Should we split target_market_share() into target_production() and target_technology_share()?

[maurolepore]

market_share() outputs two different results: technology_share and produciton.
Should we have a dedicated function for each kind of output?

This is not an urgent matter but an invitation to reflect on the design of the
package.

[jdhoffa]

This is an interesting discussion that I would like to bring back to the fore. I'm curious what @cjyetman's thoughts are here too.

In general, I think I would favour this. I can imagine that "1 metric, 1 function" may be a useful design pattern. Splitting these apart may allow us to more easily document each "metric", and what/ why it is useful, providing a unique vignette for each.

However, that would depend on us adequately defining "what a metric is" which as it stands is a totally ambiguous definition.

In the case above, the values of technology_share and target_technology_share are dependent entirely on the values of production and target_production. In that sense, are they different views of the same metric? Are they different metrics? It's pretty ambiguous.

[cjyetman]

I'm in favor of "1 metric, 1 function"

I believe technology_share and target_technology_share are 1 metric applied to two different production trajectories, one expected production and one target production

edit: Since target_technology_share also likely requires the first year expected production, then yeah I understand how this is possibly a separate metric

[jdhoffa]

The question more precisely is:
are the production (+ trajectories) a separate or unique metric from the technology_share/technology_mix (+ trajectories)

[cjyetman]

The question more precisely is: are the production (+ trajectories) a separate or unique metric from the technology_share/technology_mix (+ trajectories)

I would say yes, technology_share/technology_mix needs more information right, so they can't possibly be the same formula? Like if a piece of pie is X radians wide and it's Y percent of the pie... I'd say X and Y are different metrics.

[jdhoffa]

I agree! Just wanted a second opinion/ sanity check.

Then all that said, I would think this issue represents a reasonable and desirable feature.

[jdhoffa]

Status update:
Since this is a breaking change, and not considered a priority ATM so will likely not be worked on for the upcoming 2024Q1 release.

[jdhoffa]

George and I believe that this may be a valuable architecture for a future iteration of these packages, but not something to work on in the existing r2dii suite.

Tagging as "wontfix"