Open Source Ecology Forums

gregor

We currently have no method by which to estimate the cost of a system or design. For example, an estimate that included the following:

Maintenance cost/year, materials:

Maintenance cost/year, labour (low skilled person-hours):

Maintenance cost/year, labour (highly skilled person-hours):

Capital materials cost:

Capital labor costs (highly skilled person-hours) (list the hours, followed by the specialty where possible):

Capital labor costs (low skilled person-hours):

Cost of any non-flexfab production equipment:

Total production equipment requirements including flexfab equipment (list the equipment):

Let's face it, this current approach of just talking about capital materials cost is not really a full accounting. The other costs are not at all insignificant and we should at least be keeping track of them. We need to define high and low skilled labor, though. I think we can do so by looking at the training requirements that 90% of the population would need to get the skills needed, perhaps.

We should agree on definitions for the others too. I suggest some at http://opensourceecology.org/wiki/Solar_to_electrical_energy_conversion_system_Development#Performance.2C_specs_and_cost_estimates_for_prototype_I

This can be used for both estimating, and tracking the costs of equipment that is actually produced. I think it's clear this is a tool we should have that will help with cost reduction, yeah?.

OmniPhoenix

Yes gregor, great stuff! I like using spread sheets with variables to keep track of things and compare different outcomes.

Just some quick ideas:

Maybe some kind of complex spreadsheet could be used, it should take into account the different options like off the shelf cost (for materials, etc) and estimated costs for producing the means to produce the materials (maybe referenced from it's own page with it's calculations for the different options), etc. The cost of raw/off the shelf materials could be a variable, and it could also determine when it makes sense to buy/produce something.

It could be designed (or just make sure things are uniform enough) so a more complex process could be made that would sort the projects and help determine what would make sense to do first, and it could help determine at what point something else makes sense when the variables change.

Maybe the systems analysis stuff could use spreadsheets to track
energy/materials input/output variables to graph what the results will
look like and measure how different variables affect the system. Tracking things as a whole system can also help take into account shared costs, but it would be difficult to define how things are shared if something can only be used in one way at a time while something else also needs that thing.

It should also avoid unnecessary feature creep. Very complex spreadsheets can sometimes be too complex though which might make it too difficult to understand the results. It's kind of the same problem that any highly complex system has though, it becomes so complex that no one person can understand it completely and thus no one can recognize if the final results are incorrect due to some small bug along the way. I've worked with programming large projects kind of like this and it's a common problem. So to state it positively would be to design it to be easy to sort things so the useful information is easy to see.

gregor

Yeah, that would be good, can you make a basic spreadsheet like that in Calc? Where should we put it on the wiki?
There are a lot of things like the wiki maintenance policy, CAD standards and so forth that should perhaps be put under the category (they can remain within their existing categories too) "Organizational operation at OSE" or something similar so people know where they can see all such things so they are aware of them. When the hotCat gadget is working I will do that, I have added it to my list of things to do.

In a way we are just summarizing the build instructions/design I guess. This can be done at various levels, either to assemble a list of things that need to be bought or you could just present the number which they in total are expected to cost. We'll just come up with something reasonably useful to be getting on with I think though, we don't want it to be too lengthy.

OmniPhoenix

I could try making a basic spreadsheet for stuff, it would help if we had a list with some more ideas and specifications for what it should include. I've been quite busy lately though, but it should go faster the more clear the specifications are made. It shouldn't really matter where things are on the wiki at first, they can be moved around to be organized once things get more well defined.

So it looks like we want to have estimates of what a tool will cost before it is designed, and estimates of what it would be in different cases? And we want to have post design estimates from a bill of materials to measure costs in different cases? There are a lot of questions like this to answer for the spreadsheet specifications and for project planning.

Anthonyrepetto

I'd advise using another interface. Spreadsheets may be fine for archiving the data, but the flow between them will be easiest to follow if it's spatial, pictorial, filterable, and allows zoom. Have you looked at 'mind mapping' software, like http://www.thebrain.com/? Also, I understand that Pattern Language is a common starting point, but there have been advances in systems diagram techniques in the last 30 years, and unaddressed pragmatic realities that are familiar to fans of Wittgenstein. First, on mapping the flow of the system: in addition to labeling the 'objects', you should label their connections. If you want to go a level deeper, name their qualities, too. Now, you have nouns, verbs, and adjectives. You can add 'external criteria' (prepositions) and 'combinations' (conjunctions), to get something closer to a real language about the system. Using nested labels for generalizations of each part of speech allows an object to appear under multiple 'patterns', according to its real use. You only have to define the aspects that need to be distinguishable. So, if a project uses 2by4s, you'd list '2by4' as an object. If you have to cut these to length, add the process 'saw wood to length', which requires one of a variety of saw tools, and you must now specify the 2by4 as 'wooden', to receive the 'saw' verb. Further into that project's specifications, you need 2by6 planks; both 2by4 and 2by6 are objects, but they also fit in the category object, 'dimensional wood'. 'Metal rod' would be its own object, but if it had a function that could also be served by wood, they would have a shared label, 'support beam'. Each component, process, and quality is added as it is reached in a project, and labels group them according to need. This functional evolution of a full 'pattern language' is more likely to fit the users' needs, than a top-down lexicon and syntax. (Sorry, Alexander, Wittgenstein was right; a language has to evolve, and is only comprehensible to those who learned it as a gestalt, not as a strict diagnostic.) Plug-in measurements of inputs and outputs, from every practitioner around the world, to develop an allocation algorithm. Given this software, anyone could plug in their local resources, and model various 'production outcomes', comparing their rates and requirements to find a fit for themselves. I'll be going into this sort of mapping in my concept wiki, Anthony Repetto/Concept Log, and hopefully have some good pictures, soon.

Andrew

I think that each individual is likely to follow a slightly different process. I'm currently replicating the power cube, and I'm sure my process is different than the one used at FeF. I'm documenting my progress at http://aispina.blogspot.com/. I'm recording my costs and fabrication time here (google doc). I think a very important part of recording costs is collecting lots of data points. The credibility of OSE will grow as more people demonstrate affordable results. This implies a need for an easy to use cost recording tool. The discussion above seems to be tending towards complex tools. These may be useful for 'core' OSE documents, but I don't think will universally solve the problem of collecting cost data.

Howdy, Stranger!

Categories

In this Discussion

Tagged

Open Source Ecology Forums