Making planing software

The part in the article I was referring to was how cockshott explained how to plan a whole economy with linear optimisation, down to the choice between two kinds of products or different places to build a factory. The same scalar vector of products would still be used to plan, which would be arrived at by comparing the store prices to the labour value of products.

Its been a while since I read that paper, but why the fuck would you want to plan in money (or labour voucher terms, which is the same in real terms since you compress it into a singular item)? Cockshott shows how you can plan COMPLETELY in kind, without needing to measure things in terms of a common unit (labour or money) since in a fully linearly optimised economy, labour is just another resource. Labour would only play a special role in the distribution of product and the making of the planning vector, not the planning itself.

yeah yeah i agree with this plan, for a moment i short circuted and thought you were talking about the original kantorovich plan to implement on the usst, i apologize

got to write it in Rust, the communist programming language

...

hurh
>>>/gnussr/206

I have been writing in rust but holy fucking shit that syntax is so bad. I like the features but the syntax is just so overly complicated with their double :: and other bullshit.

scala profit off the great JVM and great libs, but I'm not sure it's the best for our goals. Do we need massive async, optimized network, custom DSL or shit like that ? do we even need a functional language ?

but all in all, if I were to start putting my productivity behind my ideas, I'd start with a news wiki+plugin, shit that help spread information and organization, not straight up planned economy

well i'm thinking the linear programming needs to be done in the fastest way possible, so something like C would be appropiate, now there shouldn't be a central node that'd be dumb, we have to assume future revolutionaries will be fighting for control of land, and be under constant attack, so a network i think is needed.
and yeah i agree a wiki is needed, what should we name it?, is google sites okay for this?, i'd think so but what do you think?

GNU Octave uses GPLK library for linear programming implementation. As far as general algorithm and outline of a planning software as well as its particular parts, a versatile tool like GNU Octave should be used.

Matlab is used in most of engineering applications and simulations, and GNU Octave as far as pure math goes rather comparable. Sure one can nitpick cases where the performance is worse, but as far as it goes, its syntax is simple and familliar to those who tried programming in C or C++.

octave.org/doc/v4.2.1/Linear-Programming.html

almost all languages have linear optimization packages, the problem is the speed of the languages and the complexity of syntax, octave is simple but also slow, we need as much speed as possible, if you really wanted symple syntax and math tools with C like speed i'd be inclined to recommend julia, but as julia is beta, and the linear programming part is i think the simplest part of the netword, so a lower level language might be better

The syntax is fine, you can't have Rust's style of memory management without a bit of extra syntax.

humanitisfag. eggsplain, plox

Although python might not make sense from a computational perspectice or from a developer's POV rhe language is far more readable for noobs, so it could potentially have a much higher pedagogical value (esp. with a densely commented sourcecode).

The question is: do you want to create a 1337 shit, or a commie shit. Since it's just a simulation, I ask u to reconsider. Project phase2 could be an actual runnable model, while phase1 could (and imo should) lure in as many from diff. bgs. as possible.

That's masturbation. It should be a proof of concept first.

That the example of the harmony function given in TANS only has one method of producing per thing was strictly for educational purpose, not because of a general logical limit to what such a method can achieve. The modifications that are tried are judged by a scoring function, and you can just modify what enters the scoring function under the header of an industry. That is, when you have e. g. two methods of producing a particular product (each with a different set of fixed co-efficients) you can just take the summed output of these two processes as the output of that industry.

Wait a minute, I'm getting my sway config up and running.
What do you guys want it in. I can go for C Python or Ada.

Attached: animeccp2.jpg (800x575, 128.13K)

I could learn haskell or scheme. Which one would be good?

That's better. I do know some octave.

okay here is my project outline for the project, respond with what you disagree with, and with what you want to add

Outline for linear optimization plan
The desired characteristics of the planning network we are going to make are the following
- Environmental sustainability: this means there need to be material caps for total amount of CO2 produced and dangerous liquid, solid, and gaseous waste
- Demand meeting: Every product must meet the demand of the population when possible (i.e. when there is no state of war, hunger, etcetera)
- Decentralization: There must be no central node to attack, revolutionaries might be in a state of war when the plan is implemented
- Democracy: the general population must have a vote in how the total labor of the country is divided, at least in key aspects
The parts of this network are going to be the following
- Data gathering: this mechanism must take important information such as total population, amount of people per profession, natural resources available, means of production and their location, etcetera and store them in data bases such that they can be used by the linear programming mechanism
- Pricing calculator: calculates the labour time given to each product as a price, such that the demand for that product clears the supply until the next production batch
- Demand calculator: this process must take past instances of purchase of a given product, and use them to calculate the approximate future demand of that product, possibly by statistical methods
- Linear programming: takes data stored about natural and man resources and demands for each product and generates an allocation of production to meet such demand

Project steps
1. making the linear programming part for a hypothetical economy of a given size
2. making the demand and pricing calculators
3. data gathering

What's the difference between one and two?

the first part is just going to take a bunch of data and organize it in equations that can be put through a linear optimization package, and that is going to yield the plan
the second part is going to take the data of purchases of a given product, fit them against a distribution, and use statistics to calculate what is the probability of the demand being a given number.
in short both are completely different mathematical beasts

thanks for the clarification.

Yea I've been coming around to it a bit but I just dislike the c++ inspired syntax. Would prefer something a little more java/c# esque tbh.

I am 99.99999999999% sure we are not going to implement our own linear optimisation algorithm. Thats reinventing the wheel, but shittier.

Since we will use something that is already implemented (in C) for our computationally intensive part, we can basically use any other language to do the other stuff and just use a FLI. Hell, if you wanted (but dont do it) you could write all the other stuff in python and then send it to a wrapped C linear optimisation function. The speed of our wrapping language really is of no concern, what it should do is serve our needs of easily writing the surrounding program of creating and calculating constraints, interfaces with outside datasources, etc.

no

In simplest terms:
If you can express a task of optimising something with linear constrains (y = 3x + 5z + 4) you can optimize it with linear optimisation algorithms. You can do this with an economy, as shown by cockshott in paper related here . But more formulas increase the time it takes (disproportionately larger to the amount of formulas). So cockshot made a harmony function to distribute resources based on a set "target amount of stuff" which would sort of aim for the target, to save calculation costs. This is not the most optimal, and with newer developments and more powerful PC's we can use linear optimisation to basically do the entire economy. It is also arguably more simple in terms of code and different algorithms and steps, making it more understandable and easier to maintain (and write).

Pyhon is fucking shit to develop anything substantial in. I will not stoop down to dynamically typed runtime checked languages, sorry, my mental health is somewhat important to me. I dont care if it is understandable to dumbfucks, dumbfucks are not going to read the python code anyway, and dumbfucks who cant lrn2statictype are, imo, also unlikely to understand things like linear optimisation, so it might as well be a black box to them, so you might as well use something better to build their black box.

Stupid and unneccecary. With linear optimisation you can add both options (as linear equations) and the optimisation will assign production to both in accordance with the goal and available resources.

What do you mean? What is your definition of "meeting demand"? Equaling store price to labour value? Fullfilling all possible demand? (impossible due to limited resources)
This would be the distribution of labour to social services (and specified which kinds of services) vs the consumer goods (which would be assigned internally according the goal of store_price == labour_value)?

eh, what?

Don't forget the resource requirements for the non-consumer side, ie social services, and expansion of the MoP. These are all things that have to be produced.
Sorry if I am mistaken because its 2 am, but shouldn't this be done in the step where demand is calculated, before the planning vector is created?

So what, just generate a mockup model of an economy (factories, transport, resources, etc) and then randomly generate some values? Then create formulas from it (really in this case you generate the formulas immediately).

I have no fucking clue how to do statistical prediction to be honest, I took a course in statistics but I only remember how to calculate boxplots.

...

No wait we could, if we wanted to, include the expansion of the MoP in the linear planning phase. It would just take some work.
You still have to take into account the costs of the social services though.

Could we theoretically just do this by demanding a constrain that the stock of stuff at the end should be at least 0 (or whatever it is) and setting the starting stock at negative whatever the cost is?
IE we need 10k bricks for roads, we say we start of with -10k bricks, we need to have at least 0 bricks, and then let the consumer good planning go along as usual? It should then aim for the normal planning ray with available resources and also produce our set constant amount of stuff.
Would this fuck up anything?

Good night gomrades

Attached: economic calculation problem solved.jpg (644x483, 100.05K)

if we are just putting a wrapper on a C algorithm, then i put foward again my suggestion of julia, it's a statically typed language with type inference, and it has the syntax that's piss easy, so development time would be reduced.
the next problem would just be making a wiki, and writing down the equations for stuff, which i'll do

yes this is it, the demand is met when in a given period of time there is enough of a product so that store price doesn't rise or fall significantly

i'm thinking more about letting people decide allocation of labour through vote in all areas not just social services, this would mean people can decide how much to invest in growth, science research, etc

it's just an algorithm that rises of decreases prices so that products don't run out in stores and there is no shortages

oh shit yes i knew i was forgetting something

yes, more or less a simmulation to see response times and where stuff is broken and needs fixing

i don't know that much either, but from what i do know you fit the data to an established formula, and use that, however i think we can get a statistics guy to take care of that, i mean there has to be a statistician among all the shitposters

i was just thinking about treating machines, factories, etc as products themselves, where you have a set stock, and just letting the algorithm decide what to build in order to meet planning

literally what?

although scrap that thing about including the growth of means of production inside the optimization itself, i don't know if it's going to send calculation time through the fucking roof, by almost cuadrupling the equations

I dont think using an obscure language is going to be too beneficial given that widely used languages have better tooling and library support.

Yea, I would class those under social services as well. Thats what I meant.

Oh ok, so just short term price corrections

Yea I was thinking it would be impossible because you can only build one factory in one place at a time but you can just limit the building of that factory as a product to one with a constraint.

The linear planning algorithm only plans for the planning vector, for the proportions of products to each other. This is a scalar. Social services (and other non-consumer areas) have non-scalar, constant resource requirements.

Rephrased/rewritten, since you cannot add a constant to a scalar goal, my though was that you can set extra variables with a restriction of having to be equal to the constant requirements of the other parts of the economy, and then simple make those variables have a y = x formula where y is the resource requirement and x is the same resource as a consumer good. This way the feasibility area would be limited to only that area where the constant requirements are met, and the optimisation for the consumer goods can still occur within it. So for my example, it would try to optimise for consumer goods, but only where 10k bricks are made in addition to the consumer good optimisation.
(Maybe I am just having a brainfart, its 3 am here, I feel like im not thinking straight).

Just because we can doesn't mean we neccecarily have to include every fucking thing we can in the global optimisation. You could optimize for product distribution between different supermarkets according to localized demand but the added benefit really is next to nothing.
But large MoP increases (that are not specialized like a one of a kind project) with different options could be included, if we wanted to.

anyway good night

I would like to put forth an addition requirement:
We need to be able to swap it out for another algorithm, especially since I am pretty sure a distributed form of linear optimisation could give us much benefit, given that we would have tons of existing, idle, not-too-powerfull computers available as computing resources, as opposed to modern businesses, which run on purpose build server farms.

now that you mention this, i was thinking of changing that ratios crap, and just replacing it with a simple restriction
Amount of product to be produced-Expected demand of said product>0

and for the rest i do agree that they need to be included as separate variables, however i need to read your comment in the morning

also i've created the wiki, i named it gnosplan since a guy in the other thread suggested that
gnosplan.wikia.com

actually the calculation problem doesn't exist and was made up by bourgeoisie economists in the early 20th century, because feels>reals, to try and discredit socialism and all the previous economics of Ricardo and Adam Smith. Change my mind.

Just… no. The entire point of cockshott's cybernetic system revolves around this. The amount of products that need to be produced should not be set as a constant value, because then it would fuck up the entirety of the price == labour_value system. Prices of products are a function of their demand and the prices of all other products. The linear optimisation function needs a function to optimise for, and if you restrict the ratios of production, you can optimise for any output. However, if you do not restrict ratios, your linear optimisation algorithm may choose to make a trillion fidget spinners, because you chose that as the optimisation.
See paulcockshott.wordpress.com/returning-to-kantorovich/
You need to restrict the quantities in order to correctly maximize production.
Otherwise, if you choose to minimize labour usage, you might be able to optimize for meeting your projected demand, but in that case your calculated demand will not be correct, because the calculated demand is calculated as a function of the total available labour vouchers people have available, and by minimizing labour usage, you also give out less labour vouchers. Even worse, the next time demand for products in real terms will be less (less vouchers, thus lower total demand) so you will lower production, eventually producing nothing. So in this last case, instead of taking what society said they want to work and what they have available, to make as much consumergoods as possible, you ignore all that and just make whatever you think society needs, which will be wrong as explained before.

The whole point of the system cockshott proposes is not to try to predict demand. You primarily look at what was produced last time, and if proportionally too much was made, produce proportionally less, and if proportionally to little was made, produce proportionally more. Then, on top of that, you can use statistic on past data to predict seasonal trends, or use your common sense to not start building factories some something that is merely a temporary hype.

Oh right, I found a simple solution for this.
Products:
Ratios:
Constant amounts needed:
Then restriction:
You simple subtract the constant amount from the amount produced and multiply it by the inverse (divide it by) of the ratio.

I tested it with the solved in libreoffice and it works. It gives production distribution where the amounts of products produce after subtraction of the constants is proportionally equal to the ratios set.

Attached: soffice.bin_2018-07-10_12-42-40.png (614x230, 11.22K)

how would you make sure quantities of products meet demand then?, the only way i can think of that can be done with ratios is to do a shooting algorithm, and that takes solving optimizing several times with different ratios to see which produces enough things.
anyway if this is to continue anyway we have to attract more people, only two posters can't program this whole thing

yeah this is not going to work, in order to how much little and how much more you have to do a shooting algorithm and thus applies

what i propose is maybe keep the ratios for some products, the ones that are key, those we maximize, but for others such as consumer ones why produce more that is needed?, the plan should just yield enough so that the labour costs of those are reduced by mop growth, and more labour can bee invested into new things, like tecnologies and such

No.
The demand for products does not only depend on their labour value and their amount, it also depends on the availability of other similar and dissimilar products, and their availability and price, as well as the amount of labour vouchers people receive. The whole point of the cybernetic in "cybernetic socialism" is that you look at the disparities in the previous manufacturing cycle between the real price and the labour value, and then change the ratios to try and approach the ideal.
I have already shown you how your proposed "idea" is mathematically impossible to plan for with linear optimisation. You need something to maximize or minimize. Minimizing labour means the population has less labour vouchers, and thus your projected demand is way off, and maximizing products is impossible as you can only maximize for a single variable, which, without ratio restrictions, would mean you gear your entire economy to try and produce your set constants, and then produce as much of whatever random product you chose as humanly possible to use up all the other resources and labour.

The whole point is to not plan for definite constant amounts of products, that's literally the whole point of cockshott's system, but to plan for a ratio of consumer goods, exactly because we cannot know precisely what consumer want, in what quantities, and how much each consumer should receive of the produced items.

Your idea boils down to making a set of definite quantities of stuff, not using up all of the available labour that society wanted to assign to producing consumer goods, and then ration it pretty much arbitrarily, because your projected demand by definition took as argument the amount of labour vouchers available in society, and then it does not use it, meaning that there is less vouchers than you expected, and thus less demand, creating a whole new completely unpredictable demand because the price elasticity of products is not equal, IE food is prioritised over new jeans.
It is this system, your system, that would require a shooting algorithm, as you would have to find a solution where you can exactly produce your projected demand with the almost exact amount of hours you calculated it for.

Cockshotts system on the other hand, does not care about whatever predictions we make, we do not have to predict. We just have to change the ratio of the vector in the direction that the price:value relation tells us to.

What line do you draw for "other consumer goods" and at what point do we produce "more than needed"?
The whole point of the consumer good market is that it produced goods that are by definition scarce, whose demand will increase if more labour time would be available to consumers. If a product gets to the point where their price does not affect their demand anymore, IE drinkingwater, because there simple is no more need for it, then it can be removed from the consumer market and produced as a social service. In this case, such a good would not be scarce, it would be post-scarce, where its price really has little impact on it, and people will just use the same amount of them regardless of if they are free or cost a little bit.

(Of course, we can use predictions on top of that to make sure we don't over or undercompensate, or to cancel out things like seasonal cycles that are known to exist)

*more labour vouchers
To rephase: it is for goods that will be bought more if people had more money/vouchers

the only problem i have with this is that moving the ratio of a product doesn't only more the amount of that product made, it also moves other products whose materials and labour force are needed in order to make that other thing you want to make more of, and i know what you are going to say "oh yes that labour and materials come out of the products you want to make less of, those are the ones you decrease the ratio for", the thing is that you have to add up all the effects of all the ratios, and after doing that you have an economy that widely changes on just a few numbers, and you have no way in which to know how anything is going to change except by doing a round of calculations, or how do you think this
is going to be done?, by trying out different ratios and using them as inputs in a shooting algorithm, this is literally the only mathematical way to check you don't over or under produce, the other option is an economy that produces things in a really uncertain way and nothing is checked, like imagine this, people buy a little less potatoes this week, oh well lets put in a ratio of 4 instead of 5, and oh it looks like people bought a lot more hemp also, so lets increase it from 15 to 20, and now you have a virtual potato shortage, with prices over the roof, because the algorithm assigned all the farming equipment into hemp instead of potato, and if you spot this in the planing stage then congratulations you have to modify the ratios and try another time, and how do you do that again? oh yeah there is this thing called a shooting algorithm

this is why no matter how you look at it there needs to be some kind of quotas in the algorithm, that would prevent the shooting and i'm not saying i want to take away the maximising function, all i'm saying is why maximize things like potatoes?, like making as much potatoes as possible is only going to decrease their value, and you already roughly know how much of the thing people buy every day, it's not a changing quantity, so why not just include it as a constant, and focus the maximizing function on other things

so this is your argument, demand is hard to predict so it shouldn't be predicted, instead we should move the ratios by feeling without checking?, i already said how this was an awful idea

i never meant this, i meant getting out of the maximising function products for which you already know the demand, and yes there is things you should absolutely maximise, but consumer goods are not one of them, and to answer this
at the point where all the prices are falling because the plan maximized things wrongly

It is this system, your system, that would require a shooting algorithm, as you would have to find a solution where you can exactly produce your projected demand with the almost exact amount of hours you calculated it for
no not really, if i set up an equation inside the linear programming that says that product=demand, then one simple round of planning will make it so that product is produced in that amount, while maximizing other things like clean energy production or growth or investment, if anything this is designed to cut off calculation because you now have less ratios to move, because that's the thing you are assuming i'm just going to set everything constant, but no the maximization still will take place, except we will no longer have a million ratios to look out for, except a few key ones, like a couple tens of thousand at best, and not a fucking billion like you want to, and thus it will be simpler

whether you want it or not you are making a prediction here, that is the amount you have to change ratios, that amount determined the number of things more or less to produce, making more educated choices than this will require marketing research

You don't need to know exactly what will happen, and it is impossible to know. It is also impossible to know with "predictions", which are based on imperfect data from the past with completely different consumer preferences and production capabilities.
It is completely by design, it is the point of it all. If some product has a way higher price compared to its value than all of the other products, that means we make too little of it, and all the prices of the other products will necessarily be lower, because the increase spending on vouchers on that product decreases the amount of vouchers than can be spend on other products (all prices add up to the total amount of vouchers after all). So by increasing the production ratio of that undersupplied product, you decrease the production of the other products, both decreasing their supply to raise their price to match the labour time, and freeing vouchers from being spend on expensive items.
We do not need to know exactly what will happen. We already know changing the ratio of a product will have the inverse effect on all other products in an equally distributed amount.

You cannot mathematically check if we over or underproduce. We cannot mathematically model consumer preferences, inter-product replacability and price elasticity and changes in those things.
Your entire argument hinges on the assumption that "projected demand" is 100% true. Not only is it impossible to predict demand for all consumer goods if we change the production of all consumer goods (since all consumer goods have the abovementioned things), but also calculating demand depends on the available labour vouchers, which in itself are a variable produced by the productive distribution of labour generated by your projected demand.

First of all, youre absurding it, implying that if you tell the algorithm to make slightly less potatoes we will have no potatoes at all. Second of all, you cannot "spot it" in the first fucking stage, because you can not predict consumer demand to reflect reality, as explained before.

Yes you litterally do. You litterally want to throw out the whole economic essense of TANS and everything cockshott wrote.

Because as I already explained, if this is the case YOU DONT NEED TO PLAN IT IN THE CONSUMER MARKET. IF IT WILL BE A CONSTANT AMOUNT REGARDLESS OF PRICE YOU CAN JUST MAKE IT FREE AND LEVY TAXES TO MAKE THEM. IT WILL BE A PUBLIC SERVICE.

Memetalk aside, but try reading cockshott and understanding it.

Now I will be off to diner. Please, try to read what I wrote. Your whole position hinges on the assumption that either all products have a finite, inchanging, demand, which will not increase if price approaches zero, which is not correct for the vast majority of things we produce, or that you can accurately predict demand for all products despite changes in consuming patterns, changes in products themselves, changes in the prices of other products (increasing or decreasing disposable income), changes in manufacturing efficiency, changes in total hours worked, etc. This is not possible. You cannot predict demand for all things from limited data while literally all the other variables changed, this would require a perfect or near perfect model of reality, which would either require omnipotence or infinite amounts of data about the past and future, both of which do not exist.

Now if you are going to keep repeating that you want to implement "cybernetics" while rejecting cybernetics, and saying "we should maximize, but not for consumer goods", when consumer goods is litterally the only thing we should maximize for, I am not going to keep responding to you.
Have fun writing your mathematically impossible god-like planning system on your own then.

you say this but you entire plan hinges on how well prices can predict future demand, so you implicitly are assuming we can know if we need to produce more or less than now, depending on whether the price is lower or higher than value, now how much of a jump would be to say that the difference between those two is related to the amount to want to increase or decrease the product?, surely you don't belive that if the price of product x is 400 hours more than it's value and the price of product y is 20 hours more than its value, that the production of both should be increased by the same amount, surely you can see that the demand of x is greater than y, and thus more x needs to increased than y, this whether you want it or not is prediction, because in your system, instead of directly predicting a number for the change demand of a product you are predicting the change in ratio for it, this change in ratio is directly related to the amount of product that changes, so whether you want it or not you are already making a value judgement, but instead of being in the number of the demand you make it in the ratio change, and if the change in ratio doesn't need to be exact, then why does it need to be that way in the demand prediction?, you already know the current demand, and you know the price, pressumably you could increase it or deacrease it in the same fashion as the ratios, explain to me how these two value judgements are different?, you even cite TANS where you well know that in that book there is no ratios, and you only modify production of things by setting higher or lower targets of the thing, WHICH I WHAT I FUCKING WANT TO DO, how is this against what cockshott wrotte when i'm literally stealing the idea from him?, so can you please end this meme of "you can't predict anything"

lets go with this for a moment, so increasing the ratio of a product increases it's production, and decreases the production of other products, yeah i think we can agree on this, now which products?, to see this lets take an example, say you have an economy of four products, a, b, c and d, they are set at ratios x, y, w and z respectively, then i modify the ratio x to be twenty more so from x to x+20 now the relation of products a to b goes from being x/y to (x+20)/y, , but as you didn't modify the ratio of y/z or z/w, now as we said some product needs to be produced less, so that something is going down, however to keep the ratios y/z and z/w that you didn't modify ALL PRODUCTS need to be produced less, the opposite also must be true iif you decrease something all other products must be produced more, so incrementing one is going to decrease all, got it?, well now imagine that not only one product needs to increase stock, but lets say 2000 need to, well you need to take into account how the increment of every single product affects the rest, and i don't think you have given the thought to the math related to this

no, you didn't understood, i'm saying that increasing hemp already decreases potatoes, and that deacreasing potatoes also increases hemp, and that both effects need to be taken into account, otherwise unexpected things are going to happen, and doing this kind of calculation for 1 billion products is VERY VERY VERY difficult for an entire economy

what i meant by "not changing quantity" is that a quantity that changed very little, this is just like healthcare is, the use of healthcare changes, due to public health crisis, changes in population and such, but normally it won't change that much, there is going to be a shit ton of products like this, people don't buy more potatoes from one month to another, or rice or any other quantity of things, all i'm saying is why can't you just make those services?, it has a clear advantage as in decreases the complexity that i talked about by reducing the amount of ratios you have to fiddle with, MY PLAN IS NOT EVEN THAT DIFFERENT FROM YOURS

you do know that tans doesn't use ratios right?, it uses expected demand like i want to, I LITERALLY STOLE THAT IDEA FROM TANS YOU MORON, i just want to have it so that it's not all products, but just the ones with unchaging demand like elementary food things, those don't change, and you use a simmilar function with ratios for the rest, and about , i did sayi wanted to get rid of them there, but to be fair it was like midnight and what i wanted to express but didn't was what i said here

now who is the one that doesn't read?, if you what i propose demand won't stay the same ALL the time, but will be modified as price changes, i never said it was going to be the same

you don't need to, it doesn't have to be totally exact, you can just produce a bit more if prices are up and less if they are down, my objective here was not to predict the demand exactly and then use that exact number to guide production, i just want to decrease the fiddling with ratios you have to do (not eliminate it mind you), because FIDDLING WITH RATIOS IS FUCKING COMPLICATED, i want you to pass that through your head

Hey mate have fun with your "im not going to do ratios" and "i dont want to maximize anything" bullshit, I am done with your retardation to be honest.
The argument that "a billion equazionz iz haaard!!" is also retarded as cockshott shows it was possible, back in 2007.

You wan to produce definite amounts of products which completely makes your prediction of prices impossible exactly because the demand for things is caused by how much labour vouchers are available, and how much labour vouchers are available is directly the same as the amount of hours worked, yet you want to predict prices and not use up all the hours, or worse yet, you want to do away with the democratic decisions of society. They chose to assign a % of the total time to increase the MoP, a % to consumer goods and a % to social services. You want to decide, for them, what ought to be "the right amount of products".

Lastly, you say I only want to spitball shit, pretending I dont realise that if something has a 100000% difference between price and value that it will be different if there is a one nanosecond difference. You pretend I dont want to use predictions and past data to make the correct changes.

Except that all products are labour, all products are highly interconnected, and the price of other products changes as a direct result of the changes in spending habits as well, not just production.

That is not cybernetics, it is not revolutionairy, its just the same that the USSR did.

Well, have fun with your stupid ideas and your autistic special snowflake programming language and your useless wikia. I will hide this thread from hereon out.

bunp

and this is why STEMfag communities suck ass

Well what am I supposed to do? How much of a community is it if theres only two people? Just stay and waste my energy?

Organize with someone in your geographic proximity? Organize with an existing organization? Start as a part of international organization as a platform for doing the research that goes in prototyping an implementation of planning software?

Maybe doing a survey of organizations that are actually interested in this kind of research. Or how this can be incorporated into existing praxis. Even if it stays on a level of series of lectures for internal education of the organization or as public lectures. It is just that there is no formal institute that trains people in this. Only scientists working on this do this to publish papers as those are a metric in science.

Fair enough, though I would have to find an organisation nearby in my classcucked country that would be interested in it, as well as someone with a background in CS who could be interested.

I really should get around to joining the sucdem party some time soon.

(wtf i appreciate your comment, it motivates me to continue learning about planning and all that)

btw

am very tired

this research is already being done, just not in the communist sense, in fact the kantorovich shit i learned in college, it's just that the way they taught it to me was as a corporate tool to use resources properly, of course the field is mired with ideology, but the information is still there, in fact some version of this software must already somewhere exist i posit, it's just that they use it in companies, in fact witness the people that will plan your economy in the future
pubsonline.informs.org/journal/moor

Then we simply need to put in the effort translating this already existent knowledge into communist theory, and of course educating people in it. Right? I'm convinced that the most important thing for us to do at the moment is relay information.

another programmer thread: