4 ms·
I work for an organisation who's business model depends on users completing a very long and complex legal form. The form is paper-based, because it requires a
by Spearchucker 13y ago
I work for an organisation who's business model depends on users completing a very long and complex legal form.
The form is paper-based, because it requires a number of signatures to make it legally binding. They brought in a consultancy to move the whole thing online. They've been at it for a year and a half. I started wondering what it was that's so difficult.
The consultancy created a form in which fields, guidance and rules were hard-coded. Part of what took them 18+ months is that the client kept making changes to form guidance.
I got myself a printed version of the form and used it to create a form-builder, which simply generates HTML. A non-technical business admin can change field sequences, guidance, and business rules using a simple UI, and publish those in real time. That took just 3 days.
When I demonstrated my proof of concept I got those 10x comments. Guess what. I'm a festering pile of shit as a programmer. But I've been doing this for 20 years, and have solved this little problem many times in those 20 years. I didn't write much code. I just copied and pasted stuff I had lying around, renamed some entities and methods, and put it onto a server.
It really is just experience.
- mooreds 13y agoI think that having the context to recognize a problem and solve it at a meta level means, almost by definition, that you aren't a "festering pile of shit as a programmer". Programming isn't just about writing great code, it's about solving problems. Sure, one problem to solve is having code be extensible/maintainable, but it isn't the only one. I don't care what shiny new code is written with the latest framework; if it doesn't solve a business problem, it's bad code.