How to make dresses (or databases)
March 6, 2024
From schmatas to schematas in two generations
It's a Sunday morning in the 1960s. After teaching school in California for a year, David comes home to Pennsylvania.
Charlie, his father, wants to put David to work in the factory making dresses. Charlie's in the rag trade, the schmata business. His associate Bill is over for breakfast.
"I don't know how to make a dress," says David.
Bill says to Charlie, "He went to college, he seems like he's smart."
"I don't know," says Charlie.
"You want to tell him?" asks Bill.
"No, you can tell him," says Charlie.
"Get a pencil and paper," says Bill.
David gets a yellow pad and a pencil.
"Are you ready? Sit down."
Bill rattles off the instructions.
To make a dress: You make the top. You make the bottom. You sew it in the middle.
"Is there any part of that you don't understand?"
"Well, there's another option," adds Charlie.
"Oh, right!" says Bill.
To make a dress: You make the front. You make the back. You sew it together on the sides.
"You think you can do that?" Bill asks.
David shrugs and nods.
"Now you can make any dress."
And that's how my Uncle David started making dresses. He worked with Grandpa Charlie, took over the family business, and expanded into apparel manufacturing for the biggest department stores in the country.
---
Now here I am making databases, using almost the same instructions.
To make a database app: You make the back. You make the front. You sew it together in the middle.
The difference is I start from the back.
I'll build out a few tables with sample data and then have Power Apps automatically generate a model-driven app for testing. And then I'll add another table and repeat, pushing the data model as far as it makes sense.
Once that's done, only then do I start building pixel-perfect, front-end canvas apps that define what the laptop, tablet, and mobile screens look like. And then sew it all together with business process flows, rules, and automations.
There's also a case for starting from the front. That's how it's done by UI designers using Figma and the like. First comes the napkin sketch, and then the database team does the implementation. At least that was how it was done before low-code tools made it so easy to iterate with back-end data structure.
---
And now, a word of caution: Knowing how to make a dress doesn't make you a professional dressmaker.
With low-code tools, you can build a full-stack application in days. But other than apps for your own use, be extremely careful about which applications you deploy into production, and for whom you deploy them.
In my previous career as a professional writer working with IT consultants, I wrote case studies, thought pieces, and editorials bundled into highly polished newsletters. We'd go through dozens of iterations before sending a PDF to the printer. But once it was printed, that was the end of it until the next issue.
With database apps for a client, go-live is just the beginning.
After go-live, you can't predict what the next bottleneck will be, or what problems will appear, or what opportunities will arise, or what skills you'll need in response.
Unless you know everything about everything, you're going to need help.
And if you do know everything about everything, you should be running a team.
That's why my plan for the month is to finish up and publish my personal portfolio apps (Certificate Skills Explorer, the Back-of-Envelope calculator for Power Platform pricing, and my final project for Microsoft Power Up).
Using those samples, I'll get work at a factory.
I think I can handle it. I went to college.