Datasets for website use cases and inclusion in Mathesar

[Zack Krida]

Hi team, particularly Brent and Sean,

My pull request to add the new sample datasets to the playground repository is available here:

https://github.com/mathesar-foundation/mathesar-data-playground/pull/20

In this message, I wanted to give some context for these and thoughts on their inclusion in Mathesar that I don't think is suitable for the public repository:

1. These use cases were derived from a Go To Market document Jeff produced. That document gives the justification for each use case's inclusion in the website, from a marketing perspective.
2. I created schemas and sample datasets to the best of my ability in an effort to represent Jeff's vision but also create realistic "proof of concept" schemas that would resemble real-world usage. Each use case has some website marketing copy that can also be reviewed. It has not been reviewed by anyone yet.
3. The idea to include these in Mathesar itself came later. The goal there, as I understand it, is to approximate functionality like Airtable templates that gives users a relevant entrypoint into our application. While we do not yet have any means for a "View in Mathesar"-style link that would allow someone to inject a DB and sample data into their Mathesar installation, including these datasets in Mathesar itself was a way of getting some semblance of functionality like that, and some continuity for marketing site visitors evaluating the product.
   
4. This is all to emphasize that these datasets were not written with inclusion in Mathesar in mind. I have not personally thought about:
   1. The criteria by which a dataset is suitable for inclusion in Mathesar
   2. The "correct" number of rows / size of a dataset for inclusion in Mathesar
   3. The proper "fidelity" for a dataset in Mathesar (lorem-ipsum style text vs. "real" data vs. plausible-but-fake data)
      
5. I imagine the above points could be contentious and up for debate by the team.
   

I say all of this in an effort to clarify that I'm not strongly opinionated on any of the work here, and that any feedback on these datasets should be focused on their suitability for Mathesar at a product-level.

Thanks! I'm very curious to see what everyone thinks.

[Sean Colsen]

I’ve been skimming over this data inside Mathesar, but I’m running out of time due to higher priorities. There’s a lot to look at here!

I can see the value of adding a way to load this data from inside Mathesar, but it does feel a bit rushed at this point. My inclination would be to hold off on including it until we have more time to give to vet it.

As an example, loading the table page for “Museum Exhibits” > “Item_Collections” gives a JS error, making the table page unusable. That’s a bug I’ll file (when I have more time). I imagine Brent’s fixes would address it. I’m not sure how much more time I’ll have to review this before we cut the release branch. And I’m very curious to poke at it more. At the very least, I think it makes a useful basis for performing QA!

​

[Kriti Godey]

Thanks for the review, Sean. The plan is for Brent to take over this work, he'll be working on it on Monday. No need for further review at this time.