4 hours ago, Oberon_Command said:

You need to imagine a world where all the physical rules are made up by humans and humans can change those rules at any time.

Regardless how phisic rules made autopilots have to drive takig care about it rules, automatic doors have to react to approached player, tanks have to aim his turrets and so on. So game AI and automation and real robotics algo have very similar or even same background. Of cource it is a huge difference that ingame virtual measuarment tools not affected by natural effects, so works with devilery precission without any  black magick involved. Also it is main huge difference - games simulates so big world with so different automation in so much instances so any real world automation system jittery smokes bamboo far away. So why not to use a well-proven algos and metodologies that intended to help build control systems that can be easyly tweaked by data driven way and in most cases self-regulated into game development? 

9 hours ago, Fulcrum.013 said:

"Director have to press a button and computer have to show reports how many loses and profits factory have today, from witch it come and exactly recomendations how to avoid losses and increase profits."

Is you really think its specs enought to make system like it without any researches?

That's not a specification, that's a user story. And it's an almost textbook example of where iterative design would work.

Assuming the data is there to be mined, you should be able to produce a simple report showing key metrics in less than a day. Then the business analysts can start to look at it and refine what actually needs to be displayed.

38 minutes ago, ChaosEngine said:

And it's an almost textbook example of where iterative design would work.

realy finally thay has purchased ready EPR system and then 18 years  made iterations to adjust it to its wishes until factory gone complete bancrupcy 2 years ago. So it is really textbook story about what agile able to do into field where even basic system analize can  give exactly recomendation without any  EPR - to be in profit thay should to shift their designers to CADs and production lines to CAM 20 years ago insted to play 20 years to antiscientific games.

11 hours ago, Fulcrum.013 said:

But specs can not be used to design [develop] a software architecture, becouse it is just a list of features without any explanation how it have to work inside or even detailed description how it have look.

It can, if one writes a thorough GDD and dubs it "specifications" (specs for short). The specs docs you've seen were not GDDs, apparently. A lot of us in games use different terminology to refer to the same thing. Specs = GDD to a lot of people in this industry.

8 minutes ago, Tom Sloper said:

GDD

GDD is kind of "technical goals" but not kind of task description. It describes goals but not how it can be reached. Instead of it  task description describes exactly way how to reach its goals. And also it is exists a  double straight line betwin GDD and task description - GDD is a goals that developers have exactly hit but bot change, while with task description thay have to make anything that found helpfull to reach GDD goals. 

Is GDD describes for example how to find intersection of curved cilinder and curved torus or wich formulas have to be used to model fire o water  effects required by GDD? 

29 minutes ago, Fulcrum.013 said:

So it is really textbook story about what agile able to do into field where even basic system analize can  give exactly recomendation without any  EPR

Hang on, do you think Agile is a substitute for system analysis? 

That's not how this works. You don't just jump feet first into coding with no analysis. You still need to figure out what it is that you're doing.

Agile just recognises that you will never have the complete picture up front and your analysis will change after you start the work.

10 minutes ago, ChaosEngine said:

Agile just recognises that you will never have the complete picture up front and your analysis will change after you start the work.

Why it have to change?

@ChaosEngineAlgo is universal thing by his definition. So in case it designed properly it will newer change.  Just look for example to 25 years old software that still alive until now. For example compare a C++ Builder ver 1.0 and C++ Builder XE 10.2. Nothing has changed. Only addition of new components, shift of compiler to 64 bits but core libraries code that drive anything else still unchanged. Other example - SolidWorks. Look at version 1.0 and version 2017. Same functionality of geometrical core and addition of new components not related to geometrical core only.  Most significant addition - finite elements modeling core that uses models produced by geometrical core as input data. And geometrical core never be changed. Just becouse it cover full set of primitives and operations over its, that allow to model any geometry. Same with finite elements modeling core. It never be changed becouse it universal. It only will have a addition of tasks that it preconfigured to solve. And so on. So agile just a tool to waste a time. Most better metodology is a decomposition of whole system to independent or low depended subsystems and universal implementation of each of it, in order into wich already implemented subsystems help to implement rest of subsystems.

41 minutes ago, Fulcrum.013 said:

Why it have to change?

Because experience has shown that it will change. Every. Single. Time.

15 minutes ago, Fulcrum.013 said:

For example compare a C++ Builder ver 1.0 and C++ Builder XE 10.2. Nothing has changed. Only addition of new components, shift of compiler to 64 bits but core libraries code that drive anything else still unchanged.

2

I think you might need to look up the word "change" in a dictionary.

Ugh, this is pointless.... I'm done with this conversation.

5 minutes ago, ChaosEngine said:

ecause experience has shown that it will change

whose expirience? Is it expirience of guys that unable to understand that list of required features have no any relation to description how to implement those features? It is nothing wonderfull. Becouse any algo can not be developed properly without complete research of field.