Agents in relation to process events

[fosterlynn]

We have a general question on if agents should have a place in processes, or do they just relate to transfers, i.e. which agent has which role (rights and responsibilities) in relation to a resource. From a gitter chat, too hard to find it now. 😞 But I think this is maybe the biggest conceptual hurdle left in the core model after our marathon discussions.....??

I think there is a lot of agency involved in processes, and several kinds of process related events require one or possibly two agents to have complete information. For example work, load, unload.

@elf-pavlik please add your thoughts or objections to this, I don't think I can do it justice and don't want to mis-represent.

[bhaugen]

Unless you got some robotic agents, your process will need some human agents.

[elf-pavlik]

I think there is a lot of agency involved in processes, and several kinds of process related events require one or possibly two agents to have complete information. For example work, load, unload.

If we model work as vf:ServiceResource, than we can reference agent as provider (owner doesn't seem to fit services well). For scheduled services (possibly modeled as planned resources) we also need provider and recipient information.

If we want to capture the actual work of loading and unloading... again we have work which we can represent as vf:ServiceResource which in turn also can reference load and unload events (we would need to define property for it, but we will also need something like that for referencing load and unload from the transportation service - output of transportation process)

Maybe for each YAML snippet where you reference agent directly from a process event I will try to write alternative where we have a resource with a relationship to an agent? If you paste couple of links to existing YAML from other issues I could give it a try.

[fosterlynn]

If we want to capture the actual work of loading and unloading... again we have work which we can represent as vf:ServiceResource

I was thinking more about this too. I think you are right about load and unload - creation of value all comes down to work in the end. It also takes work to create a service (and sometimes other inputs too).

Load and unload provider, carrier, receiver might be just context agents? Would need to do some examples to be sure, because different viewpoints might impact this, I'm not sure. Context agents have agency too, but I think you are .

Then again some people might not want to record work, but only record what was created/produced. Could be a service or a material item.

If we model work as vf:ServiceResource

I'd like to understand better why you don't like vf:work? Seems to me to clearly and directly describe what is actually happening. Which is what I think rea:EconomicEvents should do.

I think if some people would rather log their work as a service, that's fine too. That would be for situations where they think their actual work is not anyone's business (can be valid), or people are working very independently, or isn't useful to log for whatever reason.

But vf:work events can also be used in collaborative situations to coordinate the work among several people in a process, make it clear what everyone is doing so others can see it. That transparency is a useful part of an open collective process. As a more edge example, in Sensorica, they have also used work events to log results and thoughts while doing R&D, keeping running scientific lab notes so everyone can see results as they go along, see what has been tried already, etc.

[elf-pavlik]

@fosterlynn in https://github.com/valueflows/process/issues/21#issuecomment-238099929 you wrote (I copied only the vf:work input)

'@context': https://w3id.org/valueflows/v1
'@id': https://fablab.example/818f30f4-c119#process
'@type': vf:Transformation
'skos:note': making a 3d printed part
'vf:contextAgent': https://fablab.example/OhioFabLab#agent
'vf:io':
- '@id': '#input-4'
    '@type': vf:ResourceEvent
    'vf:action': vf:work
    'vf:eventDate': 2016-08-10
    'vf:resource':  '@id': https://fablab.example/3d-printing#typeOfWork
    'vf:fromAgent': '@id': https://fablab.example/FredFlintstone#agent
    'vf:eventQuantity':
        '@type': qudt:QuantityValue
        'qudt:unit': unit:Hours
        'qudt:numericValue': 2.5

We could rewrite that work event as

'vf:io':
- '@id': '#input-4'
    '@type': vf:ResourceEvent
    'vf:action': vf:input
    'vf:resource':  
      '@id': https://fablab.example/3d-printing-by-fred-on-aug-10#work
      '@type': vf:WorkResource
      'vf:model': https://fablab.example/3d-printing#typeOfWork
      'vf:location': https://fablab.example/main-hall#place
      'vf:provider':
        '@id': https://fablab.example/FredFlintstone#agent
        '@type': vf:Person
      'vf:date': 2016-08-10
      'vf:quantity': 
        '@type': qudt:QuantityValue
        'qudt:unit': unit:Hours
        'qudt:numericValue': 2.5

We would need a way to capture that 'one person working for 2.5 hours' vs. 'n persons working for 2.5 hour' If for example we would have vf:WorkResource sub class of vf:ServiceResource that implies just one person working. Now when we cast an intent, we can create planned service and reference input of a process which we want to use (following relating Intent to a Process, and to other resources via its inputs/outputs

'@context': https://w3id.org/valueflows/v1
'@id': https://fablab.example/818f30f4-c119#process
'@type': vf:Transformation
'skos:note': making a 3d printed part
'vf:contextAgent': https://fablab.example/OhioFabLab#agent
'vf:io':
- '@id': '#input-4'
    '@type': vf:ResourceEvent
    'vf:action': vf:input
    'vf:resource':  
      '@id': https://fablab.example/3d-printing-on-aug-10#work
      '@type': vf:WorkResource
      'vf:model': https://fablab.example/3d-printing#typeOfWork
      'vf:location': https://fablab.example/main-hall#place
      'vf:date': 2016-08-10
      'vf:quantity': 
        '@type': qudt:QuantityValue
        'qudt:unit': unit:Hours
        'qudt:numericValue': 0
      'vf:expectedQuantity': 
        '@type': qudt:QuantityValue
        'qudt:unit': unit:Hours
        'qudt:numericValue': 2.5

Having above, we could reference that particular instance of planned vf:WorkResource from vf:Request.

'@context': https://w3id.org/valueflows/v1
'@id': https://fablab.example/intent-2342#request
'@type': vf:Request
'@id': https://fablab.example/3d-printing-on-aug-10#work

I find it fundamental for making conscious decisions that intents have requested resource placed in IPO chain. I consider nowadays common practice of basing our choices on thinking: 'what i can get in return for my work' much inferior to approach of making choice based on reflection: 'what impact my contribution will make, and how i can benefit from doing that work'.

@fosterlynn how do you imagine casting an intent for work contribution to particular planned process which has particular planned output (very likely already 'earmarked' as input to another process in IPO chain)

[bhaugen]

Some examples of process work that do not involve loading or unloading:

• carpentry and other kinds of woodwork
• metalworking
• turning pots on a wheel
• working dough in many ways: kneading, forming loaves, toppings, fills, etc.
• cleaning
• medical and other kinds of diagnosis
• repair work

I am sure several hundred others exist.

[bhaugen]

I consider nowadays practice of 'what i can in return for my work' much inferior to approach 'what impact my contribution will make, and how i can benefit from doing that work'.

Would be good to develop that thought more precisely.

However, as I have explained a few times in VF issues, reciprocity, however indirect, will be required to keep communal mutual coordination for-benefit economies alive. Including your ability to continue to work.

[fosterlynn]

I find it fundamental for making conscious decisions that intents have requested resource placed in IPO chain.

Thinking about work resources -

If there is a vf:Resource, I see the resource not as the vf:Agent, but more as the labor power of the person. So maybe the "Pavlik can work max 21 hours per week" is a resource. Of course the agent has as much control over his/her work resource as possible - not much in slavery, more but not enough in capitalism, hopefully more in the future. (Although I do agree with @bhaugen that a community can't function without some notions of reciprocity.)

The event consumes some of that resource.

If people do calendar scheduling or need max time for planning, then the resource is useful. On a calendar, not sure of the granularity people would need for a vf:Resource, but in any case it would always be some time period. Or I could see that if you need to schedule a particular person but don't do calendar related scheduling, there could be a non-time-specific vf:Resource of a person's labor power. Otherwise there really isn't a vf:Resource yet, just a vf:model I think. And an event could reference the vf:model and the person doing the work as a vf:Resource. Not a proposal yet, but trying it out, and moving it back to the event structure we agreed on for other types of events:

'vf:io':
- '@id': '#input-4'
    '@type': vf:ResourceEvent
    'vf:action': vf:work
    'vf:resource':  
      '@id': https://fablab.example/3d-printing-by-fred-on-aug-10#work
      '@type': vf:WorkResource
      'vf:model': https://fablab.example/3d-printing#typeOfWork
      'vf:provider': 
        '@id': https://fablab.example/FredFlintstone#agent
        '@type': vf:Person
    'vf:location': https://fablab.example/main-hall#place
    'vf:date': 2016-08-10
    'vf:quantity': 
        '@type': qudt:QuantityValue
        'qudt:unit': unit:Hours
        'qudt:numericValue': 2.5

I do think we need to know an event type of work, as it has specific behavior.

@fosterlynn how do you imagine casting an intent for work contribution to particular planned process which has particular planned output (very likely already 'earmarked' as input to another process in IPO chain)

Do you mean how does a person commit to doing some work already scheduled? Or how does a person request some work or service from another person?

I consider nowadays common practice of basing our choices on thinking: 'what i can get in return for my work' much inferior to approach of making choice based on reflection: 'what impact my contribution will make, and how i can benefit from doing that work'.

I completely agree. I don't see how the vf:work approach is not that. It is all about a person contributing to a process or a context, in its most direct representation. If the person gets something in exchange for the work, OK. A service also can be generated directly to create an exchange at the same time it provides something useful, in fact we have had to do that in other use cases. Or we can think of them as work being direct contribution to creating something and service being something which is useful to a process but buries its internals either in a previous process of creating the service or as something not useful to log, more as the results of the work than the work itself.

[bhaugen]

The Quick Response Engine (QRE) was a skunkworks project for an ERP software company I worked for in the 1990's, and was also my first implementation of the set of ideas that fed into VF.

We did shift scheduling for people and equipment in that project. The inventory of such time-based resources was a shift calendar composed of 15-minute slots. It was implemented as sparse array [edit: actually a skip array, which I can't find a reference for on the Web, curiously...] composed only of slots where the resource was scheduled to be available: for a person, the shifts they were scheduled to work. Each time-based resource had a shift calendar. A process could reserve as many of these 15-minute slots as it needed, assuming they were not already taken by some other process.

[elf-pavlik]

another example of vf:WorkResource in https://github.com/valueflows/process/issues/22

[elf-pavlik]

how do you model in NRP a person making commitment to planned work input of planned process? how does it relate to that planned work input before someone committed to it?

[bhaugen]

how do you model in NRP a person making commitment to planned work input of planned process? how does it relate to that planned work input before someone committed to it?

The planned work input before someone committed to it is a work requirement for the process. It appears in on the process and on the My Work pages of people whose skills it fits with a description and a button that says "Take this task". Once somebody takes it, it goes in their "What I am working on now" section and disappears as a task that other people can take.

(Edit: also an email notification goes out to everybody with the required skill.)

[fosterlynn]

And in terms of the model, the planned work is a Commitment without a from_agent, and then when someone takes the task, we just put that agent in the Commitment.from_agent. (Not making any VF recommendations particularly.)