-Davies

Education

Putting learners at the heart of the feedback process

UCL Institute of Education Dylan Wiliam

Architectures for Enterprise Integration

View more insights in Professional - Engineering - Educational Technology - Software Engineering

View more figures from Griffith University

View the page of Peter Bernus

Read the Study

Peter Bernus discusses the Architectures for Enterprise Integration, which offers the latest approaches for guiding organizations, consultants, managers, and technical staff through the enterprise development life cycle.

Image courtesy of interviewee. January 17, 2024

Generating Workspace

Export as PDF

This video video outlines the evolution and significance of computer-integrated manufacturing systems, tracing the motivation behind their development back to the 1970s. It discusses the global push to incorporate computers in manufacturing processes through techniques such as numerical control and computer-aided design. The speaker shares personal experiences from working on a complex manufacturing system in Hungary, emphasizing the challenges faced due to overwhelming complexity and costs.

In the late 1980s and early 1990s, a collaborative effort occurred among experts to develop a methodology for creating flexible manufacturing systems, leading to the establishment of reference architectures that guide the design and implementation of such systems. A key milestone was the publication of a book titled "Architectures for Enterprise Integration," which synthesized various proposals and methodologies.

The speaker reflects on the lack of communication between systems engineering and manufacturing control communities despite overlapping interests. Eventually, efforts to bridge this gap resulted in a generalization of architectures and the establishment of international standards.

This video video also emphasizes the importance of integrating human elements within these systems and adapting methodologies to account for the interplay between humans and automation. Moreover, it hints at the rapid advancements in artificial intelligence and their implications for the future of complex system architecture. Overall, it underscores the long-standing history and ongoing evolution of architecting systems in manufacturing and other sectors.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.685 --> 00:04.505
The motivation is something that goes back to the middle

00:04.505 --> 00:09.065
of 1970s where,

00:09.565 --> 00:14.145
around the world there was this idea that we have to,

00:14.445 --> 00:17.645
use computers for, manufacturing,

00:18.775 --> 00:22.485
which include, numeric headache controlled machines

00:23.685 --> 00:25.495
that, actually cut the metal

00:26.715 --> 00:28.515
computerated design and so on.

00:30.725 --> 00:34.225
And, various parts of the world, including,

00:34.805 --> 00:38.425
US Air Force, program,

00:38.765 --> 00:42.345
the integrated computer, in manufacturing program,

00:43.605 --> 00:47.385
and several programs in Europe, have started

00:47.405 --> 00:51.235
to build this exceedingly complex systems.

00:53.325 --> 00:55.705
The way I got into this is,

00:55.985 --> 00:57.825
I was at the time in Hungary in a

00:59.255 --> 01:01.105
very good research r

01:01.105 --> 01:05.065
and d institute, which had 700 engineers,

01:05.095 --> 01:09.545
control engineers, software engineers, plus

01:10.485 --> 01:14.255
support personnel, very well funded.

01:14.675 --> 01:17.175
And we were designing a,

01:17.655 --> 01:19.335
computer integrated manufacturing system

01:19.515 --> 01:20.895
for a machine tool factory.

01:21.465 --> 01:24.215
About 110 people worked for five years.

01:26.155 --> 01:28.655
And, even though that was, you know,

01:28.655 --> 01:31.415
behind the Iron curtain,

01:31.635 --> 01:33.575
the engineers were watching each other,

01:34.365 --> 01:35.575
what is it that they were doing?

01:36.775 --> 01:40.475
And so, from, from,

01:41.415 --> 01:44.595
the Office of Air Force Researcher,

01:45.125 --> 01:49.655
which was actually at the UK at the time, award

01:50.365 --> 01:51.895
came several times to visit.

01:52.755 --> 01:54.895
We went to visit as well, us

01:55.125 --> 01:57.415
because engineers talked to the engineers,

01:57.795 --> 01:59.215
how do we solve this problem?

01:59.955 --> 02:03.855
And, when I was recruited as a relatively young,

02:03.865 --> 02:07.605
young researcher, they were asking me, come

02:07.605 --> 02:09.165
and work for us, come and work for us,

02:09.165 --> 02:12.045
because I had worked for them for a few years before.

02:13.305 --> 02:16.285
And I said, what is your most difficult problem?

02:17.745 --> 02:22.125
And it is, well, we have a project, 110 engineers

02:23.025 --> 02:26.435
working on this exceedingly complex system.

02:27.945 --> 02:31.315
Give us a methodology how to do that,

02:31.315 --> 02:33.995
because the complexity was just overwhelming.

02:36.245 --> 02:40.125
And so during the past five years,

02:40.525 --> 02:42.605
I was heading a system engineer group,

02:42.625 --> 02:46.525
and we developed a methodology system was implemented

02:46.545 --> 02:49.925
and so on, but it was too expensive.

02:50.705 --> 02:51.805
It was too expensive.

02:52.105 --> 02:55.765
Uh, in the us they were using it for, aircraft being,

02:56.545 --> 02:58.805
production we

02:58.805 --> 03:02.525
Were using for product producing other types of parts.

03:03.505 --> 03:07.365
So how would it be possible to make it more affordable?

03:07.745 --> 03:10.645
The technology is fantastic, but it's just not affordable.

03:10.905 --> 03:13.485
The return investment is just negative.

03:15.515 --> 03:16.975
And of course, many people

03:16.995 --> 03:20.215
around the world are working on this, including Germany,

03:21.175 --> 03:24.025
including France, including the us.

03:25.715 --> 03:30.615
And, by that time, my group developed a method how

03:30.615 --> 03:34.575
to develop a reference model, if you wish, for

03:35.095 --> 03:39.175
a class of flexible manufacturing systems so

03:39.175 --> 03:41.895
that it can be implemented in a reasonable,

03:41.915 --> 03:43.175
at a reasonable expense.

03:45.735 --> 03:49.905
So, Jim Nvi, who at that time

03:51.275 --> 03:54.495
was the deputy director of the Charles Child, Dr.

03:54.985 --> 03:59.665
Laboratory in, I don't know,

03:59.845 --> 04:03.745
Draper lab is a part or not part is independent at MRT.

04:04.645 --> 04:09.035
Um, but he raised the question,

04:10.095 --> 04:12.915
how would it be possible to bring together all the experts

04:13.015 --> 04:14.995
who work in this area from around the world?

04:16.275 --> 04:20.375
Jim was by the way, the, one of the main designers

04:20.395 --> 04:21.975
of the Apollo guidance systems.

04:22.875 --> 04:26.135
Uh, so he was a very, engineer,

04:27.335 --> 04:31.475
but he said, well bring the manufacturing control people

04:32.295 --> 04:36.155
and the, and the IT or computer science people together.

04:36.945 --> 04:40.315
What is it that we know how to create such complex systems?

04:41.335 --> 04:43.595
And so I was one of those experts invited,

04:44.675 --> 04:47.325
and there were others as well,

04:48.485 --> 04:50.255
many others as a matter of fact.

04:50.395 --> 04:54.135
Uh, and we started to review what is it

04:54.135 --> 04:58.255
that the different companies say how to create such systems?

04:59.195 --> 05:02.855
And as a result in the first, that was around 1990,

05:04.375 --> 05:08.715
as a result, we started to categorize what is it

05:08.715 --> 05:10.715
that people say, how to build this?

05:11.415 --> 05:14.835
And there were two types of proposals.

05:15.135 --> 05:17.715
One type, this is the structure

05:18.495 --> 05:21.755
of a manufacturing system,

05:21.895 --> 05:23.755
and including its control system.

05:24.625 --> 05:25.635
Well, that's fantastic.

05:26.615 --> 05:29.435
It gives you a pattern, but it doesn't give you a

05:29.435 --> 05:30.795
methodology how to do this.

05:32.775 --> 05:36.315
And so the type two, so-called reference architectures

05:37.495 --> 05:41.155
should be those which actually tell you how

05:41.175 --> 05:44.115
to get from the idea idea ideation,

05:44.225 --> 05:46.595
that you want a flexible manufacturing system

05:47.015 --> 05:49.515
to actually design, architecting, designing,

05:50.155 --> 05:51.675
implementing, operating it.

05:52.395 --> 05:53.575
That's another type

05:53.575 --> 05:56.495
of reference architecture relative to the ones

05:56.685 --> 05:59.085
That just said, this is how this should look like.

06:01.195 --> 06:03.735
And so this task force started to review

06:04.465 --> 06:06.855
these various architecture proposals.

06:07.515 --> 06:11.215
And the outcome of that was the book that we published

06:11.835 --> 06:14.935
called Architectures for Enterprise Integration

06:15.675 --> 06:17.055
is not for the faint hat.

06:17.115 --> 06:18.895
It is more like a monograph.

06:19.955 --> 06:23.535
Um, it is not a, not for teaching purposes.

06:23.605 --> 06:26.655
It's a monograph that documents the thinking

06:26.655 --> 06:28.535
of the task force and how to go ahead.

06:29.115 --> 06:33.915
One of the, at the time they made, the task force

06:34.895 --> 06:37.875
ifi, international Federation of Information Processing

06:38.255 --> 06:41.355
and iact, international Federation of Automatic Control.

06:42.325 --> 06:45.035
These two organizations came together

06:45.655 --> 06:48.155
and created this task force with many members.

06:49.175 --> 06:52.755
Um, most of them were from industry,

06:52.895 --> 06:55.035
but some of them for from academia

06:55.735 --> 06:58.475
or something in between an r and d institute.

06:58.535 --> 07:00.955
You can argue whether it's academia or not.

07:01.455 --> 07:03.235
And when we did this review

07:04.415 --> 07:08.075
before publishing the book that you were referring to, Fran

07:08.935 --> 07:12.315
was, you know, one of the people

07:12.975 --> 07:16.435
who said in the task force, so how it possible to generalize

07:16.945 --> 07:21.035
what people who are in this area of architecture, how

07:21.035 --> 07:23.555
to create flexible manufacturing systems.

07:24.335 --> 07:28.035
And the chairman of the task force was, Ted Williams,

07:28.895 --> 07:31.075
the chair of the Purview Consortium,

07:31.245 --> 07:33.555
which was originally created by Axon,

07:34.615 --> 07:38.835
but populated by many large engineering companies

07:40.015 --> 07:44.235
and, and control system vendors and so on.

07:44.655 --> 07:47.115
And I said, you know, I was relatively young

07:47.475 --> 07:48.635
researcher in my forties.

07:49.455 --> 07:51.875
Uh, I put up my head, actually, I know how

07:51.875 --> 07:53.155
to generalize all of this.

07:53.155 --> 07:54.635
They said there said, do it.

07:56.195 --> 07:58.295
So I did it with a colleague of mine

07:58.835 --> 08:03.095
and published, I think it's 1994, an article

08:03.835 --> 08:08.345
how to generalize This architecture Approaches,

08:08.595 --> 08:12.305
which give you a methodology for very large complex systems.

08:14.305 --> 08:17.005
And, the book that you're referring to,

08:17.185 --> 08:20.965
it took quite some time to actually publish, you know,

08:21.265 --> 08:24.565
it takes almost two years from the time you have the idea

08:24.665 --> 08:26.925
to actually publish a book about it.

08:27.145 --> 08:30.965
And then when this was presented, they said, Ari,

08:31.185 --> 08:32.765
now you are the chair of the task force.

08:33.195 --> 08:37.325
Because one person from the National Institute of Standards,

08:37.325 --> 08:41.745
techn in Burg said, dear,

08:43.285 --> 08:47.145
we have been going around in circle in ISO International

08:47.145 --> 08:49.225
Organization Standards Organization

08:49.645 --> 08:51.345
to actually develop such a thing.

08:53.225 --> 08:55.165
It would be possible to fast track

08:55.665 --> 08:57.405
as an international standards.

08:58.025 --> 09:00.285
And within six years, we fast tracked

09:00.425 --> 09:03.885
and became an international standard in the year 2000.

09:03.905 --> 09:06.285
The problem is that by that time

09:06.385 --> 09:10.565
and many others including the us, the, department

09:10.565 --> 09:14.245
of Defense developed an architecture framework

09:15.665 --> 09:19.495
called Dodo, which was,

09:20.755 --> 09:22.385
which had a similar intent.

09:24.405 --> 09:26.945
The British, of course, they improved it.

09:26.975 --> 09:31.765
They called modd and Canadians called it donoff.

09:33.025 --> 09:37.445
The Australians got it all stuff, and so on.

09:38.075 --> 09:39.205
That was just one part.

09:39.705 --> 09:43.245
But it was all manufacturing and control systems people.

09:44.315 --> 09:46.285
They had no idea what the sources

09:47.125 --> 09:48.125
software engineers were doing.

09:48.765 --> 09:52.205
Absolutely no. And the ignorance was mutual

09:54.595 --> 09:56.075
software engineers had no idea

09:57.135 --> 10:01.625
that we had this fantastic methodology, which would've been

10:02.345 --> 10:04.065
a systems engineering methodology.

10:04.165 --> 10:06.785
Really, the only problem is

10:06.785 --> 10:09.785
that a systems engineering methodology usually looks at

10:09.785 --> 10:14.745
technical systems where a manufacturing system has humans in

10:14.745 --> 10:17.225
it, and humans are always the problem

10:17.485 --> 10:18.905
and often also the solution.

10:19.285 --> 10:22.105
And our architecture framework, which was,

10:22.105 --> 10:26.085
which which we called is generalization, generalized,

10:26.415 --> 10:28.205
enterprise reference architecture

10:28.205 --> 10:30.805
and methodology covered it all.

10:31.185 --> 10:34.045
But where the ingredients came from

10:34.635 --> 10:38.365
were from these different European and other projects.

10:39.365 --> 10:42.865
So for example, the Perdu Enterprise reference architecture

10:43.885 --> 10:48.255
very eloquently, described how

10:48.395 --> 10:51.695
to divide the system into

10:51.885 --> 10:55.755
what the system actually does perform its mission

10:56.895 --> 10:59.515
and its control management and control.

11:00.575 --> 11:01.995
And these are the only two things

11:01.995 --> 11:03.155
that you have to worry about.

11:04.695 --> 11:08.155
But when you perform the mission, there are certain things

11:08.155 --> 11:09.235
that you automate

11:10.055 --> 11:12.195
and certain things that are being done by humans.

11:12.895 --> 11:16.675
And the same is true for the management and control.

11:16.935 --> 11:20.155
The Europeans computer integrated manufacturing,

11:20.305 --> 11:22.475
open systems architecture, mosa,

11:22.545 --> 11:26.075
they brought in the lifecycle concept, which says, at

11:26.075 --> 11:28.155
that time in a very,

11:29.615 --> 11:33.035
by now I would say rudimentary way, you have

11:33.035 --> 11:35.075
to define the requirement, the design

11:35.095 --> 11:36.875
and implementation of a system.

11:38.685 --> 11:39.985
We have refined that later.

11:42.005 --> 11:45.425
And the French, the French are always a bit outliers

11:45.565 --> 11:48.585
and say very interesting things,

11:48.705 --> 11:51.505
because the French have a different idea,

11:52.455 --> 11:54.525
they look at the world in a different way.

11:55.105 --> 11:57.165
For example, they don't have the word for design.

11:58.395 --> 12:00.285
They insist that they do it, but they don't.

12:00.465 --> 12:01.965
So when you ask what,

12:02.105 --> 12:04.245
how do you translate computer related design?

12:04.355 --> 12:05.925
They say co

12:09.255 --> 12:14.005
conception is not the design is design concept

12:14.705 --> 12:16.125
is come up with the concept.

12:16.505 --> 12:19.355
So to this day, the French

12:19.415 --> 12:21.795
and English engineers don't a hundred

12:21.795 --> 12:23.235
percent understand each other.

12:23.585 --> 12:27.795
Anyways, they brought a very interesting, structure to

12:28.255 --> 12:30.475
how to think about the management system

12:32.365 --> 12:35.105
of an enterprise, putting it all this together.

12:35.365 --> 12:38.875
That's g So now we had something

12:40.065 --> 12:41.225
actually in the year 2000

12:41.245 --> 12:44.825
and I, I presented it, in Toulu, we,

12:44.825 --> 12:47.025
the large ISO meeting.

12:48.565 --> 12:50.505
And they said, dear, dear, there,

12:50.925 --> 12:52.265
we should have known this before

12:52.615 --> 12:55.745
because we have been working on this anyways, that

12:55.745 --> 12:57.065
with the systems engineers.

12:57.565 --> 13:01.465
And, anyways, jumped 23 years forward

13:01.645 --> 13:04.625
or 20 something years forward, I realized

13:04.775 --> 13:08.025
that these two communities, the systems engineering

13:08.165 --> 13:10.505
and the manufacturing and control communities,

13:10.855 --> 13:12.545
they just don't talk to one another.

13:13.845 --> 13:16.505
So the only way is to,

13:17.505 --> 13:20.265
I was in the Manufac manufacturing and control world.

13:21.505 --> 13:25.805
I moved into this world and joined the systems

13:25.985 --> 13:27.805
and software engineering community.

13:29.935 --> 13:33.355
And so when you jump forward, we can see

13:34.665 --> 13:39.295
that now what was that, concept in the,

13:39.295 --> 13:44.015
manufacturing world has now infiltrated in pretty much all

13:45.415 --> 13:47.485
major systems engineering standard.

13:48.145 --> 13:51.325
It just took a cool 23 years for that to happen.

13:51.905 --> 13:56.215
One thing though, that is a kosher, I've been talking to,

13:57.425 --> 14:00.805
to people for who are doing enterprise architecture,

14:01.375 --> 14:05.725
India in the big contractors like tall lock, Martin,

14:06.095 --> 14:09.245
those sort of thing, people, I don't wanna point out anyone,

14:09.305 --> 14:10.965
but I've been talking to them all,

14:11.665 --> 14:15.735
you cannot stop creativity if you want

14:15.735 --> 14:17.695
to standardize something and this

14:17.755 --> 14:22.445
and you say, this is the only way everybody is going

14:22.445 --> 14:24.245
to turn it away, so forth.

14:24.465 --> 14:28.645
In the same way as I was mentioning that, Dodo,

14:29.605 --> 14:32.555
which was actually a defense renaming at,

14:32.555 --> 14:34.235
so-called C four ISR.

14:35.205 --> 14:37.315
Don't ask me the,

14:37.315 --> 14:40.315
those people in the no know exactly what I'm talking about.

14:40.545 --> 14:42.755
They wanted their own terminology.

14:43.015 --> 14:45.195
The defense wants their own terminology.

14:46.015 --> 14:50.915
So they created Doda based on C four ISR, and there

14:51.285 --> 14:54.465
Is a, a long list

14:55.285 --> 14:59.755
of these architecture frameworks that are currently in use.

15:01.215 --> 15:03.435
And the reason is, is humorous,

15:04.845 --> 15:09.175
because they read the book, which can be,

15:09.495 --> 15:12.575
a hundred pages, 300 pages,

15:13.075 --> 15:16.535
and then another, another 300 pages that explains how

15:16.535 --> 15:18.175
to use the first 300 pages.

15:18.755 --> 15:20.775
So therefore, people take

15:20.805 --> 15:23.895
what they understand from the published architecture

15:23.895 --> 15:26.535
frameworks, and then they develop their own

15:26.535 --> 15:27.535
because they own it.

15:29.035 --> 15:31.415
And then when they own it, they believe in it

15:31.415 --> 15:32.615
and they start using it.

15:33.355 --> 15:37.295
So for that reason, when in the middle of nineties we said,

15:38.325 --> 15:41.385
Tim, now from nist, from National Institute of Standards

15:41.385 --> 15:43.945
and Technology said, why don't we standardize it?

15:44.325 --> 15:45.625
We said, no, no, no, no, no.

15:46.205 --> 15:48.225
You cannot try to standardize it

15:48.225 --> 15:51.225
because it'll be standing against all the others.

15:51.805 --> 15:56.385
The investment in people who who have already learned,

15:56.965 --> 16:00.825
mod of, south my own favor and on is a long list.

16:01.125 --> 16:04.185
But what we can do here is this Guam framework.

16:04.455 --> 16:07.385
What is it that it can do for you?

16:07.885 --> 16:12.825
And so RA was already existing and we said, it can do this.

16:13.045 --> 16:15.825
For example, it can equally represent the human

16:15.885 --> 16:17.145
and the automated task.

16:18.355 --> 16:21.855
It can equally represent the, the,

16:22.715 --> 16:25.455
all these different types of entities that are involved

16:26.515 --> 16:28.975
in architecting an integrated enterprise.

16:30.315 --> 16:32.095
It can rep with various entities.

16:32.205 --> 16:33.815
Like a project is an entity,

16:34.815 --> 16:36.975
a manufacturing company is an entity,

16:37.975 --> 16:40.415
a car is an entity, and so on.

16:41.555 --> 16:46.295
Um, you have to think about the software and the hardware.

16:47.585 --> 16:52.415
And people said, what are you serious? That's trivial.

16:55.075 --> 16:56.575
Not as trivial as you think

16:58.555 --> 17:02.685
because you understand the computer software

17:02.685 --> 17:07.255
and hardware that the subdivision hardware and software.

17:08.235 --> 17:12.855
But when you try to subdivide, human

17:12.955 --> 17:16.545
and machine the machine software,

17:17.845 --> 17:19.825
if the machine happens to be a computer,

17:19.925 --> 17:21.105
that's what you understood.

17:21.525 --> 17:22.585
The rest you did not.

17:25.205 --> 17:28.025
If the machine is an NC machine,

17:29.235 --> 17:31.015
the software is different for it.

17:31.675 --> 17:34.775
It is the instructions for the NC machine to cut the metal.

17:36.315 --> 17:41.145
How about the humans? The humans have abilities.

17:42.185 --> 17:43.205
That's how they come.

17:44.695 --> 17:48.755
What is the human software? Oh, well, these

17:48.755 --> 17:50.085
Are the instructions to people,

17:51.985 --> 17:56.205
and it can be a definition of processes or just policies,

17:57.015 --> 18:01.465
because if you are extremely skilled, you don't have

18:01.465 --> 18:02.585
to be given a procedure.

18:04.175 --> 18:06.955
If you want to tell a friend chef how to make an

18:08.915 --> 18:10.535
and say how to flip it,

18:12.265 --> 18:14.155
they will smash everything and leave.

18:15.875 --> 18:18.185
Don't you tell me how to make an,

18:22.525 --> 18:26.305
so that means you have to adjust the human software

18:26.405 --> 18:28.825
to the human hardware in the same way as you adjust

18:29.605 --> 18:32.185
the computer software, computer hardware.

18:32.565 --> 18:33.945
And at the age of ai,

18:34.305 --> 18:37.465
nowadays things are changing really, really rapidly.

18:39.995 --> 18:44.205
I have in my life studied AI a lot.

18:45.345 --> 18:46.745
I had a National Academy

18:46.745 --> 18:50.305
of Sciences scholarship at the AI lab at MIT.

18:52.115 --> 18:54.575
And I learned a lot there

18:55.845 --> 18:59.665
and developed ideas, which I'm currently using,

18:59.945 --> 19:02.465
I cannot talk about unfortunately,

19:03.665 --> 19:05.325
but currently using in

19:07.435 --> 19:09.565
appropriate defense areas.

19:10.945 --> 19:15.365
So, this architecting complex systems

19:16.505 --> 19:17.565
has a long history.

19:18.525 --> 19:23.145
And that book in the 1960, nineties 96,

19:23.285 --> 19:26.785
it was published, was just the beginning and

19:26.785 --> 19:29.345
after that, international standards

19:30.225 --> 19:33.545
applications in various industrial consortia

19:34.285 --> 19:37.505
in ship building, in, chemical

19:37.685 --> 19:42.265
and manufacturing, continuous process manufacturing,

19:43.325 --> 19:47.105
and other areas as well, has found its,

19:47.725 --> 19:48.665
its application.

00:00 - The motivation is something that goes back to the middle of 1970s where, around the world there was this idea that we have to, use computers for, manufacturing, which include, numeric headache controlled machines that, actually cut the metal

00:26 - computerated design and so on. And, various parts of the world, including, US Air Force, program, the integrated computer, in manufacturing program, and several programs in Europe, have started to build this exceedingly complex systems. The way I got into this is, I was at the time in Hungary in a very good research r and d institute, which had 700 engineers,

01:05 - control engineers, software engineers, plus support personnel, very well funded. And we were designing a, computer integrated manufacturing system for a machine tool factory. About 110 people worked for five years. And, even though that was, you know, behind the Iron curtain, the engineers were watching each other, what is it that they were doing?

01:36 - And so, from, from, the Office of Air Force Researcher, which was actually at the UK at the time, award came several times to visit. We went to visit as well, us because engineers talked to the engineers, how do we solve this problem? And, when I was recruited as a relatively young, young researcher, they were asking me, come and work for us, come and work for us,

02:09 - because I had worked for them for a few years before. And I said, what is your most difficult problem? And it is, well, we have a project, 110 engineers working on this exceedingly complex system. Give us a methodology how to do that, because the complexity was just overwhelming. And so during the past five years, I was heading a system engineer group, and we developed a methodology system was implemented and so on, but it was too expensive.

02:50 - It was too expensive. in the us they were using it for, aircraft being, production we Were using for product producing other types of parts. So how would it be possible to make it more affordable? The technology is fantastic, but it's just not affordable. The return investment is just negative. And of course, many people around the world are working on this, including Germany, including France, including the us.

03:25 - And, by that time, my group developed a method how to develop a reference model, if you wish, for a class of flexible manufacturing systems so that it can be implemented in a reasonable, at a reasonable expense. So, Jim Nvi, who at that time was the deputy director of the Charles Child, Dr. Laboratory in, I don't know, Draper lab is a part or not part is independent at MRT. but he raised the question,

04:10 - how would it be possible to bring together all the experts who work in this area from around the world? Jim was by the way, the, one of the main designers of the Apollo guidance systems. so he was a very, engineer, but he said, well bring the manufacturing control people and the, and the IT or computer science people together. What is it that we know how to create such complex systems? And so I was one of those experts invited, and there were others as well,

04:48 - many others as a matter of fact. and we started to review what is it that the different companies say how to create such systems? And as a result in the first, that was around 1990, as a result, we started to categorize what is it that people say, how to build this? And there were two types of proposals. One type, this is the structure of a manufacturing system, and including its control system.

05:24 - Well, that's fantastic. It gives you a pattern, but it doesn't give you a methodology how to do this. And so the type two, so-called reference architectures should be those which actually tell you how to get from the idea idea ideation, that you want a flexible manufacturing system to actually design, architecting, designing, implementing, operating it. That's another type

05:53 - of reference architecture relative to the ones That just said, this is how this should look like. And so this task force started to review these various architecture proposals. And the outcome of that was the book that we published called Architectures for Enterprise Integration is not for the faint hat. It is more like a monograph. it is not a, not for teaching purposes. It's a monograph that documents the thinking

06:26 - of the task force and how to go ahead. One of the, at the time they made, the task force ifi, international Federation of Information Processing and iact, international Federation of Automatic Control. These two organizations came together and created this task force with many members. most of them were from industry, but some of them for from academia or something in between an r and d institute. You can argue whether it's academia or not.

07:01 - And when we did this review before publishing the book that you were referring to, Fran was, you know, one of the people who said in the task force, so how it possible to generalize what people who are in this area of architecture, how to create flexible manufacturing systems. And the chairman of the task force was, Ted Williams, the chair of the Purview Consortium, which was originally created by Axon, but populated by many large engineering companies

07:40 - and, and control system vendors and so on. And I said, you know, I was relatively young researcher in my forties. I put up my head, actually, I know how to generalize all of this. They said there said, do it. So I did it with a colleague of mine and published, I think it's 1994, an article how to generalize This architecture Approaches, which give you a methodology for very large complex systems.

08:14 - And, the book that you're referring to, it took quite some time to actually publish, you know, it takes almost two years from the time you have the idea to actually publish a book about it. And then when this was presented, they said, Ari, now you are the chair of the task force. Because one person from the National Institute of Standards, techn in Burg said, dear, we have been going around in circle in ISO International Organization Standards Organization

08:49 - to actually develop such a thing. It would be possible to fast track as an international standards. And within six years, we fast tracked and became an international standard in the year 2000. The problem is that by that time and many others including the us, the, department of Defense developed an architecture framework called Dodo, which was, which had a similar intent.

09:24 - The British, of course, they improved it. They called modd and Canadians called it donoff. The Australians got it all stuff, and so on. That was just one part. But it was all manufacturing and control systems people. They had no idea what the sources software engineers were doing. Absolutely no. And the ignorance was mutual software engineers had no idea that we had this fantastic methodology, which would've been

10:02 - a systems engineering methodology. Really, the only problem is that a systems engineering methodology usually looks at technical systems where a manufacturing system has humans in it, and humans are always the problem and often also the solution. And our architecture framework, which was, which which we called is generalization, generalized, enterprise reference architecture and methodology covered it all.

10:31 - But where the ingredients came from were from these different European and other projects. So for example, the Perdu Enterprise reference architecture very eloquently, described how to divide the system into what the system actually does perform its mission and its control management and control. And these are the only two things that you have to worry about. But when you perform the mission, there are certain things

11:08 - that you automate and certain things that are being done by humans. And the same is true for the management and control. The Europeans computer integrated manufacturing, open systems architecture, mosa, they brought in the lifecycle concept, which says, at that time in a very, by now I would say rudimentary way, you have to define the requirement, the design and implementation of a system.

11:38 - We have refined that later. And the French, the French are always a bit outliers and say very interesting things, because the French have a different idea, they look at the world in a different way. For example, they don't have the word for design. They insist that they do it, but they don't. So when you ask what, how do you translate computer related design? They say co

12:09 - conception is not the design is design concept is come up with the concept. So to this day, the French and English engineers don't a hundred percent understand each other. Anyways, they brought a very interesting, structure to how to think about the management system of an enterprise, putting it all this together. That's g So now we had something actually in the year 2000

12:41 - and I, I presented it, in Toulu, we, the large ISO meeting. And they said, dear, dear, there, we should have known this before because we have been working on this anyways, that with the systems engineers. And, anyways, jumped 23 years forward or 20 something years forward, I realized that these two communities, the systems engineering and the manufacturing and control communities,

13:10 - they just don't talk to one another. So the only way is to, I was in the Manufac manufacturing and control world. I moved into this world and joined the systems and software engineering community. And so when you jump forward, we can see that now what was that, concept in the, manufacturing world has now infiltrated in pretty much all major systems engineering standard. It just took a cool 23 years for that to happen.

13:51 - One thing though, that is a kosher, I've been talking to, to people for who are doing enterprise architecture, India in the big contractors like tall lock, Martin, those sort of thing, people, I don't wanna point out anyone, but I've been talking to them all, you cannot stop creativity if you want to standardize something and this and you say, this is the only way everybody is going to turn it away, so forth. In the same way as I was mentioning that, Dodo,

14:29 - which was actually a defense renaming at, so-called C four ISR. Don't ask me the, those people in the no know exactly what I'm talking about. They wanted their own terminology. The defense wants their own terminology. So they created Doda based on C four ISR, and there Is a, a long list of these architecture frameworks that are currently in use. And the reason is, is humorous,

15:04 - because they read the book, which can be, a hundred pages, 300 pages, and then another, another 300 pages that explains how to use the first 300 pages. So therefore, people take what they understand from the published architecture frameworks, and then they develop their own because they own it. And then when they own it, they believe in it and they start using it.

15:33 - So for that reason, when in the middle of nineties we said, Tim, now from nist, from National Institute of Standards and Technology said, why don't we standardize it? We said, no, no, no, no, no. You cannot try to standardize it because it'll be standing against all the others. The investment in people who who have already learned, mod of, south my own favor and on is a long list. But what we can do here is this Guam framework. What is it that it can do for you?

16:07 - And so RA was already existing and we said, it can do this. For example, it can equally represent the human and the automated task. It can equally represent the, the, all these different types of entities that are involved in architecting an integrated enterprise. It can rep with various entities. Like a project is an entity, a manufacturing company is an entity, a car is an entity, and so on.

16:41 - you have to think about the software and the hardware. And people said, what are you serious? That's trivial. Not as trivial as you think because you understand the computer software and hardware that the subdivision hardware and software. But when you try to subdivide, human and machine the machine software, if the machine happens to be a computer, that's what you understood. The rest you did not.

17:25 - If the machine is an NC machine, the software is different for it. It is the instructions for the NC machine to cut the metal. How about the humans? The humans have abilities. That's how they come. What is the human software? well, these Are the instructions to people, and it can be a definition of processes or just policies, because if you are extremely skilled, you don't have to be given a procedure.

18:04 - If you want to tell a friend chef how to make an and say how to flip it, they will smash everything and leave. Don't you tell me how to make an, so that means you have to adjust the human software to the human hardware in the same way as you adjust the computer software, computer hardware. And at the age of ai, nowadays things are changing really, really rapidly. I have in my life studied AI a lot.

18:45 - I had a National Academy of Sciences scholarship at the AI lab at MIT. And I learned a lot there and developed ideas, which I'm currently using, I cannot talk about unfortunately, but currently using in appropriate defense areas. So, this architecting complex systems has a long history. And that book in the 1960, nineties 96,

19:23 - it was published, was just the beginning and after that, international standards applications in various industrial consortia in ship building, in, chemical and manufacturing, continuous process manufacturing, and other areas as well, has found its, its application.

Faculti AI can make mistakes