ADNY: Lost in Translation

**Daily Gullet Staff** · September 16, 2005

The Daily Gullet is pleased to present the second of three excerpts from the just-released book The Seasoning of a Chef: My Journey from Diner to Ducasse and Beyond (Broadway, 2005). Part one is here.

by Doug Psaltis with Michael Psaltis

What made me realize I was in the right place [Alain Ducasse New York, aka ADNY -- ed], and that I would fight to stay there, was the kitchen itself. It was levels ahead of any I had ever been in. In just about every other kitchen I had worked in at least part of the accomplishment was stretching to overcome physical limitations. There was no creative rearranging to overcome design shortcomings, and no corners were cut in creating this kitchen. It was not nearly the biggest kitchen in New York, but everything from the hundreds of pen-tip size fiber-optic lights that would provide light above the stations to the hand-crafted stove was designed for excellence. It was clear that the cooks wouldn’t be challenged to exceed the limitations of the kitchen, but to reach the level of excellence it made possible.

When you entered from the dining room, the kitchen’s automatic swinging doors opened to reveal a dazzling black tile floor and gleaming stainless-steel equipment. It was an open kitchen with a large window that overlooked the main dining room, and there was even a private dining room in the kitchen. The massive stove took up much of the kitchen. It had three flattops and one plancha, a mirrorlike metal griddle. All of the stations would work around the stove. The pass would be on the side closest to the dining room. The chef would stand between the stove and an island (the servers’ pass), and the servers would pick up plates that he put onto the island. The dishwashers would be in a separate room off the kitchen, and the front of the house would have a separate room for its own preparations. It was hard not to appreciate the stylish side of the kitchen. It is as much of a show kitchen as those on the Food Network, but this one is truly functional. As a cook, I appreciated the details more than anything.

Attention to detail is what Ducasse is all about. The thought that went into everything was incredible, and this wasn’t just regarding the kitchen. The day after we unpacked the pans, we started moving in the products—dry goods, produce, and meats. Also on the scene that day were some of the staff for the front of the house. Part of their day was spent being fitted for their uniforms, which had been specifically designed for the restaurant. The uniforms, custom designed for Ducasse, were both elegant and stylish. The front of the house would also have a second set of custom-designed uniforms that were to be worn when they were setting up. For a restaurant to purchase uniforms for servers for preservice work was unheard of. These weren’t just aprons. They were full outfits, nearly as fancy as their serving attire.

The first two weeks at ADNY were a blur. From the very first day, though, everyone knew his place in the kitchen. There was a solid chain of command, a rigid order. The positions in Ducasse’s kitchen are based on the traditional French model. There was, of course, Alain Ducasse at the top, then Didier as the chef de cuisine, then Olivier as the sous chef, then there was a chef de partie of each station: the entremetier (in charge of hot appetizers in Ducasse’s kitchen), the gardemanger (cold apps), the poissonier (fish), and rotisseur (meat)—the pastry station, which also baked all of the bread, was separate. On all of these stations, the chef de partie was the saucier (in charge of the sauces) and the manager. He was truly the chef of the station, responsible for everything and everyone, from ordering, receiving, and evaluating the principal products for the station, to the dishes we sent to the pass, to how we cooked and dressed. Then there was a demi chef de partie, who did most of the actual cooking during service, and then there were a few commis—the lowest-level cook, who did whatever was needed. Everyone knew his position, and everyone answered to the man above him and for the man below him. A lot of kitchens are focused and competitive, but this was more like the marine corps than a restaurant. We were there to do a job and to do it absolutely perfectly. It didn’t matter if you were good buddies with the guy working next to you or if you absolutely hated him. All that mattered was doing more than was expected of you.

The first time I heard the sous chef yell in his harsh broken English, “What the fuck is this! You insult me with this shit! This is shit!” it was over a broken sauce, in which the elements hadn’t emulsified correctly. It was a scream of incomprehension, something that you might expect from someone who had just watched a stranger walk up to his mother in public and slap her. This was in the middle of the day when we were only cooking through the menu and plating the dishes so that we could understand them and perfect them. The recipient was a French cook who was the chef de partie of the station I was working on. He hung his head a bit but kept working. The restaurant wasn’t open. We weren’t even serving special guests.

I didn’t know much more than oui and merci when I joined ADNY. I had started at the bottom of the kitchen, as a commis, but after just a few weeks I was moved up. The French guy who had been our chef de partie was suddenly moved off our station. Now I ran the station as the unofficial demi chef de partie, even though I didn’t know French and that was the only language used during service. For the first few days running the entremet station, I’d hear Didier call an order and I’d have to pause as I tried to figure out what he said before reacting. Olivier stayed near me most of the service to translate. “Hey, homard, lobster, vas-y,” he’d say. I was constantly behind, almost unable to keep up. I had to pick up French right away -- à la minute -- or I’d be gone. So, I took one of the first printed menus home. I translated as much of it as I could (mostly by calling Greenie, who had returned to Los Angeles after giving up on a small restaurant in the West Village that he helped run), and after the first service I had learned at least the main ingredient in every course that I was involved in. I went back to work the next day ready to listen for homard, petits pois, poularde, and more. But the menu had changed and only a few of the words I knew remained useful. I was back to memorizing in the time between work and sleep.

+ + + + +

Halloween 2000 was in most ways like every other holiday for me: I was working. What made this day different, though, was that by midnight we’d be certain of our fate. I knew by the time service was over that night, when the New York Times was first available on the newsstand, we’d either be celebrating or facing changes.

I’d been in charge of the entremet station for a few months, even though I still hadn’t been given the official title. Just before service that night, while I was working to finish the entremet station’s mise en place, Didier came by the station. “En place, Doog?” he asked. “Doog” was how all the French guys said my name. But I didn’t mind, as it was a lot better than “dog,” which is what they thought my name was for the first two weeks. “Oui, chef,” I replied.

At ADNY just as the service began, while all of the cooks were checking over every little detail, the kitchen’s lighting was changed. All day the room was brightly lit, but for service only the lights that were necessary were used. The lights accentuated our focus. Hundreds of little fiber-optic bulbs shone in the precise locations where we would work while the rest of the space was unlit. All of the distractions were removed. The stage was set, the show was about to begin. Didier began to call out the orders—no one else spoke—and we all moved. First just a few of us and then everyone. There was little noise, except the sounds of cooking.

The level of focus in the kitchen that night, as on most others, was extremely high. I was managing all of the orders we had coming in to our station and what we were sending to the pass, and I realized that we were running out of ravioli. As soon as I recognized that, I had Brendan (“Soda Pop”) start rolling out dough. It would take only a few minutes to get the dough rolled out and cut, so we wouldn’t miss a beat in the service. But, while running the dough through the machine, Soda Pop mangled it. A pasta machine has several settings, with the highest number being the largest space between rollers and the lowest number being the smallest space. When you first start running the dough through the pasta maker you start with the highest number. Then, as the dough is being rolled out and thinned by the machine, you decrease the setting. The process is gradual. If you start by rolling the dough through a low setting, you will likely shred it. When you’re done, you must always turn the dial back up to the high setting so the next person doesn’t ruin his dough. No one had done so on this machine and Soda Pop was left with a mess of green dough. Having no more spinach dough to make the ravioli with, Soda Pop started rolling out the egg dough. Within a few minutes, he had enough ravioli to get us through the night.

I knew that Didier had seen what was going on; he didn’t miss anything. When I sheepishly sent the first egg dough ravioli to the pass, Didier immediately yelled out to me, “Doog, qu’est-ce que c’est cette merde?” (Doug, what is this shit?) He pushed the plate with his hand and nearly knocked it off the pass. His face was bright red. Just then Ducasse, who was with us in the kitchen that night, started yelling his name from another side of the kitchen. I had gotten used to hearing Ducasse yell “Didier” rapidly about ten times whenever he was unhappy with something. That was his way of calling him over to fix something, which could be anything from how a cook was searing a piece of meat to how something was being stirred. The new ravioli were reluctantly used and we were off the hook. They weren’t spinach dough ravioli, but at least we had some and we hadn’t lost more than a beat or two.

What Soda Pop couldn’t have anticipated -- mostly because I was keeping track of the order tickets for the station -- was that we were about to run out of spaghettini as well. So, when he used the egg dough for the raviolis, he had closed the door on making new spaghettini. Telling Didier that we had no more spaghettini for the night, as Ducasse stood right behind him, was easily the hardest thing I had yet to do in that kitchen. I knew nothing about tension before I stood in front of the two of them as they exchanged several heated words in French. I stood there waiting for a verdict. While I hadn’t been directly involved in the pasta fiasco, I was running the station and so this was my fault. A handful of painful minutes later, an alternative was found -- we would stretch each portion of pasta that we had and add more veg to the dish -- and we moved on. After service that night, Brendan was very apologetic and thanked me for being in the middle of the fire. I understood what he meant, but all the same I told him his thanks weren’t necessary. We had withstood the heat, so it really didn’t matter as long as we did better the next time. Regardless of how hard you try, how much work you put in, everything can’t always go as planned. Later that night, when the newspaper came out, the news we were awaiting only confirmed that fact.

+ + + + +

This is the second of three excerpts. Part one is here. Doug Psaltis is the Executive Chef of Country Restaurant, which will open soon in New York City. He has cooked in some of the world’s finest restaurants and with some of the most acclaimed chefs. Michael Psaltis is a literary agent in New York City. He works with both fiction and nonfiction authors through his own literary agency, and also heads up a division of Regal Literary that is dedicated solely to food writers and cookbook authors. Copyright © 2005 by Doug Psaltis and Michael Psaltis. Excerpted by permission of Doubleday Broadway, a division of Random House, Inc. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.

maggiethecat · September 19, 2005

For the first few days running the entremet station, I’d hear Didier call an order and I’d have to pause as I tried to figure out what he said before reacting. Olivier stayed near me most of the service to translate. “Hey, homard, lobster, vas-y,” he’d say. I was constantly behind, almost unable to keep up. I had to pick up French right away -- à la minute -- or I’d be gone. So, I took one of the first printed menus home. I translated as much of it as I could (mostly by calling Greenie, who had returned to Los Angeles after giving up on a small restaurant in the West Village that he helped run), and after the first service I had learned at least the main ingredient in every course that I was involved in. I went back to work the next day ready to listen for homard, petits pois, poularde, and more. But the menu had changed and only a few of the words I knew remained useful. I was back to memorizing in the time between work and sleep

I have to give you all the credit in the world for this proactive French immersion course -- there's nothing like necessity to provide the fire in the loins to learn a new language. Were you never previously in a position where you had to learn kitchen French?

project · September 20, 2005

What Soda Pop couldn’t have anticipated -- mostly because I
was keeping track of the order tickets for the station -- was
that we were about to run out of spaghettini as well. So, when
he used the egg dough for the raviolis, he had closed the door
on making new spaghettini. ...
 
A handful of painful minutes later, an alternative was found
-- we would stretch each portion of pasta that we had and add
more veg to the dish -- and we moved on.
←

So,

<UL>

<LI>

Ran out of raviolis.

<LI>

In the effort to make more raviolis, a mistake with a machine

wasted some dough.

<LI>

Ran out of dough.

<LI>

Ran out of spaghettini.

<LI>

Since was out of dough, couldn't make more spaghettini.

<LI>

Making more dough would take too much time.

<LI>

Found an alternative.

</UL>

Okay.

Actually, in a draft of a comment to Part One, I wrote but did

not post:

Capacity Planning.

How many loaves of French bread should the restaurant order?

The bread doesn't keep well at all.

So, ordering too much bread results in waste.

But ordering too little bread can result in running out during

the day and disppoint customers and hurt revenue and earnings.

French bread is just an example; there are many cases of how

much to purchase or prepare, and then the cases can interact.

Good answers here could help earnings.

</blockquote>

So, ADNY running out of dough was suprising to me only in that

it could actually happen at ADNY!

So, ADNY had a problem in "capacity planning" but with dough,

spaghettini, and raviolis instead of French bread.

I'm not surprised that a restaurant would have this problem; I

am surprised that ADNY would.

But the problem was at least a little bigger than just simple

capacity planning of how much of some one item, e.g., French

bread, to order:

So, for the evening service, need some materials, e.g.,

dough, spaghettini, and raviolis.

Then some of the spaghettini and raviolis need to share the

same supply of dough.

I don't know just what is in the dough, but it likely at least

has some flour, of some type, in it.

Given dough, the spaghettini and raviolis can be made quickly,

but making the dough takes time, apparently too much time to

be done during the evening service.

So, before the evening service, there must be some

planning for the requirements for the dough.

So, we have a case of what for some decades in some industries

has been called material requirements planning (MRP).

This particular MRP problem has only a few materials -- flour,

dough, spaghettini, and raviolis; but it is easy enough to see

that many manufacturing processes may have many more.

Actually, ADNY must have many such MRP problems.

Since many of the problems share flour, eggs, butter, salt,

milk, etc., the problems are related and not all independent.

If the problems were independent, then the whole could be

decomposed into separate smaller problems, and that could ease

the planning.

But, at least to some extent, there is just one big ADNY MRP

problem and not many small ones.

Small or big, ADNY has some MRP problems.

Of course they do.

Easy enough to see that nearly any restaurant, factory,

process plant, etc. has MRP problems.

Some of the problems are complicated.

But, the flour flows into the dough, and the dough flows into

the machines and into the spaghettini and raviolis.

So there are some flows.

Each single end to end flow looks like a chain.

There is some jargon, then, to say that the planning is at

least a supply chain problem.

In each chain, there are:

<UL>

<LI>

Time Delays.

Time to make spaghettini and raviolis from dough appears to be

nice and short, but time to make dough or to get a new supply

of flour is relatively long.

So, the quantities of spaghettini and raviolis may not have to

be planned in advance, but the quantities of dough and flour

do.

<LI>

Shelf Life.

A sack of flour has a long shelf life.

Likely the shelf life of dough is much shorter, and spaghettini

and raviolis, each very short.

So, mostly can stock up on flour; then the cost would be

capital and storage space.

These can be significant and have been the motivation behind

just in time deliveries (which in general are risky),

but capital and storage space (even at the price of New York

City floor space!) for flour, etc. may not be very significant

at ADNY.

If stock up on dough, spaghettini, or raviolis, then can get

too much and can have some waste, in labor, materials, etc.

<LI>

Capacity Limits.

There are some limits on capacities, e.g., how fast the

machine can make raviolis.

<LI>

Uncertainties.

The operation is just awash in uncertainties.

Handling these is likely the biggest challenge.

One example of uncertainty was the mistake in using the

machine to make raviolis, the mistake that, then, wasted dough

and caused a dough shortage.

Other uncertainties can be in staff illness, accidents, what

the customers order, late vendor deliveries, the weather, etc.

</UL>

So, there are supply chain problems with uncertainties.

When things do go wrong, then sometimes there are some

alternatives.

Such decisions are sometimes called recourse.

So, there is a system that evolves over time with various

costs, limitations, and uncertainties and that has some

decisions with recourse.

Hmm ....

Some of the customers at ADNY will be familiar with such

things in parts of finance.

What to do is known fairly well in general terms and in many

details.

Of course, for any good restaurant, and certainly for ADNY,

the basic restaurant planning has already done good work on

having the front of the house, back of the house, number of

staff in each category, standard supplies, etc. all nicely

balanced, i.e., given the size of the front of the house,

plenty, but not always a gross excess, of everything else.

Further, for the planning for each day, there are likely some

easy approaches with good effectiveness.

E.g., since ran out of dough, from now on each day have a

little more.

Also put on a list to remind the staff always to check the

settings on the machine and, thus, avoid ruining dough.

Wasting dough is bad; running out of dough is terrible.

But, more should be worthwhile.

Really, for planning that is important and has to be done

continually, there should be some fairly solid -- efficient,

reliable, etc. -- processes in place.

For this, a first-cut might be:

<UL>

<LI>

Basic Data.

So, start with the basic restaurant planning data on size of

the front of the house, size of the back of the house, numbers

of staff, etc.

<LI>

The Menu.

Likely some menus will need more dough than other menus.

So, in planning the dough, etc., will want to consider the

menu.

<LI>

The Customers.

For the number of customers, reservations reduce the

uncertainty here.

Otherwise, there are some ways to use historical data to

estimate the numbers of customers.

For a given number of customers, there is still the

uncertainty of what they order.

</UL>

Then, with such considerations in hand, should be able to

evaluate what the results would be.

Then, if there is too much chance of running out of dough,

should be able to see this.

And should be able to see how much dough would need in order

to have small enough chance of running out.

Such first-cut planning should help avoid situations such as

running out of dough and having to apply the recourse

of changing what is served, or worse yet, having to tell a

customer that an item on the menu is not available.

I'm surprised that there is no standard software for such

restaurant daily operations and planning.

So, with such software:

<UL>

<LI>

Process.

Could get a process that organizes the work and makes

it more reliable.

<LI>

Historical Analysis.

Could get a means of recording data useful for analyzing what

has been happening over time and, thus, doing better planning.

E.g., when a Maine lobster dish and a Dover sole dish were

both on the menu, what fraction of customers ordered [A]

lobster or sole?

What menu items resulted in selling what wines?

When significant snow was on the ground, what happened to

business?

<LI>

Menu Item Capacity Planning.

For each menu item, balance the cost of [A] running out during

the service and having too much and wasting the excess --

could be of enormous importance for many restaurants; could be

of value for ADNY in the case of highly perishable very

expensive ingredients.

<LI>

Quanties to Order -- MRP.

For a given candiate menu, determine quantities needed for

dough, Dover sole, Maine lobster, rack of lamb, filet of

venison, Boston lettuce, butter, cream, eggs, milk, etc. for

acceptably low chances of running out.

<LI>

Work Levels.

For a given candidate menu, in one minute intervals during the

service, find how busy is each machine, appliance, person,

station.

</UL>

Carrot Top · September 20, 2005

:laugh: Ei yi yi, project. I think that you have such a deep interest that you might want to really get a job in a restaurant.

And luckily for the world, restaurants (though not usually ADNY) do have lots of openings, often, for those who want to work and who want to apply themselves.

Good restaurants do have a number of planning systems. A great number. Again, I will refer you to Amazon under the Professional Books. Seek "Restaurant Management and Operations" and you will find them.

There are well-thought out policies and procedures that cover each tiny element of detail that happens. One of the driving systems of the kitchen is a Production Schedule, which is made up based on anticipated needs which are based on knowledge and experience. Which I would guess they have at the restaurants mentioned in the book above. There is also menu engineering, pricing structure, food and labor costs, etc. etc. Running a restaurant is a complex undertaking but people do know how to do it.

But guess what. Life happens. People walk in the door to be fed, and people are all different.

Nothing is completely stable, ever, in the restaurant business. And software, and systems, will never be able to iron it out.

It's just that darn aggrevating human element that messes things up, you know? :biggrin:

For goodness sakes, if people would just do as anticipated, the thing would run right ship-shape.

People. Life. Food (which in its own way is also alive).

They simply refuse to be totally controllable. Unless you aim to be a McDonald's and even then I imagine that sh** happens.

And that, makes for creativity often. . .and it makes for challenges in the kitchen often. . .and it makes for a day where at the end of it anyone who has faced these challenges which will never be finally captured before they happen. . .it makes these people go home knowing that they used their brains and training well, to put out a good meal that would make people happy.

Life. It happens.

And when life happens, and it can not be pinned down onto a production schedule nor a written menu, this is where writings such as these stories above are useful.

They show the reality vs. the planned. They show how one can manage it and manage it well, if one does feel they need a useful reason beyond entertainment for the reading.

P.S. I keep having this feeling of deja-vu in these discussions with you, and finally I realize what it is. I feel like we are those two movie reviewer guys on TV who keep taking completely different viewpoints on everything. It works for them.

Maybe we should go on the road. :laugh:

Edited September 20, 2005 by Carrot Top (log)

Shalmanese · September 20, 2005

I'm not surprised that a restaurant would have this problem; I
am surprised that ADNY would.
←

Lots of restaurants use SCM software to a certain extent but it's not as useful as you think.

1. SCM doesn't scale down well. It's a major pain to enter into the computer every morning what the weather is like and whether your restaurant was reviewed and a whole bunch of other factors that affect your turnout. If your changing menus every day or week, then you also need to enter that into a computer every time.

2. The variation is huge. It's not uncommon to get an entire dining room which, for no reason all of a sudden discern they are enamoured with a usually, not very popular dish. The quality of your forecast is inherently limited by the variation in your data. Theres no point in using more techniques more sophisticated than your noise level.

3. Restaurants have an amazingly efficient way of dealing with excess capacity in the form of "daily specials". As a result, theres not much incentive to pare stock down to the bone.

4. Small restaurants just don't have enough data to do proper SCM. You really need to be shifting in the order of a million units a day before the benifits really start kicking in.

It's certainly used but there are very good reasons why sh*t still happens all the f*cking time.

Nentony · September 20, 2005

I read the two excerpts Sunday and read with interest the discussion of part one. Projects way of looking at things is so different than mine and I was thinking about it on the way to the bookstore. When I got there the bookseller had my copy(I had called ahead) and another new release she thought might be of interest. It's called Don't Try This At Home, Culinary catastrophes from the worlds greatest chefs. LOL. I thought, Project would hate this book.. Anyway, Thanks for the interesting discussion.

Tony

Carrot Top · September 21, 2005

I read the two excerpts Sunday and read with interest the discussion of part one. Projects way of looking at things is so different than mine and I was thinking about it on the way to the bookstore. When I got there the bookseller had my copy(I had called ahead) and another new release she thought might be of interest. It's called Don't Try This At Home, Culinary catastrophes from the worlds greatest chefs. LOL. I thought, Project would hate this book.. Anyway, Thanks for the interesting discussion.
Tony
←

Yes, that "Don't Try It At Home" looks fun, doesn't it. :biggrin:

Just noticed that this was your first post, Tony, so hello hello and welcome! :smile:

Doug Psaltis · September 21, 2005

For the first few days running the entremet station, I’d hear Didier call an order and I’d have to pause as I tried to figure out what he said before reacting. Olivier stayed near me most of the service to translate. “Hey, homard, lobster, vas-y,” he’d say. I was constantly behind, almost unable to keep up. I had to pick up French right away -- à la minute -- or I’d be gone. So, I took one of the first printed menus home. I translated as much of it as I could (mostly by calling Greenie, who had returned to Los Angeles after giving up on a small restaurant in the West Village that he helped run), and after the first service I had learned at least the main ingredient in every course that I was involved in. I went back to work the next day ready to listen for homard, petits pois, poularde, and more. But the menu had changed and only a few of the words I knew remained useful. I was back to memorizing in the time between work and sleep
I have to give you all the credit in the world for this proactive French immersion course -- there's nothing like necessity to provide the fire in the loins to learn a new language. Were you never previously in a position where you had to learn kitchen French?
←

French terms were used in just about every serious restaurant that I worked in before joining ADNY, but this wasn't just some terms. The entire service was run in French, and most of the cooks in the kitchen came directly from France. Even the dishwashers spoke French. It was a challenge.

project · September 21, 2005

Carrot Top, Shalmanese,

No, it's not that way!

What you are objecting to, I didn't say or mean!

The main theme in your comments is that there is a lot of

what I called uncertainty in restaurant operations.

Right.

Of course there is.

And, I addressed this issue, although briefly.

Gee, guess what I wrote wasn't long enough!

Here is where it appears you are getting off the track:

It appears that you are assuming that I would want to be able

to say to a restaurant owner:

Four times last week you ran out of important supplies during

the shift and had to tell customers bad news.

That cost you.

And, 29 times last week you had to throw away supplies that

were too old to use -- i.e., past their shelf lives.

That cost you, too.

That was a lot of cost -- waste.

Via some planning software, we can eliminate all that waste.

</blockquote>

So, you seem to be assuming that I would need to be able to

say the above or nothing at all.

No!

Clearly I could not deliver on the claim of "eliminate all

that waste".

I would not make such a claim.

And I would not need to.

You want to assume that, then, there is nothing to be done.

No!

There is still plenty to be done, plenty of savings to be

obtained!

Basically you are assuming that the only progress for such a

restaurant would be to plan with exact calculations and, then,

in practice execute that plan exactly or nearly so.

Sure, that might work for some school feeding 2000 captive

obedient students each day!

Since the uncertainty does not permit executing a prior fixed

plan exactly, you want to conclude that nothing can be done.

Uh, this is a common mistake.

Some very serious people in some very serious situations have

made this mistake.

Still, it's a mistake, possibly a very serious one.

For a simple way to see that there is a mistake here, we need

only notice -- as both of you have been emphasizing -- that

restaurants around the world, using just intuitive manual

methods, work in situations of enormous uncertainty everyday.

Easy conclusion:

Working with uncertainty really is possible.

Then you are in effect jumping to conclude that these

intuitive manual methods are the best that can be done working

with uncertainty.

Think about it:

Do you really want to draw this conclusion?

Of course you don't!

But, what to do, how to work with uncertainty better than

intuitive manual methods?

Not many people have gotten into such things.

Well, via the largesse of the US DoD, your tax dollars have

paid for long wide deep streams of research on this question.

You see, the US DoD realizes that there is a lot of

uncertainty in national security.

In particular, the DoD believes about the mostly deeply in

"sh*t happens", "the fog of war", and "no war plan ever

survives the first contact with the enemy", and much more on

uncertainty.

In particular, at times the US DoD has been big on how to

handle uncertainty in various operational contexts

plenty general enough to cover, say, restaurant operations.

Here is an example:

During the Cold War, a claim was that the the Soviet forces

would not be able to find the US nuclear powered ballistic

missile firing (SSBN) submarines as they were operating at

sea.

Well, clearly this claim cannot be literally true:

In principle it would have been possible to get a big net and

drag really quickly!

I won't go into more (I could, but then I'd have to ...).

But, it was clear enough that for the Soviets to find all of

the US SSBN's at once would be somewhere between impractical

and impossible.

And they would definitely want to know just where each and

every one of those SSBN's was, very, very definitely,

before attacking even one of them with a nuclear weapon!

But, there was a concern about another scenario -- a

good George C. Scott impression would go well here, along with

some good Peter Sellers in a wheel chair!

See, if there really were nuclear war with the Soviets and if

they really did find a US SSBN, then easy enough to get people

to agree that can scratch one SSBN.

That is, in nuclear war, the only real safety of an SSBN is

not to be located!

If they cannot all be found at once, maybe they could be found

slowly one at a time and then destroyed as they are found?

Further, maybe somehow the world could get into a situation of

global nuclear war but limited just to sea?

Then the SSBN's could be picked off slowly one at a time.

If so, then how long might they last?

At times, this was regarded as a "controversial" scenario!

But, suddenly the US Navy very much wanted an answer.

Yup, they did.

Your tax dollars hard at work keep you safe!

At the time, I was working in such things supporting my wife

and myself through graduate school.

The Navy wanted their answer in two weeks.

No one much saw what to do.

I had some ideas and got to work.

"Two weeks?".

Hmm ....

That was about right because just after that deadline my wife

had already scheduled a short vacation for us, one she very

much wanted, at Shenandoah.

So, two weeks it had to be!

Now, it wasn't just the submarines I needed to consider; it

was full nuclear war, but limited to sea.

So, I had to consider missile firing submarines, attack

submarines, surface ships, airplanes, etc., on both sides,

with everything on each side potentially shooting at

everything on the other side.

That's a lot of shooting!

No way would it be possible to calculate exactly when each

shot was fired or its result or even when each SSBN got shot

at!

Uncertainty!

And, I had two weeks!

How to do that?

Well, I was not the first to consider looking for little tiny

submarines in great big oceans!

In WWII, a guy named Koopmans (I don't recall his first name)

wrote a study called OEG-56 where he said that, give me the

area of the ocean, the speed of the target submarine, the

speed of the search submarine (surface ship, airplane, etc.),

and the distance at which the searcher can detect the target

and then I'll tell you the probability distribution of the

time until the next detection!

Moreover, that distribution will be in the family called

exponential.

Well, Koopmans was not asking for much information, and that

fit well because no one had much information!

Or at least what Koopmans was asking for was about all I had

time to work with!

Then, make a few mild assumptions, use some work of A.

Kolmogorov (right, a Soviet mathematician -- I'm not making

this up!), tap lightly, and what falls out is a continuous

time discrete state space multidimensional pure death

(that's what to call it!) Markov (right, a Russian) process

subordinated to a Poisson process (can sometimes get these in

French restaurants).

As Kolmogorov knew, the probability distribution for the

number of US SSBN's can be found directly from an exact

calculation in terms of a matrix exponential.

However, in this problem, the matrix would have been

absurdly large.

So, do what S. Ulam did -- to Monte Carlo we go!

A few days of writing software, and I was done!

The Navy got their results, and my wife got her vacation, both

on time.

My work was later sold; I could tell you who bought it, but

then I'd have to ....

But, what I found was the probability distribution, not the

exact situation.

So, it was all working with probability distributions, not

exact results.

Net, it really is possible to do a lot with uncertainty.

But, as both you have emphasized, all during the evening

service at a restaurant, the staff is reacting at the last

minute to suddenly revealed uncertainty.

What would be the best way to do that?

Again, the work is heavily based on working with probability

distributions.

As, I mentioned, a lot is known.

It's called "stochastic optimal control" or "Markov decision

processes".

A short but okay description of the field is, how to make the

best decisions over time under uncertainty.

Yes, there is some at Amazon!

Consider E. Dynkin (at Cornell, student of Kolmogorov), D.

Bertsekas (at MIT), S. Shreve (at Carnegie-Mellon, Bertsekas

student), R. Rockafellar (University of Washington), W.

Fleming (Brown University).

And consider the ORFE group at Princeton.

Uh, won't find any VEFE in those sources (oh, poor VEFE, poor,

poor misunderstood emotionally abused passionate poignant

poetic VEFE!).

Curiously likely some of ADNY's customers from Wall Street

understand at least some of stochastic optimal control, if

only because the Black-Scholes option pricing model is a very

simple special case.

Uh, they have not always known as much as they should have;

the Long Term Capital Management (LTCM) disaster was a result

of not really understanding.

But, at LTCM some economists were involved, and how to believe

that an economist could understand such things!

In principle stochastic optimal control would definitely be

the gold standard for how to run a restaurant!

Uh, those books on restaurant management, they covered that,

right?

But, I didn't leap to say to apply stochastic optimal control!

What I suggested was mostly quite simple!

Actually, I did include a role for Monte Carlo as in the

evaluation of the survivability of the US SSBN fleet!

Such Monte Carlo is not always too difficult to do:

I did if for global nuclear war at sea in two weeks!

Also I didn't recommend applying supply chain

management (SCM), supply chain optimization, or the

advanced onerous but coveted multi-echelon supply chain

optimization!

This last has to be essentially a very narrow special case of

stochastic optimal control.

Supply chain management software, e.g., from SAP, Manguistics,

Oracle, might be okay for GM or Boeing and might actually in

principle be able to cover restaurant operations, but I did

not recommend just calling up your local SAP office!

Monte Carlo is not so bizarre:

It's been a standard part of spreadsheet software for at least

a decade.

I do remain surprised that ADNY would run out of dough!

Carrot Top · September 21, 2005

Well, okay, project. :biggrin:

You have definitely left me speechless this time.

...............................................

Except for one note.

If computer systems did manage to control everything, life would be very boring.

And fewer "real" creative opportunities would arise if there were no longer any problems to be solved. Creative opportunities within the real metier, not within the computer field.

There's many a delicious dish that has come about from the serendipitous, which is part of the flip side of total control.

And the drive that leads one to create something that is "above the usual" generally does not start with the idea of control or of money.

It starts with a love of beauty and surprise. Not the final dollar that is to be made.

....................................................

But of course one would have to have actually cooked for a while to realize that, maybe. And not be scared to color outside the lines while doing it.

:rolleyes:

Edited September 21, 2005 by Carrot Top (log)

Carrot Top · September 21, 2005

Doug,

Can you tell us, what were your "reasons" for writing the book?

Thanks.

Edited September 21, 2005 by Carrot Top (log)

project · September 21, 2005

If computer systems did manage to control everything, life
would be very boring.
And fewer "real" creative opportunities would arise if there
were no longer any problems to be solved. Creative
opportunities within the real metier, not within the computer
field.
There's many a delicious dish that has come about from the
serendipitous, which is part of the flip side of total
control.
←

Don't think it's that bad!

There is an example in cars:

One of the real nonsense problems in the US automobile culture

was fuel flow and ignition timing.

These were done with poor accuracy, especially with cold

engines, and the result was wasted fuel, dirty air, worn dirty

engines, too much in maintenance, etc.

Then about 20 years ago we got computers controlling fuel

flow, ignition timing, etc.

It was terrific progress on wasted fuel, etc.

But, that use of a computer to control fuel and ignition

didn't ruin the creative or enjoyable aspects of using a car!

Sure, in some science fiction scenario, computers will be so

advanced and powerful that humans will be left just holding on

to the power plug as the last means of control over the

computers!

But that's a very long way in the future!

There is quite a bit of computer control now.

E.g., the Internet we are using to connect with eG basically

depends on computer control, say, in the packet routing

algorithms.

The Internet is being tough on the old circuit switched

telephone network, but without the Internet eG would have to

be, say, just an old bulletin board system, and those

were so clumsy to use they just were not effective enough to

catch on.

The role of a computer I was outlining in restaurant

operations would say how much dough to have on hand, etc., at

the start of the evening service.

This computer role would not much reduce "'real' creative

opportunities".

E.g., what to do with the dough, what sauces on the spaghetti,

what to have in the sauces, what wines to pair with the dish,

etc. would still be from the existing creative means of ADNY.

Your "total control" sounds regimented.

Do you feel regimented driving a car that has a computer

controlling fuel flow and ignition timing?

The Ford Model T had ignition timing and fuel flow

choke controls for the driver; do you feel that handing

these over to a computer was a loss?

Do you use computer spell checking?

How about Google; last I heard, they had 125,000 computers.

You know, essentially all of eG is on Google!

There gets to be some question why the data should also be on

eG's servers!

ADNY is using a lot of control systems now:

With all their terrific equipment, they will have some

terrific refrigeration, and that will have a control system

for temperature, humidity, and frost removal.

Similarly for their heating, ventilation, air conditioning

(HVAC) system.

In much of the electronic equipment they have, there is a

power supply that converts standard wall alternating

current to direct current.

This thing has to be a control system because its main job is

to keep the direct current at a specific voltage whatever the

load -- and the load varies quickly from no load to full load

and back frequently.

The new fancy low temperature cooking, possibly being used by

ADNY, needs a control system to keep the temperature quite

precise.

Certainly the ADNY bread ovens have control systems for the

temperature and maybe the humidity.

The standard wall alternating current is delivered with a

control system.

Similarly for the pressure of the natural gas used in the

stove tops (assuming ADNY has gas burners -- I would guess

that they do).

Sure, there should be the "serendipitous" in cooking, with as

much "beauty and surprise" as you want!

Or, if "beauty and surprise" are the goals, then so be it!

Gee, a car with computer controlled fuel and ignition might

get used for driving in Upstate New York to look at the Fall

leaves; there is no conflict here!

At a restaurant with a lot of creative activity, a good system

based on stochastic optimal control for making better

restaurant operation decisions under uncertainty should notice

that the quantities of ingredients -- from all those creative

efforts -- were highly variable and then, roughly, have higher

stock levels to reduce the chances of running out.

So, the chefs deglaze with red wine, reduce, add some red

current jelly (at some restaurants, maybe not ADNY!), toss in

butter, etc., all in variable amounts, and with a good system

it seems that somehow there is always plenty of red wine, red

current jelly, butter available but not a gross excess!

For

But of course one would have to have actually cooked for a

while to realize that, maybe. And not be scared to color

outside the lines while doing of it.

</blockquote>

you're preaching to the choir here!

Anyone proposing stochastic optimal control for restaurant

operations is definitely not "scared to color outside the

lines"!

But, you do touch on what likely is a serious point about high

end restaurant cooking:

Gee, in my cooking, once I get a dish the way I like it,

usually after what I regard as far too much in time,

money, and effort, on a computer I type in what I did.

Then when I do backup for that computer, I "get happy, happy"!

My main goal is not to create the dish but to eat it!

So, sure, a restaurant could do something similar:

Once the chefs have had a good creative period and

gotten a good dish, they can record it, teach it to the

relevant staff, and have it available to put on the menu

whenever that seems appropriate.

But, maybe some restaurant chefs actually do not want

to do this and, instead, want always, always, to be

creative for every dish served -- never write down what

they did, every serving like a jazz improvisation, and no two

servings ever just alike!

Okay, if that's what they want!

And, in that case, stochastic optimal control should do well

having on hand all the novel ingredients they might want to

use!

But, should mention, that in such an application, at some

point will have to ask the restaurant owner:

About that sushi quality tuna:

It's on the list.

We're keeping it on hand.

It's been used for some creative dishes, but the last use was

seven months ago.

It doesn't keep very well; the kitty cats have been enjoying

it!

Uh, could you mention a number, what you regard as the

cost, in US dollars, for an instance when a chef

suddenly asks for this tuna and it is not on hand?

If this cost is $500,000, then there will be plenty of tuna on

hand; if this cost is $0.50, there will be much less tuna on

hand at least until the chefs start to use it again.

If you want to try some different numbers, then for each we

can give you reasonably good estimates for what it will cost

for supplies for the next three months.

The estimates will be especially good for what you would be

looking at here, the comparative cost.

It's your judgment.

</blockquote>

Fairly generally, that's about how cost and creativity balance

out.

The issue really isn't regimentation or staying within the

lines, but there are some costs.

I have a view on this; might as well type it in since there is

no chance I can type too fast for Jason's servers!

Can take a family of four where both parents work 80 hours a

week at $10 per hour.

So, they gross $80,000 a year, and, for a family of four, and

commuting to possibly four jobs, we are sure that they can

spend it!

Now, if we can use computers to get productivity up by a

factor of 10, then one parent can stay home with the kids, the

other parent can cut back to 40 hours a week, they can gross

$200,000 a year, and again we can be sure that they can spend

it!

Heck, they may even go to ADNY once a year!

If we can use computers to get productivity up by another

factor of 10, then they can gross $2 million a year, have high

end private schools for the kids, an 8000 square foot house,

nicely furnished, a vacation home in the country, and a 40

foot boat for fun on the water.

If we can use computers to get productivity up by another

factor of 10, then they can have a nice yacht, a 20,000 square

foot house, a really nice place in the country, Ivy League

educations for the kids and after a few years retire and

pursue creative activities in cooking, music, painting, etc.

If we can use computers to get productivity up by another

factor of 10, then they can join with others in launching,

say, a telescope with seven mirrors, each mirror 100 yards in

diameter, the mirrors separated by 10 miles, to one of the

Lagrangian points where it can have a stable home and quietly

calmly look into the clouds on distant planets, etc.!

So, we are up to four factors of 10, a factor of 10,000.

This means that a person can accomplish in 12 minutes what

takes a 2000 hour working year now; we're talking a lot

of increase in economic productivity!

Okay, set aside that telescope and consider just three factors

of ten:

That means a person can accomplish in two hours what takes a

2000 hour working year now; we're still talking a lot

of increase in economic productivity!

And, you do want Ivy League educations for the kids,

right?

So, how are we going to accomplish in two hours what takes

2000 hours now?

About the only way we have in mind is just to automate

everything in sight.

So, we will want people managing computers managing computers,

..., managing computers doing the work.

Some of what these computers will be doing is stochastic

optimal control.

What the humans will be doing is creative activities in

cooking, music, painting, etc.

But, to get there, we have a long way to go.

Carrot Top · September 21, 2005

At this point I think our discussions are bordering on the rude, considering that the subject here is supposed to be the excerpt from the book.

If the entire thing is deleted, I won't be too surprised.

Two final comments though.

I prefer to drive a car with an manual transmission. German or Italian, please, for then you can feel the road. Chokes are fun, too, particularly on boats and lawnmowers.

And standardized recipes are in standard use in fine restaurants.

Let's stop, project, unless it has to do with the book. It's not right.

Shalmanese · September 21, 2005

Carrot Top, Shalmanese,
 
No, it's not that way!
What you are objecting to, I didn't say or mean!
 
The main theme in your comments is that there is a lot of
what I called uncertainty in restaurant operations.
Right.
Of course there is.
And, I addressed this issue, although briefly.
←

Right, OK, you certainly seem to know your stuff (or, at least managed to name the right names. PM me the actual math if it's not classified because I'd be interested to look over it). However, the question you were asked to solve is not the same problem that ADNY faces. ADNY can't order a poisson distribution of dough, ultimately, it has to condense down to a single estimate.

And you still haven't addressed the other points I raised, namely:

1. It doesn't scale down well. The information gathering resources are so onerous it overwhelms the benifits.

2. You cannot get below the noise floor. No matter what you do, there is no theoretical way to get below the noise floor. Even if you knew the perfect distribution ahead of time, it's impossible. And since the noise floor is very high, theres very little meaningful information you can extract with any method.

3. Theres not much point anyway. Everything can be dumped off as specials.

4. You don't have the information neccesary to build a statistical model. A dish may be on the menu for a month or so. You can't do anything meaningful with 30 sample points.

Think about airline bookings, the airlines essentially have the same problem, they need to book enough seats so as to fill a plane yet not so much that they overbook. They can run a billion statistical tests over it yet overbooking is still reasonably common. What makes you think a restaurant can do any better?

project · September 22, 2005

At this point I think our discussions are bordering on the
rude, considering that the subject here is supposed to be the
excerpt from the book.
If the entire thing is deleted, I won't be too surprised.
Let's stop, project, unless it has to do with the book. It's
not right.
←

Of course we actually are on topic:

I am accepting the claim made in the thread about Part One

that the point of the book was to provide information about

restaurant operations, so have been discussing restaurant

operations.

The central problem in Part Two was running out of dough, and

this is a good example of an important issue in restaurant

operations -- having enough without having too much. E.g.,

it's a problem in inventory control.

The relevant research on such questions in operations is

called, appropriately enough, operations research, and

what I have been describing are some of the most solid parts

solidly in the middle of operations research.

E.g., I did mention the Princeton ORFE group, and there the

"OR" abbreviates just operations research.

Will bet you dinner at ADNY that the ORFE group -- e.g., E.

Cinlar -- would confirm that an important problem in

operations research is inventory control, especially with

uncertainty. [Disclosure: Uh, one of Cinlar's best students

was one of my professors!] Want to take me up on this bet?

You'll have to bring your checkbook. Be careful -- ADNY will

set you back.

In particular, I mentioned Poisson processes: These

are crucial and powerful in inventory control and, more

generally, for estimating arrivals -- e.g., when the next

order for Poisson, or dough, will reach the kitchen at ADNY.

Cinlar has one of the nicest treatments of Poisson processes.

As a professor, I taught operations research, including

Poisson processes, in one of the better MBA programs; that

these topics are crucial for operational planning in business

is so solid that granite looks like pastry cream. Once, as an

expert witness in a legal case, I applied the material to

hospital operations, e.g., estimating when they would run out

of beds -- close enough for you? If the books you mentioned

on restaurant management did not make clear the importance of

this material, then you have learned something important on

eG, have made some professional progress, and are now ahead in

the restaurant industry. You will learn nothing of higher

quality about restaurant operations anywhere. An MBA will

cost you; you just got for free some of the highest quality

material there is in any MBA program. That this material is

important for business planning, someday possibly including

restaurant operations, is just rock solid -- totally beyond

any question. We're talking crown jewels of civilization here

-- it doesn't come any better than this, not in this solar

system. E.g., it's fully respected at Princeton, and they are

fully correct.

Supply chain management is also solidly within operations

research. That in principle supply chain management applies

to restaurant operations is as true as apple pie. In

particular, the core of supply chain management is just

inventory control. Further, part of the Manguistics software

is just C-PLEX which is optimization, by R. Bixby at Rice

University, and one of the best pieces of operations research

software.

There isn't a lot on applications of operations research to

the operation of restaurants, so my operations research

examples, with lessons for restaurant operations, were from

other fields. My mention of Monte Carlo simulation to

evaluate plans in restaurant operations is one of the most

appropriate suggestions for progress in planning in restaurant

operations. Huge literature shows solidly that Monte Carlo

simulation is one of the best tools in planning and sizing

facilities operating under uncertainty, including for reducing

chances of things going wrong; that this methodology would

never be helpful in restaurant operations is absurd.

Uh, you recall, commonly a restaurant kitchen has a queue of

tickets that have arrived from the front of the house, a queue

of dishes from the kitchen not yet carried to the tables, etc.

Restaurant operations are awash in queues. Monte Carlo and

Poisson processes are among the most important -- likely the

two very most important -- tools in analysis of systems of

queues. Uh, I kept it simple and omitted swindles, importance

sampling, generation of random numbers! In your restaurant

career, you just made some professional progress.

Want some real and solid progress on restaurant operations?

Well, this is what some of the best of it looks like, and it

is NOT in Escoffier!

What I said about economic productivity is important for the

restaurant business: Broadly in the economy, more in

automation is coming. One result will be that the comparative

cost of traditional bench work by craftsmen, artisans,

and artists will rise quickly. That is, even if we do get a

factor of 10 increase in productivity, not all prices will

remain in the same proportion; instead, some things will get

comparatively cheap and some, comparatively expensive. E.g.,

the DVD player in my new computer just failed. It's within

warranty; maybe I can get it replaced. But, I can buy a new

one for $20. How far will $20 go at ADNY? A meal at a place

like ADNY, a loaf of handmade French bread, a piece of

handmade goat cheese will become comparatively more expensive.

Ask a reference librarian what Google's 125,000 computers are

doing to their careers. Fewer and fewer people will be able

to pay for, or do, such bench work. The future is more and

more in driving automation. Including in the restaurant

business. Net, you just made some more professional progress.

We very much are on topic, and any conclusion

otherwise would be naive and shortsighted. With the Internet,

this really very much is a Brave New World, but meeting

the challenges is what the stuff between our ears is for!

I'm still surprised that ADNY ran out of dough!

Carrot Top · September 22, 2005

In particular, I mentioned Poisson processes: These
are crucial and powerful in inventory control and, more
generally, for estimating arrivals -- e.g., when the next
order for Poisson, or dough, will reach the kitchen at ADNY.
Cinlar has one of the nicest treatments of Poisson processes.
As a professor, I taught operations research, including
Poisson processes, in one of the better MBA programs; that
these topics are crucial for operational planning in business
is so solid that granite looks like pastry cream. Once, as an
expert witness in a legal case, I applied the material to
hospital operations, e.g., estimating when they would run out
of beds -- close enough for you? If the books you mentioned
on restaurant management did not make clear the importance of
this material, then you have learned something important on
eG, have made some professional progress, and are now ahead in
the restaurant industry. You will learn nothing of higher
quality about restaurant operations anywhere. An MBA will
cost you; you just got for free some of the highest quality
material there is in any MBA program. That this material is
important for business planning, someday possibly including
restaurant operations, is just rock solid -- totally beyond
any question. We're talking crown jewels of civilization here
-- it doesn't come any better than this, not in this solar
system. E.g., it's fully respected at Princeton, and they are
fully correct.
←

Beds are not food. Beds are a stable product that sit in one place in relationship to the user.

Food comes in to the professional kitchen and it is of varying weights even though you may have ordered it one weight. Because of time limitations you may have to use it. Food comes in and some of it needs to be trimmed off more than the day before. Therefore the processes of the day require operational adaptations "a minute". Just to give two quick examples of how beds and food are in different categories of reality.

The user of the bed takes the bed as is. They do not ask for it without shallots or with a cream sauce, also, again "a minute". Another example of differences in the relationship between beds and food and users.

As far as "being ahead now" in the restaurant industry because of your information: first of all I do not care where or what you taught because I have no great respect for MBA programs "just because" we are supposed to. MBA programs to me have the same odor that VEFE has to you. With the difference perhaps being that although they have an odor to me, I realize and accept their appeal to others and do not have a problem with that at all.

As far as your informing me that Princeton is fully correct, well they may be. Or not.

All I know is that the partners of Goldman Sachs (who might have some spare change in your book of knowledge and power, perhaps a bit since they are the employers of these hungry MBA's who enter into the business world panting in anticipation of their glory to come whereby they themselves might have a shot at being a Goldman Sachs partner). . .the partners of Goldman Sachs chose me to be their executive chef when I was in this industry. They chose me because I did my job and did it well. They demanded daily that this job be done to the level that they demanded every job be done in that corporation and I must say that it is a rather high level of excellence demanded. They chose me even though I have a ninth grade education.

Did they choose the Princeton MBA? No. Why? Because the average Princeton MBA does not know how to run a foodservices operation. Because the average Princeton MBA is so wet behind the ears in terms of real life that when they enter any part of Goldman Sachs they are considered rather silly "newbies". Some of them make it there, some of them don't. Some of them go off to teach in other MBA programs, I guess.

So you will teach me of restaurant operations?

There is a saying in the writing business: "Show, don't tell."

Show me that you can feed the partners of Goldman Sachs and their guests as well as I did and get promoted to VP in the Operations Division because of that performance. Show me that, then I'll listen to what you tell.

By the way, "Poisson processes" to me means what you do to the fish.

.................................................................

Oh. And P.S. Thank you for informing me that ADNY requires a fat checkbook. I might have been able to guess this as it was part of my job to visit places just like ADNY so that we could keep in the loop with our own kitchen and service.

And the reason why, finally, I don't really need your help in "getting ahead" in the industry is that I left it.

I made enough money while I was in it. Don't need help to get ahead. :wink:

Edited September 22, 2005 by Carrot Top (log)

project · September 22, 2005

Right, OK, you certainly seem to know your stuff (or, at least
managed to name the right names. PM me the actual math if it's
not classified because I'd be interested to look over it).
However, the question you were asked to solve is not the same
problem that ADNY faces. ADNY can't order a poisson
distribution of dough, ultimately, it has to condense down to
a single estimate.
And you still haven't addressed the other points I raised,
namely:
1. It doesn't scale down well. The information gathering
resources are so onerous it overwhelms the benifits.
2. You cannot get below the noise floor. No matter what you
do, there is no theoretical way to get below the noise floor.
Even if you knew the perfect distribution ahead of time, it's
impossible. And since the noise floor is very high, theres
very little meaningful information you can extract with any
method.
3. Theres not much point anyway. Everything can be dumped off
as specials.
4. You don't have the information neccesary to build a
statistical model. A dish may be on the menu for a month or
so. You can't do anything meaningful with 30 sample points.
Think about airline bookings, the airlines essentially have
the same problem, they need to book enough seats so as to fill
a plane yet not so much that they overbook. They can run a
billion statistical tests over it yet overbooking is still
reasonably common. What makes you think a restaurant can do
any better?
←

It isn't so easy to know how much to say.

Here I keep trying to give short answers and later say more.

Analyzing systems that operate under uncertainty is not taught

well very often.

There is some relevant software, but mostly it's not among

the really big sellers!

In some cases, new applications could be of value.

In principle -- and I believe at times in practice -- there

should be applications in parts of restaurant operations.

The math I did for SSBN survivability was decades ago -- when

there was still a Soviet Union!

I don't still have access to it, and it would be imprudent to

spread it around freely anyway.

Continuous time discrete state space Markov processes are

treated at various levels in the relevant literature.

Cinlar has a good relatively elementary book.

There is more in Dynkin's classic books.

You would want to look for the Kolmogorov forward and

backwards equations.

For queuing applications, should keep in mind exploiting

embedded Markov chains.

When they start using Laplace transforms in tricky ways to get

closed form solutions of distributions in tricky queues, feel

free to turn the pages more quickly!

When they start talking about non-regular Markov processes,

Martin boundary theory, and monotone class arguments, look for

a simpler book!

For

ADNY can't order a poisson distribution of dough, ultimately,

it has to condense down to a single estimate.

</blockquote>

Right.

Again, mostly we cannot hope to have ADNY never run out

of dough.

Our goal, then, has to be more modest:

For a relatively simple modest first cut goal, although

possibly progress over just intuitive manual methods, would be

[A] to get the distribution of dough needed and using that

distribution, to select the "single" amount to make to have an

acceptably low probability of running out.

For simple approaches we basically need the distribution to

calculate probabilities of running out as needed to select the

"single" amount to make.

Some caution is needed:

Distributions really are important, and given a

distribution we can calculate probabilities, but it does not

follow that a necessary step in every application is to find a

distribution!

A second cut would be to balance the cost of running out

versus the cost of dough wasted from having too much.

A third cut would be to include what you mentioned, uses for

leftover dough the next day.

This use of the excess the next day clearly can help reduce

waste but, still, for most inventory items, fundamentally does

not change the need to find and use a distribution.

Part of what is nice is how much Poisson process theory gives

us from just some really meager assumptions, ones that are

just qualitative and, thus, relatively easy to check in

practice just intuitively! This topic is called the

axiomatic foundation of Poisson processes and, for

applications, is nice.

Cinlar's book has a nice treatment (uh, I do have a change at

one point in his proof).

 

You have remarked on data:

That is a solid concern.

The amount of data needed can vary enormously.

E.g., the work I did on global nuclear war limited to sea (sit

down for this one) had as input data just some numbers on one

sheet of paper!

For other seeming relatively simple problems, the amount of

input data can be staggeringly large.

You mentioned entering data on the weather, etc.

One issue is, the analysis may not know what to do with, say,

a snow storm!

Sure, quite generally, the more data we have, know how to use,

and do use, the better we will do.

Still just because some data is available doesn't mean we have

to use it if we don't know how.

We are not obligated to type in just everything every minute

24 x 7 even if in principle it could help!

 

Another issue is the computing:

Sometimes the computing is really simple.

Other times, the computing is a total mess!

Broadly one of the more important issues in current practical

computing is what infrastructure makes applications easier to

do?

There has been a lot of progress; more is on the way; still

more is needed.

 

On 1., it depends on what "it" means!

Sure, if "it" is supply chain management (SCM), then

there is a well deserved reputation for taking a team of 200

people three years just for the initial implementation at one

company!

Some of the people doing the implementation arrive at work in

new Corvettes or Ferraris!

Further, on-going usage can remain expensive.

And the installation can be "brittle" meaning that small

changes in the company's operations can cause big breakage,

long delays, and high costs in the SCM system.

But such examples are for some fairly comprehensive coverage

of some fairly complex processes at large companies.

Here the effort is nearly all in just getting, checking,

transmitting, and storing the routine business operational

data.

E.g., need a list of items to order, candidate vendors for

each item, possibly really complicated pricing based on

volumes, other items, total ordered from the one vendor, etc.,

descriptions of people and their skills, processes and the

machines, capacities, data base schema, screen forms,

encryption and security, backup and recovery, etc., all before

we consider a Poisson process or necessary conditions for

optimality!

However, if "it" is just a small focused application of the

basic applied mathematics, then the effort can range from

something reasonable down to just a few keystrokes on an

engineering pocket calculator.

 

For what could be done at a high end restaurant such as ADNY

likely would require some investigation.

So far in the US, people who concentrate on high end topics in

engineering such as stochastic optimal control are reluctant

to spend time in a restaurant to "bridge to practice".

So, for the high end topics, less is known about the potential

for practice than we could wish.

But, the effort for a real application doesn't have to be

huge: I was doing a quite serious and valuable industrial

application of stochastic optimal control before I went to

graduate school. At one point I flew to Cornell to discuss

something, at the end mentioned a second problem, got a 90

second explanation of stochastic optimal control theory, ran

for my cab, and by the time my plane landed understood it all

well enough to start writing software.

Someone in the restaurant business might do something similar.

 

On 2., that is a common intuitive conclusion from classical

filtering theory in electronic engineering, but, actually, it

is not very general!

There are many places where it is possible to do some quite

clean powerful valuable analyzes in situations with really

messy noisy data.

One broad collection of examples is the work on

non-parametric (distribution-free) statistical tests of

about 60 years ago; this work has not been much used in

classical engineering but is everyday stuff in parts of the

social sciences.

These non-parametric tests can at times do terrific things

with meager data and meager assumptions in really messy

situations.

Actually, the Poisson process is something of an example:

Some simple intuitive qualitative considerations can be enough

to conclude that some arrivals really are a Poisson process.

Then all that is left is the estimate of the arrival rate, and

the obvious estimator really does have high quality and is the

one to use.

An estimate might take no more than 100 arrivals; might get

something useful for only 20 arrivals.

Then Poisson theory can say a lot.

One of the nicer tools, with some terrific properties, is just

cross tabulation.

At times can get some terrific results with meager data in

messy situations.

Of course, one of the classical examples was checking the

fairness of a roulette wheel.

A roulette wheel is noisy, but it didn't take much data

to show overwhelmingly that there was cheatin' going on!

Again, people work with lots of uncertainty everyday;

mathematics should be able to contribute to the work and to

improve on simple manual techniques without insisting that all

the noise be removed from the system first, and

mathematics can do this.

 

For 3., how much needs to be done on inventory control varies.

E.g., for beef stock, maybe they make 100 liters at a time,

reduce it to demi-glace, freeze it in ice cube trays,

pack it in humidity proof containers, and store it in the

freezer.

One batch may last them three months, and the shelf life is

nearly forever.

When it's half gone, make another batch.

Their chances of running out are next to zero; the cost of the

space in the freezer is low enough.

Not much more to do!

That's one extreme.

 

For "specials", sure, that often works, but at ADNY, likely

won't get last night's leftover biscuits as dumplings with

today's lunch chicken!

Yes, I know:

At 3 PM each day, each French restaurant in Manhattan is

supposed to get a delivery of 50 quarts of hand grown, hand

picked, hand selected picture perfect varietal raspberries at

$50 a quart, serve three quarts or so and turn the rest into

raspberry coulis and use it to make nice red lines on white

sauces or to lubricate between the layers or chocolate cake!

We're talking some expensive little red lines and

lubrication!

 

There are other extremes:

We could think of various cases of very expensive very

perishable fresh seafood and produce.

Here we would like to know about the trade-off between [A]

expected cost of running out and expected cost of what

gets thrown away or, perhaps, used for a less valuable

purpose.

Here there may be some significant complications from

interactions:

When both Maine lobster and Dover sole are on the menu, what

is the Poisson process parameter for each?

 

On 4., what you say is partly true but not always!

I never mentioned a "statistical model"!

What I did with global nuclear war limited to sea was

statistical, but one sheet of paper had plenty of data!

Really, then, the idea of building a "statistical model" is a

bit crude; sounds like something from a motel seminar in

regression analysis for marketing studies!

 

Having only 30 data points is not impossible!

With 30 points, can do non-parametric tests with probability

of type I error any positive whole number multiple of 1/30.

So, can get type I error as low as 100/30 = 3.33%, and that's

not so bad.

Just 30 points may well be enough for a usable estimate of the

arrival rate of a Poisson process from which one could say a

lot about the probability of running short.

 

At a restaurant with reservations, have a good start on when

the customers arrive.

Of course, they don't arrive exactly on time, but there should

be plenty of data for a rather good description of the

difference between the reservation time and the actual arrival

time.

Next, there is time to get them seated, get past any round(s)

of drinks, and place orders.

Now the orders arrive at the kitchen.

Given the menu and for each dish how the kitchen does it, at 9

PM, for each worker, what is, say, the distribution of the

length of the work queue?

So, does anyone look really too busy?

Anyone look really not busy enough?

If so, then might move around some staff.

Similarly for the major pieces of equipment.

Does this particular menu promise, with probability too high

for comfort, to cause a queue too long at some one piece of

equipment?

The quantities of data readily available should make it

possible to answer such questions, and the answers might be

useful.

 

For

<blockquote>

Think about airline bookings, the airlines essentially have

the same problem, they need to book enough seats so as to fill

a plane yet not so much that they overbook. They can run a

billion statistical tests over it yet overbooking is still

reasonably common.

What makes you think a restaurant can do any better?

</blockquote>

Uh, you might find that at times some airlines deliberately

did overbooking!

Airline reservations and pricing have been addressed fairly

seriously in what they call yield management!

Nothing I have said here would be news to the experts in that

work.

Such work has also been done for hotel room reservations; some

people in that field actually have heard of stochastic optimal

control theory (trust me on this one!)!

 

It does appear that, again, you are wanting to see the applied

mathematics and computing get a perfect solution -- no empty

seats and no overbooking.

Such perfection is usually asking too much, more than is

possible even in principle.

The most that can be asked is to do the best possible; that

can be much better than the results from manual or

simple methods yet still noticeably (perhaps frustratingly)

short of perfection of having each seat filled and no

overbooking.

 

Giving up right away is too pessimistic and not useful.

Again, humans, using just simple manual methods, do this stuff

everyday; there should be an opportunity to do a little

better!

project · September 22, 2005

Carrot Top,

I have not been questioning your qualifications at all.

Congratulations working on the southern tip of Manhattan.

The time I worked down there, the traffic was terrible!

Beds are not food. Beds are a stable product that sit in one
place in relationship to the user.
←

Right:

In an analysis of hospital sizing, mostly the beds are given

and fixed.

The beds are not the source of the uncertainty.

The uncertainty is from how many patients arrive each day for

each medical specialty!

Then these patients have to flow through the hospital.

Even if have beds enough in total, may have shortages of

capacity in individual departments -- burns, intensive care,

general surgery operating rooms, etc.

At times, hospital sizing has been taken seriously; the work I

and a colleague did was taken fully seriously in a serious

situation.

A restaurant is similar in the sense that customers arrive

instead of patients; food orders go to various stations

instead of patients to various departments, etc.

The hospital beds would correspond to, say, the restaurant

tables.

The challenges of the uncertainties are similar.

Monte Carlo simulation is simple enough: It is what happens

in the old board game Monopoly when moves are determined by

rolling a die.

So, if have a suitable description of some uncertainty, then

can use a computer to generate appropriate random numbers and,

say, see how a hospital or restaurant would work during 10

years of operation.

That is, in a few minutes on a computer, we can get data from

10 years of experience.

If the data shows something is wrong too often, then we can

make some changes and try again.

It can be useful.

The methodology of Monte Carlo simulation for doing hospital

planning and evaluating the chances of a hospital being too

full to provide needed care should also be applicable to

aspects of restaurant planning.

That's a more complete description of the connection between

hospital planning and restaurant planning.

For such Monte Carlo studies, there is a nice book by Maisel

and Gnugnoli.

Their book has an extended discussion of an application.

As I recall, they did either a medical clinic or a social

services office.

In claiming that such methodology could help in planning

restaurant operations, I am on solid ground.

For bringing in a Poisson process, that is technical.

Apparently Shalmanese is familiar with the subject.

At Goldman-Sachs, at one time I would have referred you to F.

Black.

However, there will no doubt be people there now who know what

a Poisson process is; Goldman-Sachs may have hired some people

from the Princeton ORFE program, and they will know!

A Poisson process has to do with arrivals.

In a restaurant, there are arrivals of many kinds.

At a McDonald's, all during open hours, there are arrivals of

vehicles, mostly cars, with customers.

There are arrivals at each deep fryer, burner, oven, station,

refrigerator door, etc.

There are arrivals of orders for mixed drinks, bottles of

wine, and each item on the menu.

Roughly a stream of arrivals form a Poisson process

provided [A] in each interval of time, the distribution of the

number of arrivals in that interval does not depend on the

time of the beginning of the interval and the number of

arrivals in intervals of time that do not overlap are

independent. So, a Poisson process has [A] stationary

independent increments. Short. Simple. Nice.

 

Commonly in practice we can check just intuitively if [A] and

are true. Just from these qualitative assumptions

it is possible to say that there must exist a number, call it

the arrival rate, so that, with this rate, for any

interval of time, we can just write down the probability

distribution of the number of arrivals in that interval.

Moreover, the times between arrivals are all independent and

have the same distribution, an exponential distribution. We

can also write down the expected number of arrivals and the

variance (Goldman Sachs calls this volatility) of the

number of arrivals. For [A], it is possible to weaken that.

 

For more, given two independent Poisson processes, if we

combine the arrivals, then we get another Poisson process.

So, if at ADNY the arrivals of raviolis is Poisson and the

arrivals of spaghettini is Poisson and if they are independent

(could be some question here), then the arrivals of uses of

the common dough is Poisson. So, right away we know a

lot about the arrival process for dough and have a good

start on being able to calculate the chances of needing, say,

more than five pounds of dough for one shift.

A careful calculation might consider a little more, but this

is a good start.

If I were at ADNY, were working for someone with a 14"

Sabatier, were responsible for having enough dough for the

shift, and had any doubts about five pounds being enough, then

I'd do some such calculation to be more sure!

If I had an excess of dough, then I'd use the calculation to

show that, even with the five pounds, there was a 1% chance of

running out and offer to raise that to 5% or 10%.

And, I would not have to wait a few hundred days and collect

data to estimate that 1%!

Now we're getting somewhere!

When I read Part Two, this is what I thought should have been

done.

 

So, for any of the arrivals in a restaurant, could check [A]

and , collect some data, estimate the arrival rate, and

then have some quite finely detailed information on the full

distribution.

 

E.g., maybe a restaurant has Meusault by the glass. Well,

then, right away we can say that at least a decently good

first-cut approximation to the distribution of the number of

glasses they sell in one shift will be a Poisson distribution.

If we can estimate the rate in that distribution, then we can

quickly calculate the probability of selling more than 16

glasses of Meusault. Doing this and checking the Meusault

stock, we might discover that the chance of having enough

Meusault for the shift is only 80%. So that we don't

disappoint 20% or so of our Meusault customers, either we

should get some more Meusault before the shift or select a

different white Burgundy to offer by the glass.

It's simple enough and could be useful.

 

I saw plenty of weakenesses in the MBA programs.

Still, the MBA programs are a major part of what is there.

They are better than nothing, do have some good content, and

commonly are taken quite seriously.

For business education, as one of the best restaurant managers

in the US once said to me about a certain red Burgundy wine,

"You won't find better."

Missing something that is in an MBA program and needed could

be costly!

Good, bad, or otherwise, the content of the MBA programs has

been considered about as carefully as we can expect for such

things in our society.

These programs have long commonly taught material very close

to what I have been discussing.

Net, the material should be taken seriously.

 

A little more generally, in the US, the research universities

are one of the most important sources of high quality

information.

We're rarely going to get better information from politicians,

TV commentators, newspapers, magazines, movies, movie

celebrities, motel professional training seminars, popular

books People with 77 Effective Habits, software

salesmen, government departments, management consulting firms,

gurus, stock brokers, novelists, TV talk show hosts, mutual

fund salesmen, etc.

Carrot Top · September 22, 2005

Good morning, project.

Nice weather we are having today.

.............................................................

Your critical analysis and explanation of the Monte Carlo and the Poisson processes undoubtedly is correct.

Unless there is someone out there that can say differently, and I doubt that, but who knows. Life is full of surprises.

What I think the problem might be with the notion of applying these concepts to restaurant operations, however, might be VEFE. The dreaded VEFE again, indeed, or even in this case paradoxically the very lack of it that exists in the sort of very rigid controls that you describe.

People enjoy the tussles, the wants, the needs, the excesses, that occur in everyday life. It allows them the opportunity to problem-solve on their own.

If there are no problems to solve, then what would life "feel" like to people?

No VEFE actualized, no fun.

The sort of rigid controls that clean everything up nicely are beautiful in concept. But in reality, who wants them? Unless, of course, you are talking about war. War between either countries or businesses, it is much the same. Someone wants to kill off someone else, and in their each individual minds for very good reason.

But usually the restaurant business is full of people who want to create and nurture through food. It is full of hidden VEFE-lovers who probably did not even know they were.

For an example of this, I must point you to the other thread currently running on eGullet about Doug Psaltis. Supposedly about Doug Psaltis' book, again.

Lots of discussion there, but most of it not on his book. Most of it on him, himself, as personality. As celebrity chef. Over 4500 views of this discussion last time I looked, and what is this discussion? Is it about what he cooked or created including the book? Is it about a "subject"? No, it is sheer fun gossip about a person. This, is what people like. Would they be interested in the same story about some guy down the street? Nope. For they would not get the same sense of self-importance and of being in some sort of imagined "in-crowd" if the story were only about the guy down the street.

So again I am afraid that VEFE rules, project.

The thrill of the everyday, the gathering of the slavering crowds, the very excitement of feeling human, a part of things, part of the problem-solving process, this feeling that exalts itself in VEFE or alternately in real life drop the V/EFE is what makes people get up in the morning.

VEFE, control, perfection. All anyone really wants is a sense of being human and of being a part of things.

And what that translates to is: Monte Carlo is preferable as somewhere to travel to on vacation, and Poisson is much more interesting when it is something caught jumping from the sea waiting to be cooked into a tasty delightful dish to be brought to the dinner table.

Edited September 22, 2005 by Carrot Top (log)

Nentony · September 22, 2005

What she said. By the way, I'm really enjoying the book.

Tony

LindaJ · September 22, 2005

Well, ADNY is really a pretty small business and this discussion of Operations Management is more related to some place like, say, GM. Maybe Yum brands needs to forecast how many pre-frozen jalapeno poppers they need on hand for the next millenium, but that's way more boring than actually talking about restaurants like ADNY.

I have an MBA and this is an awesome discussion that helps me remember why I didn't go into Operations.

I agree with previous posters that the discription of the French restaurant heirarchy has increased my understanding and the book looks pretty cool.

project · September 23, 2005

Nice weather we are having today.
←

You mean category 5 hurricane Rita?

Actually, here in New York, the weather is nice; as I recall,

you are in TX?

Your critical analysis and explanation of the Monte Carlo and
the Poisson processes undoubtedly is correct.
←

In case you want to check if I am correct, I've provided

plenty of references.

You mentioned Amazon:

It should be able to give you ISBN's.

What I think the problem might be with the notion of applying
these concepts to restaurant operations, however, might be
VEFE. The dreaded VEFE again, indeed, or even in this case
paradoxically the very lack of it that exists in the sort of
very rigid controls that you describe.
←

Again, the "controls" do not have to seem "rigid":

No doubt you have made plenty of use of measures of weights

and volumes, timers, thermometers.

For means to estimate the amount of dough needed at the start

of the evening service, for what I mentioned about

probabilities, it might be possible to have just a little

graph on a single sheet of paper; you have other methods more

common in the restaurant industry; in either case, using the

methods is no more "rigid" than using a thermometer.

No, it is sheer fun gossip about a person. This, is what
people like. Would they be interested in the same story about
some guy down the street? Nope. For they would not get the
same sense of self-importance and of being in some sort of
imagined "in-crowd" if the story were only about the guy down
the street.
So again I am afraid that VEFE rules, project.
←

Yup, it does so appear!

On "what people like", from Pretty Woman we heard that

it is "sucking up"!

I believe that you are closer to correct.

Yup, you have described the vicarious escapist fantasy

experience (VEFE) of gossip about celebrities!

On "what people like", your "in some sort of imagined

'in-crowd'" has some support:

There is an E. Fromm book that explains that the main issue in

life is getting a feeling of security in face of the

realization that we are vulnerable to the hostile forces of

nature and society. For this security, he says only four

means have been discovered: (1) love of god, (2) love of

spouse, (3) membership in a group, (4) suppression of the

feelings with passion, etc. Your remarks about in-groups,

gossip, etc. are likely mostly part of (3). It may be that we

should put some of VEFE and, sometimes, food in (4).

Fromm omitted what I would mention: (1) money in the bank, (2)

knowledge able to overcome the hostile forces and put money in

the bank, and (3) power that can yield money in the bank that

can buy knowledge and its results as needed.

Some months ago, eG had a change in forum policy that reduced

the chances of eG being used to arrange social gatherings.

So, with your "4500" and "gossip" you are saying that eG is

being used for vicarious "social gatherings"?

Yup, as we heard in Jurassic Park, "Yes, they do

form herds!".

I don't do gossip:

I don't read tabloids, read movie celebrity magazines, or

watch movie celebrity TV programs; I don't gossip on the

telephone, at the grocery store, or over a back fence; on the

Internet and eG, I don't either read or write gossip.

Never have; never will.

Some people really actually do not like gossip!

Again, I can enjoy VEFE:

It's good for light entertainment, e.g., movies.

But I come to eG for information on food, trying to get

documentation, instruction to let me be a better cook.

That's one of the main stated purposes of eG, and it really is

why I come.

For any ulterior hidden unconscious motives -- nope!

VEFE does not help me be a better cook.

Some people have enjoyed Part One and Part Two.

I didn't, not even just as VEFE.

Formula fiction is about an interesting person, with a

serious problem, who makes admirable and successful efforts to

solve the problem, and gets the girl.

However, for Part One, the guy used a simple hand tool for a

simple manual task, injured himself, and maybe put raw human

blood in the raw oysters -- which would convert expensive

seafood to dangerous medical waste.

For Part Two, the guy did not use simple solid means to

estimate the amount of dough needed, ran out, and caused a

problem for one of the world's best restaurants.

In both cases, not very "admirable" or "successful".

Maybe I'm supposed to like the story because these simple

failings make it easier for me to see myself in just that

situation.

Nope:

Like billions of other people, I've used hand tools without

getting hurt.

Millions of restaurant workers don't run out of supplies

during crunch time; running out at ADNY is sickening;

and, for me, I have studied, used, taught, and created solid

techniques for how to handle uncertainty in real situations,

including avoiding running out.

In formula fiction there should be admirable and successful

efforts; here we have someone hurting themselves and hurting

one of the world's best restaurants.

To me, no fun reading about such things.

I'm not getting any enjoyable VEFE from someone who fumbles

the ball, drops the ball, trips on the ball, falls on the

ball, loses the ball, ends up face down in the mud under

people who recover the ball.

Not me.

Your

So again I am afraid that VEFE rules, project.

</blockquote>

does raise an interesting question:

Why so much VEFE?

What can we do with VEFE?

What is it with VEFE?

Again, actually, it is not true that "VEFE rules".

You didn't make it as a chef serving just VEFE.

Instead, you had to serve real food.

For an answer to my questions, I would say:

There are buyers and sellers.

Since VEFE is cheap to provide, sellers love to have people

buy it.

The media, "Yes, they do form herds!", and they have

nearly 100% consensus on providing just VEFE.

But, I can't use VEFE; again, it doesn't help me learn how to

cook food and puts no food on my table or in my belly, and

these failings I very much do NOTICE.

So, for a good movie, maybe I'll buy some VEFE.

And, some good classical music.

Otherwise, I'm not buying.

My view is that at least 99% of the time people are buying

VEFE from the media, these people are getting ripped off.

I've been surrounded, drenched, bombarded, attacked with

people trying to sell me VEFE or, in school, even forcing me

to take VEFE, but I've never tried to sell VEFE.

But, maybe there's a chance!

Once I was in a drug store waiting on something, saw a huge

magazine rack, noticed what seemed to be a lot of magazines on

romance, and wrote down the titles.

It was a huge collection; I got maybe 50 titles.

Looking, I noticed that the three most common words in the

titles were secret romantic confessions but that there

was no magazine with the title Secret Romantic

Confessions.

So, that's a start, right?

To heck with a magazine teaching people how to cook.

Instead, have a magazine for people who are having affairs or

imagining that they are!

I know, I'm doomed:

I keep concentrating on material challenging between the ears

when all the action is below the shoulders -- the heart, the

gut, and lower still.

I very much did not like Part One and would very much

have preferred some nice photographs on how to shuck oysters

safely.

I very much did not like Part Two and would very much have

preferred some solid information on restaurant operations,

information on how to be successful, not on how to fail,

information certainly available from one of the world's best

restaurants.

To me, going to ADNY and leaving with lessons in failure is

degenerate, dysfunctional, and disgusting.

I don't touch cigarettes or illegal drugs, and I take high

quality beer, wine, and VEFE in only small quantities.

There is a curious point:

There are nearly no titles with vicarious escapist fantasy

experience, not on eG or in the media.

Instead, nearly always there is a title that has face value of

being real, rational.

So, however much people like VEFE, at face value the content

is nearly always something else.

Basically, VEFE can't show its face in public.

It appears, then, that the media top brass want to convince

themselves that people don't notice and don't mind and,

really, are happy, happy.

Well, I'm saying that I do notice, right away, do mind, very

much, and am torqued, pi**ed, angry.

For ADNY, Alain Ducasse has done well building a reputation as

one of the world's best chefs.

Part Two, however, has in a fairly responsible position in his

kitchen someone who injures himself with routine tasks with

basic tools and who makes serious and nearly inexcusable

mistakes in elementary parts of the work.

What is amazing is how Ducasse got talked into cooperating in

having his operations described in this way.

That evening, a customer had the right to demand,

Where's the !@#$%^ )(*&^%$ spaghettini!

</blockquote>

I'm confident that Ducasse doesn't get many such questions, in

his whole career hasn't gotten many such questions.

The real situation, not the escapist fantasy one, is

crucial and includes -- he's good, it's crucial that he is,

his career'd fall like a shot duck the first week he wasn't.

The work of his kitchen might be a crown jewel of

civilization; thus, to me the description in Part Two is just

sickening.

Carrot Top · September 23, 2005

Okay, project. In many of my past posts I have been tongue-in-cheek with you. Why not? You are fun to be tongue-in-cheek with and certainly a challenge to be tongue-in-cheek with.

This will be a short answer today, for I have actually got to find time to do other things for some bizarre reason.

Yes, you are right that it was rather inexcusable for the kitchen to run out of anything that would affect the guests dining "experience".

And yes, the excerpts of the book were written in a certain way that could invite the sorts of comments you wrote.

I can understand your anger (if that is the right word, if that is the word you used, it was intimated anyway but I do not have time right now to go back and check for the exact words you used) because I am prone to it myself.

I have a different way of taking on these things than you do, though. And I hate to say it (let's not even go there as to why I used that phrase although I should not hate to say it and indeed I probably don't really hate to say it :laugh: ) but as I think of it, generally I do it in the way a woman is traditionally considered to rather than the way a man would. I smile and go "around" whatever it is, giving a tweak here and a poke there and a small push there till the final hopefully better conclusion is finally reached, rather than just directly address the "problem" smack-head on.

Yes, in my "career" I learned to be outspoken, bold, and someone who would not be challenged outright by big strong testosterone-loaded guys in the kitchen. But to do it daily just in life does not feel right to me. So I go "around".

This can and does work. It is actually the "traditional" strength of women. And it can be more palatable for many people to accept challenge that is personally directed at them when shaped in this kinder fashion, also.

It is wrong that the workings of the kitchen did not go as efficiently and as well as they could have. A good manager that was on-task could and should have been able to straighten this out, though, with the tools that we do have in use. No reason why not. Again, if you are curious, PM me and we can talk more specifically about what tools are used for controls in the kitchen. Or maybe start a new thread on eG, that would provide a variety of responses that would be more "all-inclusive" of methodology.

Yes, it is "sickening" that that should have happened. But ultimately, overall, that is the best that is out there. ADNY, even, is subject to imperfection according to the excerpt from the book. It's my feeling, though, humans being humans, that whatever tools one gives them (us, me even to be inclusive :biggrin: ) they will f*** up sometimes. And of course, that makes a "story". One that can be written and sold.

And in looking at this sort of thing in a broader or more philosophic way ("this thing" being my proposal that operations controls can never really perfect human endeavor) I often think that if every single thing in the world WERE to be made "idiot-proof" we all would simply, in that moment of time, POP! disappear off the face of the earth as if in some science fiction novel. For idiots are part of us, an integral part of the chaos that ties the universe together. If they were eliminated, where would we be? POOF! I like to imagine we would just disappear.

I've lived in many places but never in Texas. You can read my bio if you are interested.

................................................................

Oh. Still, I enjoyed reading the book excerpt. :smile:

Edited September 23, 2005 by Carrot Top (log)

Carrot Top · September 23, 2005

Time to serve coffee and dessert, project, after this many-course meal we've had on this subject and then I am "outta here". The grass on my lawn has grown about an inch and the mower (you know, the one with a choke) refuses to do the work by itself.

For dessert I will offer thoughts of honey. Honey, dripped gently onto anything, makes it beautiful and attractive. Plus, you can catch more flies with it when trying to prove a point.

Vinegar, though bracing and useful, people tend to avoid. Unless it is well-oiled and temptingly seasoned.

................................................................

For coffee, here's a thought that hopefully will be strong enough to make me walk out the door of this fun though somewhat offbase discussion and not walk right back in.

I could be wrong about the overall concept of what this book should be.

And you, could be right about it.

Being wrong at least fifty percent of the time in life seems a good thing to me.

Why? Because it gives others so much happiness when they can be right.

And that, is an excellent thing.

kitchenmage · September 24, 2005

The user of the bed takes the bed as is. They do not ask for it without shallots or with a cream sauce, also, again "a minute". Another example of differences in the relationship between beds and food and users.
←

I was going to ask project to pass me some of whatever he's drinking, but I think I'd much rather share yours... :hmmm: :huh: :rolleyes: :biggrin:

Sign In

ADNY: Lost in Translation

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Similar Content

Recently Browsing 0 members