|
I have to think about this. Back in the days when rapid prototyping meant checking things out with (SHUDDER) Basic, or Tcl, or later, Hypercard, I actually learned something along the way that came in very handy later in a different context. So, a dumb problem came up --- Really dumb --- how to make the world's simplest online calendar so that someone could just enter event dates (there are less than 10 a month) and not wait until the very last one came in, for an email to go out. Key dates are known and it's the 13th and no January calendar yet. So, it's like convert date formats, which Numbers refused to do, and then pair that column with days of the week. Before I went back to the days of "Unix Text Processing" and the insanely useful "paste" command. That came from the days when I studied that book and put a ton of it into practice. It was a guide top shell and built-ins via example. Doing this, I thought that perl would solve it all, and despite not having used it in countless years (I retired long ago) it started coming back to me. But the shell stuff worked better than a spreadsheet (software from hell) and perl, which can have some incredibly poetic solutions. I promised myself to re-read what I found. Because it's "neat". I learned a lot before even usenet and Google, and that has stuck for the most part, because "I had nobody to lean on". |
Now, for the most part, AI makes great progress. I only need to step in when a human, albeit with limited knowledge, needs to step in