On the note of missing mental models, and somewhat related to "Individual Tasks" - The model that all tasks can be executed on equally effectively by every individual on the team and implying that any quantifiable capacity is therefore fully fungible among team members a.k.a "Not understanding team and task nuances and context".
Mental models can be brought to life with data, especially if it is done in real-time.
One of the challenges with the software industry is that we have a pretty stunted view of what is possible with data, and tend to only see the negative aspects of measurement.
For good reason, I might add. We've equated measurement with metrics and completely forget that in every other field people look to data as a reasoning tool.
It's really ironic that we find this such an alien concept in our own field :)
That's a great call. I'd say "garden" is better than farm. The main issue with the mfg thing is that you are likely 1) building the factory, 2) the factory "is" the product. Which is a bit different than replicating things, etc.
I think it is less about finding an analogy to work with but starting with a neutral set of first principles and mathematical models and applying them carefully to software development as it is.
Software development is different from all these other areas in fundamental ways, but it does not mean the underlying "physics" are fundamentally different.
The notion of stability for example, is a very general, and very powerful one, that has very different applications in software compared to industrial production.
This is great amount of content for my week this week, I've read every single one!
No worries John, keep'em coming!
On the note of missing mental models, and somewhat related to "Individual Tasks" - The model that all tasks can be executed on equally effectively by every individual on the team and implying that any quantifiable capacity is therefore fully fungible among team members a.k.a "Not understanding team and task nuances and context".
Mental models can be brought to life with data, especially if it is done in real-time.
One of the challenges with the software industry is that we have a pretty stunted view of what is possible with data, and tend to only see the negative aspects of measurement.
For good reason, I might add. We've equated measurement with metrics and completely forget that in every other field people look to data as a reasoning tool.
It's really ironic that we find this such an alien concept in our own field :)
That's a great call. I'd say "garden" is better than farm. The main issue with the mfg thing is that you are likely 1) building the factory, 2) the factory "is" the product. Which is a bit different than replicating things, etc.
I've been exploring this subject in more detail over on my substack:
https://open.substack.com/pub/thepolarisflowdispatch/p/flow-metrics-are-not-manufacturing
Would love to hear your thoughts..
I think it is less about finding an analogy to work with but starting with a neutral set of first principles and mathematical models and applying them carefully to software development as it is.
Software development is different from all these other areas in fundamental ways, but it does not mean the underlying "physics" are fundamentally different.
The notion of stability for example, is a very general, and very powerful one, that has very different applications in software compared to industrial production.
https://www.polaris-flow-dispatch.com/p/stability-first