No.

Researcher will write.

Yes I will work really

well just to make

the life of me here just

recent years or

so but I ration.

You know really research you know

what are you

sure you're welcome thanks Nancy.

Welcome everybody.

Thanks for coming.

So this is a picture from students
from my class last year so

I want to emphasize I don't
have all the answers this is

kind of how we were collaborating last
year but I'm really interested in how to

bring different disciplines and different
people together to communicate and

share information more efficiently and
effectively and make better decisions.

So I found that I do know that I might
take the opportunity to this kind of I'm

probably one of the order or system
professors around I've been at this for

a little while in and
out of practice a little bit.

I thought I'd just take a little bit of
a journey through my own experience of

trying to think about and
develop these kinds of tools and.

As I get towards the end all talk
about where I see some future work

as being necessary.

So yes I got a master's degree at
University of Illinois Chicago and

this was kind of my first introduction
to design computation a list.

I don't list script and
Autocad I was in Chicago.

So I got influenced by Mr Vander Rose
international style and

so this was a generally a script in.

Auto CAD they could.

Just automatically generate
these different forms based

loosely on the rules of
the international style.

So I left there I went to work for
a few years but

was really intrigued by that idea of how
the computer could get involved in helping

us to generate and explore designs but
there was no analysis on that side.

It was all about generating designs and
having the designer pick one and

look forward so I after a few
years in practice went to MIT and

started to become interested
in this idea of analysis and

and large part because of George Stein and
some other folks who were there.

I became ingrained with
this idea of emergence.

And what does that really mean and so

just as a sort of simple example
here we can look at this and

we can sort of see nothing we can
look at this we can see a chair.

We can look at this we
can see window we can

look at this we can see kind
of both chair and window.

So this functionality emerges
out of the geometric form and

so I was interested in career
get the computer to kind of play

a part in this role of interpretation and
analysis as well as just generation.

So I built and
built a system I was actually.

On a virtual collaborative network
back in the mid to late ninety's.

So I was pretty early with
Mr Vishal lectured research lab to

to develop an online collaborative system.

And it kind of looked like this
essential way it was a fairly so.

World of blocks but we have different.

Actors in the box something that looks
at the world interprets webspace spatial

design something looks
at the world interprets

inclosure something that interprets
structure something that interprets light.

And all sort of mediated through this
central design world where each person

would create their own filter or agent and
then would go and and interpret and

then structure structure and change the
world and interpret instructional worlds.

Different humans and agents could
situate themselves around the world.

So these are the constructors kind of on
the left hand side a space constructor

might interpret the need for
more space and

then go ahead and add elements to try
to create more space for example.

So I had some pretty simple management
systems which would go in and

look at the world and understand
the evaluation from each perspective and

understand who could go next and
it would help the guy to design process

like this I would try to improve
these different perspectives.

So I left on my T.
and was really interested.

But this was a toy problem I felt like I
was a long way from having any real impact

in practice I had went out to Stanford
University and started a Ph D. there and.

One of the first things that Martin Fisher
my advisor at the time suggested I do is

go out into practice and
find a real problem and

I think I was really great advice and
something I try to do with my Ph D.

students as well as get them in grained
in practice to find a real problem.

So the way I did that his research
was around forty modeling.

So I spent a lot of time gathering
models from the different designers

contractors bringing them together
bringing in the construction schedule and

constructing these kind of animated
movies that would show us how for

example to build the inside of the.

Concert hall or in this case the outside.

And so through that process.

I became very aware of data flows
different people's needs for

different models and if you remember my
mit project was all about this kind of

central model around which different
folks would come in and add and

remove things I started to see the world
a little differently here is a little bit

more of a federation of loosely
coupled distributed representations

each of which was necessary with
then dependencies between them.

So on showing here is kind of the flow for

the construction schedule the flow from
the geometry from the different people and

what was necessary to go in and
build this forty model.

So my thesis built on this this this idea.

I called it narratives.

I looked it up in a dictionary and
a narrative is an accounting of events and

the relationships between them.

And I felt like that sort of applied for
what I was thinking about here.

So I developed this narrative
formalism which essentially involves

representations some model.

Some reasoning and some dependency
between that model and other models.

Kind of combines our
reasoning algorithms for

different analyses say our representation
our AI of Caesar or bends and

then structuring and sequences you
these interactions in terms of process.

So I return to Disney Concert Hall and

I got involved in the Frank Gehry
building there was wasn't

a right angle to be found on the entire
project unfortunately so we had lots and

lots of conditions of needing to build
connectors between the beams and

the concrete slabs things like these
called Deck angles and everyone

was different and my job initially was
to go through and find those conditions.

Manually And I said hey there.

There must be a way to formalize
these relationships in a way that

we could do this automatically.

So I took this idea of

the narrative of saying these deck
attachments depend on steel and concrete.

And I broke it down in
terms of a series of

kind of lower level dependencies where
I would analyze the geometry and

break them down into features
extrude some geometry to look for

intersection between the slabs and
the beings where there were intersections

I would go ahead and automatically
construct the deck angle attachments.

This is good software prototype a kind
of drag and drop think of this is

kind of a predecessor to grasshopper
This was back in two thousand and

one two thousand and two.

And so there were the kind of deck
attachments which came out of the process.

And did so
in terms of research validation and

I claim power in generality of this method
by showing that on the actual concert

hall they missed a whole bunch of deck
attachments that we had to then go and

fix in the field and my system was able
to identify them automatically and

for further generality I I found I
thought this was kind of compelling

that many of these lower level algorithms
that I used to find deck attachments.

I could really use in a separate different
ways to do other things like in this case

I was looking for cantilever conditions in
the complex ceiling panel of the concert

hall and was able to take many of those
same algorithms as as well as some others

and stitch them together in different
ways to look at those conditions.

So.

I finished the Ph D. then hung
around as a postdoc for a year or

so and was able to get a job at
Stanford as an assistant professor and

I put together a really multi just.

Structural engineers architects mechanical
engineers mathematicians computer

scientists.

Management Science folks I
got a bunch of funding from

digital laboratory Boeing
laboratory shouldn't be there.

They didn't fund this work but they've
certainly helped me since I've gotten here

and work with these people on
some projects I'll show you now.

So you know the first thing is really
identifying what the problem wasn't

trying to come up with some metrics that
we could start to measure our impact

on our success.

So we ran some case
studies observations and

practices the hey what's
really going on out there.

What does the design and decision making
process look like we found this was a case

study that's the one that typically there
conceptual design process is when most

of the major formal decisions have been
made about the building authorities and

that's when it all reassure things
like that was happening in about five

weeks highly architecturally dominated.

Very few formal explicit performance
objectives were being defined very little

performance analysis was being performed
you know basically there's a look cool and

how much do we think it
might cost to build and

that's what they're
basing the decision on.

So you know kind of jumping back into
theory looking at this is an analysis.

Of decision methods it's became clear that
every good decision requires processing

a lot of information very effectively in
order to look at more alternatives from

more perspectives we need to do all
of this much more specifically and

explicitly and that we're having trouble
processing all of this information.

So the following tools will start
to try to address this issue.

Kind of different pieces
from different perspectives.

Of the problem.

The school is called Pip or

process integration platform this
is work by reads and ask you.

And essentially this is
an online cloud based

process mapping tool that allows
you to actually engage in.

Executing the process as
well as mapping them out.

So essentially projects today we typically

collaborate through
hierarchy's of you know F.T.P.

sites or something like that what we're
doing here is just adding an extra simple

piece of information which is dependency
between these files and folders.

So if just to kind of help you
interpret this each of these nodes

talks about what the node is
when it was built who did it and

what tool is used to
construct their view it.

And so this is a decision making process.

You know on teams objectives preferences
alternatives impacts values and

then each of these will break down so

often that the alternative generation
process is a complex process in and

of itself that ends up with some rabbit
models those rather models go in to

daylight analysis energy analysis exciter
if we dig dig into the daylight analysis.

It has its own process and

ultimately all this information flows
up into a decision at the end there.

So we use the the map to plan out
the processes execute them and

then you can see these colors green and

red to help students understand
where they were in the process.

So the pip actually with this really
simple additional piece of information of

dependency between information
started to it help us we believe

communicate share and understand our
processes better just in terms of

the communication the collaboration
I kind of are you described it.

But having everybody on the same page
about the flow of information and

where we stand in terms of the status.

With this extra piece of dependency
information we can also start to search

these processes.

So here's an example where I'm searching
for an input of some files that have

the strings arc and I have C. in them and
some output for strings that have L.C.A.

or life cycle assessment in them so
I was able to find you know a series of

different processes in the database go in
and copy a particular process pasted into.

Our process to help you share information.

And then there's this idea
of understanding that we

could start to develop communities around
different processes start to score them

talk about how long they've been they
take how many times they've been used.

And then we said we were also able
to construct some other analyses of

the database.

This is a class studio
project sensually this is

a different project pedestrian bridge
train tunnel university lab and

then each row within Here
is a different student.

So we're basically trying to
map who put out information and

who consumed information.

So you can see here this is
a fairly tight integration of

everybody sharing information
with everybody in the process and

down here these projects are perhaps
a little more distributed not sharing

information as effectively as many times
really just in this case just measuring

number of information handoffs nothing
about the content in this case but

it was interesting to the projects
that were graded in terms of their

just quality at the end of the actual
performance of of the design and

we could see at least the visual
correlation between tight integration and

high scores and
loose integration and low scores.

We ran some students given a confined
design problem in a short period

of time a couple of hours and had them
designed with this tool half the students

got PIP with all the dependencies
others are asked to design with just

a folder hierarchy without understanding
of what the dependencies were and

we found that as we
coded their discussion.

Sorry about that we found more
trend discussions more actual

iterations they were generating and
analyzing more alternatives.

When working in the context
of the process maps

their information was more consistent and
they were less statements of confusion.

We also tested.

Sometimes if they could actually go and
copy previous projects.

Into their project and design in that way
and we found that there were better scores

of projects and less mistakes made when
they were able to do that and then finally

it's understanding a project that kind of
just showed you being able to look at this

data and perhaps try correlation between
say integration and project performance.

So it was really about communicating
dependencies between information and

getting people to to do that
integration them selves.

This next project is called pyro or
process integration design optimization.

And this is about formalizing integrating
automating those those dependencies.

So this is a tool called Phoenix
model center something that

several of you in this room.

Use And here we're wrapping Frank
areas digital project parametric CAD

tool energy analysis doing
some preprocessing for

structural analysis doing
structural analysis and Arabs.

G.S.A. tool and then using one of.

Phoenix's own genetic algorithm optimizers
through our run this process so

essentially generate geometry
analyze it for energy analyze it for

structure look at the results and decide
Should you change a structural member

change a material or
change your commentary.

So we do this around a very
simple case study at first and

this is the kind of results you can get
three looked at I relocate a classroom.

We can spin this any way we want
to we can make that window bigger

we can make the room wider longer.

We can change different structural
members in these different

green members with
different structural sizes.

And you get results that look like this
these each line here represents a design.

And so if you follow all the lowest cost
designs you can see what leads to them.

Longer building lengths lead
to lower cost design as well.

Why is that as we made the building
longer we constrain the floor area.

So the building became thinner.

And therefore it was a shorter span so
it was cheaper just cheaper to frame and

build a long thin building
than a square building is

what that's telling us
you could see other.

Places like call them
sections good designs groups

go through all different columns
sections so it really doesn't matter.

What column section you choose.

You'll get to a good design.

This is a predator front here showing
the trade off between first cost and

lifecycle cost where each dot
here represents a design.

And so we're really looking for

designs along here it just depends on
your preferences between structural and.

Lifecycle performance cost
uncertainty is not not considered of

course but where I find we are often in
the design practices we're back here.

We're not optimal anywhere we're sort
of floating around a non-optimal space.

And these tools can help us
get to that printer front.

So the couple students dig in deep on this

some bend welly when after a thermal
analysis the late and energy.

Analysis and did a lot of work in terms of
really working through this data flow so

you can model in an parametric
model in digital project.

And have your files prepared and

push through energy analysis and
daylight analysis automatically.

So you can actually optimize for
the US and

for us figure Phleger looked
at the structural analysis and

he looked at a lot of steel frames for
particular stadium roofs and

developed a bi level optimization method
where essentially if you look at this this

frame we can optimize in a number of ways
we could just put different members on

on these different different size
members on these different members and

just optimize the section size we can
also optimize the shape by making

the larger or our or

this length larger or we can optimize
the topology or we have different members.

Fanning in different ways for us.

Just look at the first two but
found you needed different optimizers for

each to do sizing you needed.

It's really.

Called a discrete optimizer that can look
at different sections sizes discretely.

With a shape optimize you want to
continuous optimizer that can look through

geometry continuously and both of those
require different types of optimizers.

So that's why I've looked at
communicating design process of looking

optimizing at generating lots
of information we finally and

we need to bring all this information
together and make good decisions

great if we could optimize it
tends to not work out that way.

There are lots of objectives which are
hard to fit into an optimization scheme.

Excuse me so
I created a tool called resurgence.

Which I see is sort of the intersection of
some interesting things that are going on

and in our world these days.

One is social networks
bringing people together and

online communities around certain
topics we have document management.

You know Dropbox things like that
places to put our information online

and manage it.

We have task management
management actually you know this

is scheduling planning keeping track of
tasks and how they're being executed.

And finally have decision models formal
ways of bringing information together

wearing it.

Understanding it and
making decisions with it.

This is a little bit
sequence of as possible.

So how does research work well the first

step that you can often do is you'll
define your different stakeholders and you

ask them to weigh their goals which goal
is most important to them for example.

For on this particular decision and

you can see in this case what we're trying
to decide is for this decision of what we

should prefabricated what we should
build in in place on a project for

these are all the different goals that the
different stakeholders are concerned with

and these are the different
stakeholders that we are considering.

Once you go through the analysis process.

You end up with view.

Like this would show you
overall where the value is for

all of these different things
that we might prefabricated So

for example we could see that
substations for example there's lots and

lots of value here perceived for
if you were to prefabricated.

What's the value well we can see the brown
cost savings schedule savings but

then there are all these other reasons
we why we might want to prefabricated

to in terms of creating standardization or
flexibility for the future.

So this was the output of the team's
analysis we can also filter this

information through here I've just clicked
on the designers and I've just used their

call priorities to figure out which
decisions would make most sense for

just the designer and so I can help
negotiate and come to consensus on a team

so and then we'll end up with views like
this which in this case is a different

decision looking at what structural system
to pick for that for the project and all

of the reasons why they chose structural
steel short span because of its speed of

erection first cost impact on day lighting
integration of lateral options etc.

So.

I argue we make the basis of a lot of
our decisions right now on cost and

schedule and a lot of the other things
that are harder to kind of formalize and

bring into the same discussion with
cost and schedule gets left out of

a lot of decision making processes and so
we says it's just trying to formalize all

that rationale and show the other reasons
why a decision may or may not make sense.

So I I I work with Bill McDonough
a little bit early on trying

to understand their flows and he created
this diagram that I I quite like and

I show myself every now and again just
remember that this is an ongoing process.

I've shown you about ten years.

Work up to this point.

And it's you know kind of
still in the fog area but

start to understand how we might use
these tools to enable good form or

collaborative decision making for me.

So what you know what are some of
the outstanding issues I'd try and

transfer from Stanford
here to Georgia Tech.

What are some of the things that I see or
are still missing and

and where where do we need work.

So let me just explain this kind of
complex diagram for you for a moment.

That the dark lines the black is that
decision making process that I have for

I talked about a few times already.

Right.
And so you go through in a Figure

out of your teams are you
define your goals you define

your alternatives you
way this information.

You understand the impacts you create
a value you go ahead and you build.

Well we should see a particular as we
start to bring all of this stuff online we

should be starting to learn from from this
from all of our decision processes as

well as from the performance of what
we're actually doing in in the world and

so all of that data should
start to feed back.

So should we should understand you know
are we actually getting the building

performance that we
predicted we would get.

And what does that tell us about
what our weight should be and

what the impacts actually are as opposed
to just what our tools are telling us.

And if we understand this.

What alternatives should we
really be considering and

what goal should be looking at and if we
understand that who should we bring on to

the design team at the beginning to
help us get to this best best design.

So there is kind of this.

Ideally this reinforcing
loop of getting better and

better information by making decisions
capturing that rationale as well as

tracking what's going on in the real
world and having that feedback in.

There is there's a number of kind of
other and related issues with this and

I've really been working and
dealing with a lot of you on all.

All of these issues.

So one is just the organizational
aspects with the livery methods

work best the different contacts to
make to bring people together and

actually make good decisions and

there's also this soft notion of trust
which quickly part is pushed I'd works on.

How do you get people to really
share their information in the right

way through this process in
order to make good decisions.

There's lots of work in terms
of generating alternatives.

Coming up you know we sit down.

We just build parametric models.

Is there a method or a system system
with which we should approach

building these parametric models we're
looking through the right spaces grammar

shape grammar is and graph grammar
is how do we break out of the top

colleges prison I guess the parametric
systems can put us in and

really explore larger spaces with grammars
and then rename reinvestigating so idea

of expert systems and how can this help us
generate better alternatives more quickly.

And their issues in terms of performance
in terms of model accuracy certainly

Friedan Jason are working a lot on these
issues to rule checking space in tax

different ways to analyze our designs for
things beyond just energy and

structure economic analysis I
know that the decision theorists

like to try to take all of our objectives
down to one objective which is money and

help us make objectives with that So
how do we take something like daylight.

Or views and reduce that to a number
that we can make decisions with.

Up on decision making game theory bringing
multiple people together how to engage

them in formal decision making processes
as well as the uncertainty work in these

decisions and how do we how do we use
that to make decisions and then finally

there's this concept of integration how
do we bring all this together to do this

all more efficiently effectively
there's arbitrament G.I.S. platforms.

Lifecycle management are process based
ways of connecting this all together and

I'm becoming a kind of big
advocate of systems engineering

model based systems engineering
as one discipline that's kind of

step back and
is trying to develop the theory and

the formalisms to bring all this together
into one system that we can execute.

One one question as we're starting
to develop more and more tools more

more methods and how do we pick the right
method for the right challenge.

So we do have we have different
challenges right we have some fairly

simple challenges like
what overhang should I.

How big my overhangs be on the window.

Some more complex decisions
like what structural system for

a new construction project.

And then some really complicated decisions
like what program massing should we have

for a campus.

So each of these challenges should
probably require a somewhat

different method.

And we have different methods you know we
have kind of current practice methods of

what we do today.

We have design thinking which tries to
somewhat structure a formal design process

but not terribly formal.

We have more formal decisions
like choosing by advantages or

weight rating calculate.

We have optimization norm of
decision theory things like that.

So the question for me is we
provide some guidance or understand

which method to apply in which
challenge which one leads to the most

effective decision making for each
challenge open research question for me.

So how do you answer these
these questions moving forward.

I've been developing a course for

you Mike I'm sort of thinking
of it as a curriculum.

Called design space construction.

And I've tapped into the vertically
integrated projects.

Concept of Ed quite well here and that

really you know I'm a little frustrating
to teach students that kind of get good.

And then they go off and you start the new
students and they kind of get good and

they go off.

So this is the idea of trying to
create a multi-year project where

I'm students can come in
as undergraduates and

hopefully continue through multiple years
up through the ranks of academia and then

we also want to bring in professionals
into this class into this process to help

us understand the challenges and
give us good problems to work on.

So trying to integrate computing civil
engineering construction architecture

other folks as well as some companies
are getting involved in helping us.

Do the project.

So we develop some quick to help you
understand how to walk through these

different steps of building a design space
on a simple challenge and then we bring

in the designers and they apply these in
practice on more complicated challenges.

Not to show an example of
one student project here.

So this was you can many of you
probably recognize the names of some of

the students there.

So here they're working on a Perkins and
well project for a hospital facade.

The challenge is essentially here's
a floor plan a part of this facade to have

this aerated kind of edge
windows here bathroom bed

and we're interested in
optimizing this building form for

things like comfort external views
of quality functionality of space.

So the first thing they do is they analyze
the constraints under understand what

you know what's fixed what they can change
as well as what some of the drivers on

the design might be so
they know how what analyses to focus on.

And so here we are we're just trying to
understand how big this window should be

Reeses how big the shadow
box should be for example.

They get together in teams and
they develop process.

Diagrams this is using system all modeling
language describing a high level process

which breaks down into lower level
processes for how they're going to go

about doing their design process and then
developing what I call block diagrams or

think of in this kind of schema diagram so
here are looking at the hospital room

which has ceilings and beds and
floors in the bed has a viewing point.

Etc So defining the actual data model
that we're going to use in this analysis.

So we've been working with grasshopper

number of interesting tools
emerging in this area but

here is part of their model for creating
more logic for creating the geometry.

For laying out different analysis
points for analyzing daylight.

For setting materials.

For analyzing view angle and
cone of vision.

For creating a shading system for

analyzing the thermal
daylight performance.

One sort of thing a skill.

I try to teach is you know building
the right model for the right challenge.

If you try to build the entire tower and
do all of this analysis I just showed you

you'll get nowhere but recognizing that
each of these rooms are repetitive.

So you can do a lot of
analysis on just one room and

then you hear they're just building it
up into a tower so they can analyze it.

Statics and so
that's the overall grasshopper script.

And when they run it.

We're using optimizers story.

So such if you remember that predator
front I showed you before we're trying to

block less daylight in
other words get more view.

We're trying to have less heat dance
we kind of want to be down here.

And here's our pro front showing us that
at some point as we lead in more and

more late more and more light.

We're going to be increasing
our heat gain and

that's where the design team the Perkins
and will design was that they brought us.

And so you can see that we're
starting through our optimization.

To explore the space and
find better performing designs.

So the students then go and document.

Instances of each of
the design alternatives.

The finding them very clearly what
the parameters where the LED to them.

What are the goals that
we're going to measure for

how we're going to measure them.

What are our constraints.

Gathering all that data into analysis
different alternatives different goals in

the performance data for
each of them starting understand

sensitivity like you know what are the
what are the moves that are making us.

Get less heat gain or more he came for

example understanding
the data analyzing it.

And then ultimately weighing all this
information making decision about

what the best design is.

So that's the idea of this design space
construction curriculum I've really reach

out to all of you to help me
in that I view it as a multi.

Disciplinary collaborative effort of
really getting people together applying

research on of these design processes and
understanding what works.

So I've also been working
on the Condell building

right across the street a labor
of love and frustration.

I'd say at this point but
we all are familiar with the design but

I just want to bring this up to highlight
what I think is really important.

I show you that picture at the beginning
of the semester at the beginning of

the lecture excuse me of all those
students cobbled around a laptop.

I think this is really important to
develop a really flexible highly

digitized space that we can
quickly reconfigure and so

what we're looking at is half
of the first floor of the canal

building will be called this flex
space will have dividing partitions so

we can open it up and have one big
space or smaller spaces and you can see

just some of the configurations
that we imagine being able to.

To reach.

Configure and use the space in
different ways to really get

people connected to their
information better.

So I thought I just started concluded with

a new chapter in my life that
has been really interesting.

And that is through my collaboration with
Perkins Well particularly the design space

construction class.

And they've they separate So
this decided to start looking for

a director of research at Perkins and
well.

And so I've begun that process.

And so right now if you go to Perkins and
Wells website and you and

you Google research or not.

Google or search for research you get kind
of all sorts of stuff about different

research centers they've done.

There is a fair amount of research going
on of Perkins and well John is off.

Now others are doing that as you know but

the firm lacks kind of direction and
clarity and focus in terms of what

they're trying to do in terms of research
and they see that as a real priority.

I'm to try to develop that.

That focus and

develop better collaboration with academic
institutions to further their research.

So how how we kind of looking
at Perkins and Will has a number

of practices very diverse firm but
it kind of breaks down and

the way I like to think of it is
these four major practices education.

You know we're bring
the people up workplace where

we do the work health care we take care
of ourselves and make us healthy and in

the cities that we all involve And so we
have a number of kind of sub practices K.

through twelve higher education for
example in each of these practices.

And then we're developing channels of
focus and expertise in research in

a number of areas design process building
technology resiliency kind of new

sustainability I think health energy and
water and so forth.

We're looking for
people internally in the firm to.

Head up each of these channels and

then to interact with the practice
in a big way to try to create create

more research and and understand
research problems at the beginning and

then actually implement
research results at the end.

So I'm I'm in the process of trying to
create this organization right now.

A couple things to point out here so we're
looking at having what are called Channel

leaders for each of these channels
they all have a knowledge manager.

And then we're looking at creating
fellows reaching back a day to me and

pull folks in to have a fellow within
each of these different channels or

kind of building this up slowly but
that that would really help to create that

collaboration that we're looking for and
then this thing over here called Area

research we recently created
a nonprofit five a one three C.

entity which allows us to
collaborate on a basis in

indifferent research projects to go for
federal funding and and so what

we see here is basically you know we could
bring in money through area area with then

contract out the Perkins and well as well
as to other academic institutions and

researchers to do the work and then we
would publicize the work through area.

I'm up on a process guide should've been
kind of clear with all the process maps.

I think got purpose one needs a lot
of help in terms of structuring these

processes creating our information and
the formal ways that we can reuse them

some trying to kind of instill this
research process throughout Perkins and

well which is essentially starting
with the practical problem.

Identifying what the existing
knowledge is that we can build on.

If the existing knowledge of
solves a practical problem.

Game over.

We don't need to do any
research if it doesn't.

We asked for research questions.

We'd.

Choose research methods
we design research tasks

we create answers to our research
questions and validate them.

We claim a contribution to knowledge and
relate that explicitly to the existing

knowledge and then we talk
about its practical impact and

I see it as a very
important collaboration.

I mean that I think the designers in
the professionals have a kind of better

grip and
sense of what's going on on this side.

And academics you guys have a much better
grip on what's going on that side.

So it's about bringing these two
groups together executing these

processes reusing the information
more effectively.

So my last slide.

In my role at Perkins and
well I'm looking to really

reach back across the freeway over here
to develop a number of collaboration.

With Georgia Tech and other institutions
and these are just some A list of some of

the projects that are you know we really
started to have some conversations with

folks about Perkins Well designers and
Georgia Tech people working together.

And I'll stop there.

Forty five minutes.

Thank you.

Chuck.

It's

a good an excellent question.

If you talk to Phil Harrison.

He's like you know put it
out there keep moving and

six months it's going to be old.

Anyway.

So that's his perspective.

If you talk to a lot.

So of designers and
the people doing the research.

They're not so sure they see you there.

It's.

There's kind of two issues one is it's
a competitive advantage for Perkins and

well and they want to keep
internal The Perkins well and

sometimes they want to not even
share that widely right there.

It's what makes them valuable in the firm.

So they are a little afraid.

So that's an ongoing open question.

Kind of a nice thing about area
is that it's short circuits a lot

of that it's a nonprofit
it has to be open.

It has to be public and any research that
goes through there has to be published.

So that's one way to kind of.

And that conversation quickly and
if it does go through Perkins and

well then we probably have to
have that conversation each time.

About how the knowledge that's
created is being handled.

One moment if we start.

Right right.

Yeah

right.

A

big issue.

I guess is in developing these tools
take the we sit decision to write.

In there.

I formalize things like
that you have a team and

each team consists of stakeholders design.

They keepers and decision makers that
level of formality allows me to create

certain roles downstream that facilitate
the process and make it go more quickly

but I don't formalize what the specific
objectives are how to measure them.

And I leave that up to the team and.

Still that's a fair amount of work for
the team we go through and

formalize all these objectives
gather all their alternatives

gather all the information about them
weigh them make the decision and

it's really hard to get them to even
go through that level of additional

formalization so sometimes will
reuse a decision where will create

a decision with a bunch of objectives and
it will copy that the next decision.

So those objectives
are formalized if you will.

So I guess my my answer is it's
a bit of iterative process

of working with you to find out what the
right level of formalization is what you

can formalize ahead of time and leverage
and what you need to leave to the team.

It's an ongoing struggle like we
sit in isn't flying off the shelf.

And so the only way I know through that is

like in the design space
construction curriculum try it.

Iterate build models try them out

part of it's just getting the design team
thinking that way if they can think that

way they can create the formalizations
it doesn't actually take that long.

But they have to be trained in doing it so
you have to train the the designers and

then part of it's the tool
formalizing right things.

One winded not very clear answer but.

If you do you going in.

Do you need another where is our

life.

Right.

Sounds like a great research question and
I think I you know I showed

you that crazy graph of like challenges
against methods and design exploration.

I think there are certain decisions
like you know where should we go get

sushi deny that or
certainly probably better handled with

the loose dynamic of you know Facebook and

then there are going to be other decisions
which are more complicated require more

structured information clearly
understanding the uncertainties and

the tradeoffs and
all that that I think we need more from so

we read

well this is being filmed not
that this is being filmed or it.

Actually I offered the recent thing
at the beginning of the process.

Seminar.

The couple planning a space management in
the great wisdom chose not to go that way.

I think we really did run into issues of

different people with different goals and
different priorities and

you know we're six months
delayed off a part of that.

So I think you know it's a prototype tool
it's not exactly ready for prime time.

Perhaps but
I think something like that could

have facilitated this conversation
earlier we would have had this delay.

Talk.

I agree and
maybe it relates back to our freedom.

Question What do you formalize So
in in the East Asian thing for

example you can certainly create
a stakeholder called on our occupant.

And that's kind of what we did and
when we finally did

use we sit in on the kid elbowing so you
can include that and I completely agree.

They need to be stakeholders they
need to find their objectives

we need to figure out how important
they are versus the other objectives.

For.

I think so.

I mean that said.

So the essential thing
starts with figure out.

Who you care about who the team
is it's a first step right.

Second step is finding the objectives were
the constraints and one of the goals.

Third step is weighing the goals.

For step is weighing the stakeholders.

So at that point once you're done with
those four steps we know who's involved.

What they care about how
much they care about it.

How much we care about them.

So that's those are your requirements.

Sure.

That's what I guess I'm showing here

and I showed it here as well.

Right here.

SARS lot of slides right there right so
here's here's the outcomes.

Outcomes is the effect just the design or
outcomes as they affect everybody.

Missing a lot of the stuff you guys
are working on in terms of uncertainty and

in those issues but.

Both.

So we are we have.

First thing is values placed on the
performance metrics stakeholders how much

they care about objectives.

Second thing is alternatives how they
perform on those objectives third

thing we do what's called Weight rate and

calculate when we multiply how much
people care about the objective

by how well an alternative
performs I'm not objective and

that creates a sort of proxy for
value and.

It's all free.

Really you know.

Thanks very much for coming in.

Thank you very nice to see
you were just looking.