New value equation discussions [was: Milestone 2 - Pay day !!]

1 view
Skip to first unread message

Bob Haugen

unread,
Sep 21, 2015, 8:25:56 AM9/21/15
to SENSORICA, PV characterization project forum
I propose that we wait to start the Loomio discussions I listed below (#2 and #3) until we have fully resolved and understood the discrepancy issue (#1 below). See Lynn's thread entitled "PVC Distribution Milestone 2 parallel test".

I believe that the resolution of that issue will shed light on issue #2, transparency of the value equations in NRP and spreadsheets.

And I believe it will also have repercussions on issue #3, Tibi's Role ontology. That issue is a lot more complex and may seem to be unrelated to #1, but I think at least that it will be difficult to focus on #3 while #1 is unresolved.

I said I would kick off the loomio conversations this morning, but would rather wait. If anybody else wants to kick them off sooner, go for it.

On Sun, Sep 20, 2015 at 11:06 AM, Bob Haugen <bob.h...@gmail.com> wrote:
I see at least three sets of issues in this message (below) and would like to separate them for discussion, and give them accurate subject headings.

1. Discrepancy of results between Lynn's value equation and Tibi's spreadsheet, which seems the easiest to resolve.

2. Transparency of the value equations in NRP and spreadsheets, or value equation math in general, which is a bigger topic.

3. Tibi's Role ontology, which is an even bigger topic, as I hope will become clear.

Is it ok with everybody to separate these?  If so, I propose that we start a separate email thread for the first one, and start Loomio threads for the other two. If anybody else sees any other issues, or wants to separate them differently, please go for it. 

Lynn is looking into the discrepancy, but has some other work to finish up first, so she'll kick off that thread when she gets to it. I can start the Loomio threads by tomorrow morning latest, unless somebody else wants to do it first.

P.S. I think each of these issues is very interesting and may shed light on some other as-yet unresolved issues. You gotta get ready to understand the details in each case, though.


On Sat, Sep 19, 2015 at 7:13 PM, Tiberius Brastaviceanu <tiberius.br...@gmail.com> wrote:
Hi Lynn, I don't get the same results when I run it in the sandbox.

I can verify my calculations in the spreadsheet again. I essentially modified the calculation for Milestone 1, which was tested quite a lot. But I can't say that it's perfect.. I'll do my part.

A few months ago we had a discussion about the transparency of the value equation, the way it is implemented in the NRP-VAS. I don't know what goes on mathematically, behind these lists and the filters. Therefore, it is impossible for me to formally compare the one in the NRP with the one on the spreadsheet, the later is transparent, because you can access the code in each cell. Not getting the same answer tells us that they are not the same. There might be an error in the spreadsheet, but I can't go into the one in the NRP.
I think we need a math editor in NRP, at the UI level, when we build the value equation, and we need to renter the algorithm to everyone, once a value equation is implemented.

See the value equation agreement for the project. At the end of this doc you have the mathematical algorithm for calculating distributions. This is what I implemented in the spreadsheet. See if you are doing the same in the code. Perhaps get another programmer's advice, or second opinion... Once you render/surface the algorithm from the value equation I can tell if they are formally similar. If you need more clarification on this, I can help. I do have some ideas about how to render the formalism from your code.




IF YOU ARE INTERESTED IN THE VALUE EQUATION

To shed some light on the problem in this case,
Bob and Lynn, you can correct me

We need to filter the data that goes into the value equation grinder.
From the database, we apply the first filter, the Project.
After that, we need to make sure which Process within the project is concerned.
Within these processes we have Types of work (documentation, R&D electronics, mechanics, etc.)

In the PV project, portions of the budget are allocated to Processes (Outreach, Coordination and Facilitation, Prototyping) and the value equation assigns weights to Types of work (electronics, mechanics, software).
 
After this filtering is done, we need to place the data into different components of the value equation algorithm, which should be formally sound. This part is totally obscure for the user, including those who have admin access, like myself, I think it is hard codded.



IF YOU ARE INTERESTED IN THE LARGER DISCUSSION ABOUT VALUE EQUATION, -- HIGHER LEVEL

Thoughts for an increased level of abstraction, cover more cases and make it more flexible
In the very beginning I proposed the Activity ---> Role system
The idea was that a Role is a cluster of Activities.
Activities can be: writing, reading, experimental work, manual work, meeting, communicating, etc.
A Role is any cluster of those.
Example: Role/Administration = {writing, reading, meeting, ...}
In order to structure activities and roles, I proposed to use an ontology.
The Activity ---> Role system would be emergent.
Roles would be used in value equations.
Bob rejected the ontology idea and went simple with Type of work. So we have Projects, Processes and Types of work. But there is a problem here, because there is no distinction between activities and roles any more. I can do writing, or documentation in different contexts: R&D or as part of administration, etc. In other words, I can do the same activity in different roles. In the context of the same project, multiple people can log documentation in different roles. Should all documentation work be rewarded the same way? Most people will say NO, because technical documentation requires different skills than administrative documentation.
At this moment, Documentation is a Type of work, and it becomes harder to distinguish on the type of documentation in order to reward it differently. If the value equation runs on Roles, it becomes more flexible.



On Sat, Sep 19, 2015 at 8:10 AM, Lynn Foster <foster...@gmail.com> wrote:
Tibi, check my test value equation PVC M2 test 2 - the types of work are defined on the bucket rules as filters by resource type. 

There are two levels of filtering, one that is defined more permanently in the value equation, meant for things that could apply over and over to many distributions, and there is the process or work order or deliverable selection that happens at "run time" because the nature of those selections is that they will always change.  No matter which level a filter is defined or entered, they are combined when the distribution is run.  It is an AND for the geeks among us.

(You are defining a new value equation each time because of the % estimates applied to the buckets, which is fine, but originally the thought was that projects would define a value equation and use if for much of their work.  It works either way.)

On Fri, Sep 18, 2015 at 10:28 PM, Tiberius Brastaviceanu <tiberius.br...@gmail.com> wrote:
Lynn, we can select specific processes to input time contributions data into the value equation, but we can't go further, to select specific work types within a specific process. 
In the milestone 1, we put all the R&D into one bucket, electronics, software and mechanics, because they had the same $/h rate. Milestone 2 is different, we need to consider these differently, and they were logged in the same process. 

On Fri, Sep 18, 2015 at 11:24 PM, Tiberius Brastaviceanu <tiberius.br...@gmail.com> wrote:
We can select processes for inputing data into 

On Fri, Sep 18, 2015 at 7:58 PM, Lynn Foster <foster...@gmail.com> wrote:
​>I did a calculation for the distribution for Milestone 2. I did it in the spreadsheet, since the value equation engine cannot support the way data was logged, for this budgeted project, a forward-looking value equation. We haven't reached the level of abstraction that can support all the different use cases with the Value Equation. More conceptual work needs to be done...

I'm curious - What makes you think the value equation can't support the way the data was logged?  It worked fine for the first distribution.  I know of no reason it won't work for the second.  It is only a question of what percentages are chosen for the buckets - the original estimates ("forward looking") or the re-calculated actuals, and the logic is the same for the value equation, it doesn't care.  You have divided the buckets more finely this time, but that is easily handled by the type of work filter.  Am I missing something?  (I haven't dissected your spreadsheet calculations in detail.)

I think it is still a good idea to do the spreadsheet and the NRP distribution in parallel though, as it is only the second run for Sensorica on distributing income, and this would be general best practice for a major new function like the value equation.

I put in a couple test equations, as I wasn't initially clear what percentages you wanted to use.  Once your data is corrected, I'd like to experiment and see if we have differences, and if so, what is the cause.



On Fri, Sep 18, 2015 at 5:24 PM, Tiberius Brastaviceanu <tiberius.br...@gmail.com> wrote:
As Jim pointed out, Daniel has programming hours?!

This is a logging mistake by Daniel. you can catch that in the Milestone 2 sheet I will correct it in the NRP
9/15/2015DanielPV characterizationProgramming for product
9/15/2015DanielPV characterizationProgramming for product

I will make the changes now. 

On Fri, Sep 18, 2015 at 5:49 PM, Tiberius Brastaviceanu <tiberius.br...@gmail.com> wrote:
Hi all, 

​​
I did a calculation for the distribution for Milestone 2. I did it in the spreadsheet, since the value equation engine cannot support the way data was logged, for this budgeted project, a forward-looking value equation. We haven't reached the level of abstraction that can support all the different use cases with the Value Equation. More conceptual work needs to be done... 

So, here are the results. 

Daniel$1,103.33
Ahmed$865.39
Jim$1,295.56
Tibi$783.23
John$99.67
Thomas$420.54
Abran$57.27
Maria$138.29
Adam$72.62
Tammy$164.09
$5,000.00

See my calculations in the spreadsheet. 

Please check to see if everything is OK. There are places where I do proofs... If you can't follow the calculation I can help you understand. 

Thank you all for your collaboration. 

NOTE: We need to deliver as fast as we can. There are another 10K that we can get from this project! 
Don't forget to continue to do outreach to get more participation. 

--
Go to SENSORICA home
https://sites.google.com/site/sensoricahome/home
Go to our Working Space
https://sites.google.com/site/sensoricahome/home/working-space
--
You received this message because you are subscribed to SENSORICA group.
To post to this group, send email to Sens...@googlegroups.com
To unsubscribe from this group, send email to
Sensorica+...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/Sensorica?hl=en?hl=en
---
You received this message because you are subscribed to the Google Groups "SENSORICA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to Sensorica+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
t!b! 
co-founder of SENSORICA, 
an open, decentralized and self-organizing
value network (an open enterprise)


--
Go to SENSORICA's project page
http://www.sensorica.co/home/what-we-do/projects/pv-characterization
---
You received this message because you are subscribed to the Google Groups "PV characterization project forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pv-characterization-pr...@googlegroups.com.
To post to this group, send email to pv-characterizat...@googlegroups.com.
Visit this group at http://groups.google.com/group/pv-characterization-project-forum.
To view this discussion on the web visit https://groups.google.com/d/msgid/pv-characterization-project-forum/CAH3XNQ%2BjOgDe9rSMwQoWNC%2BhX6sA4PrN3Typ2PeNtYz4qdCs1Q%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.




--
t!b! 
co-founder of SENSORICA, 
an open, decentralized and self-organizing
value network (an open enterprise)





--
t!b! 
co-founder of SENSORICA, 
an open, decentralized and self-organizing
value network (an open enterprise)






--
t!b! 
co-founder of SENSORICA, 
an open, decentralized and self-organizing
value network (an open enterprise)


--
Go to SENSORICA home
https://sites.google.com/site/sensoricahome/home
Go to our Working Space
https://sites.google.com/site/sensoricahome/home/working-space
--
You received this message because you are subscribed to SENSORICA group.
To post to this group, send email to Sens...@googlegroups.com
To unsubscribe from this group, send email to
Sensorica+...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/Sensorica?hl=en?hl=en
---
You received this message because you are subscribed to the Google Groups "SENSORICA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to Sensorica+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Reply all
Reply to author
Forward
0 new messages