The other day, I was asked "what I program" by an interested friend. I had to use the metaphor of how a word processor is a kind of editor that automates the parts to writing English essays and how I'm creating an editor that helps me program in my favorite language by streamlining the parts I like to be in my programming habits.
The moment I did that, my brain went wild. I immediately jumped into law automation, my favorite lost realm. It looks like others wish for that, too. Metaphorzing what I do is a slippy slope in my head. I was infected with it for a while.
First I'd like to share some reminders that helped me get out of it. Then I'll share the question that stumped and silenced it and helped focus and re-shape it for the better. It's a good reminder that the current startup world is a little bit "flip"-oriented, here. And that other alternatives to the prime gate to that, VCs, exists... here. Finally, it's not all fun and games, but lots of hardship, but it also depends on how you frame it, here.
The question that saved me from my mind freeze in la-la land was: how could my "purported" ideas for automating law apply in exact parallel ways to software? There. Think about that for a while. Hard, isn't it, to swallow one's own bitter pill.
...
I think I gotta acknowledge how case law tells you exactly how everything should be, down to exactly the range of what coffee temperature should be when served. And it's all supplied in a highly stylized manner. I think it is important to acknowledge the irresistability of all that... and just deal with it after I direct the cannon at myself first.
How can I parse new projects? Can I use documentation and use instructions to re-purpose a project? Can I recompile an algorithm in a different language? Voice recognition aside, can I use natural language as a limited domain interface? Can I have smart GUI tools that go beyond syntax highlighting, code references, completions, reminders of method definitions, error detection and debugging? What would that be like? Can I have enough tooling to make many projects fast, or a large project manageable by one person? Can I embed a kind of engine that could auto-correct "obvious" errors? Can I direct the engine towards an unknown project and have it manage it, as well; maybe even change it as needed?
That kind of stuff... first. I must also resist all social implications and focus on making them relevant in the first place. Goodness, so many things to be distracted with. But it's ok. I think I can do it.