Wow, I’m up to the step where I show what the diff output looks like. OK, that predisposes that I turned the current and previous babystep code into text files. So, it’s time to fire up the FileSystemObject and the TextStream object again. I made heavy use of them in the first half of the project, but mostly for reading. This time, I’ll be opening files for writing, and then they will be immediately read back in. After read in, there will be an object in memory that represents the new content for between the babystep tags in the current post. And as we’ve done recently, we will use the RegEx object to replace the babystep tag pattern match with the new content. And the resulting content gets updated back into the table, and voila! The application will be done.
Right now, both halves of the project have lots of output. It’s all really just debugging output. When I combine the two halves of the project, the output will actually be made invisible. Instead, it will reload the same discussion forum thread you are currently looking at, but will force a refresh. The process will not be automatic at first, so that I can retroactively apply it to discussions that already exist. Think clicking a “babystep” link over and over until the program is fine-tuned. It will be safe to re-run, so if there’s still adjustments to be made, no real damage is done. I think I’ll make a backup of the table beforehand just to be on the safe side. And this same system can be used to add a program code colorizer and beautifier if that ever becomes a priority.
This will be absolutely fascinating to watch how it will affect my work. It is central to the way I work, and plan to maintain my focus and stay engaged. The best way to learn is to teach, and this is the best way for me to teach. It forces me to be systematic, and allows me to review process. It creates a play-by-play archive of a project, recording my thoughts at different stages. It will help other people when they work on similar projects, and will help me by allowing review and feedback by my peers. I’m sure professional programmers will cringe at most of my VBScripting work, and liberal use of global variables to avoid parameter passing. But these initial projects are not about purity. Neither are they about long-term manageability. They are about breathing life into a site quickly and starting to build an excitement level, that will justify me switching to my ideal platform, at which time I will be going through some very interesting learning processes, and documenting it all here in the babystep style.
Process is an important characteristic of a project that rarely gets proper play. Programmers don’t like to reveal their follies, and the book publishing model taught us to be efficient with our examples. Rarely would you re-print an entire program example to show how just a few lines of code changed from one page to the next. But that’s exactly where the value lies. I can’t count the number of times I looked for code examples on the Web, and had difficulty viewing the code out of context. Seeing it built up from scratch, especially when you go in steps of just a few lines at a time, can make programmers out of even the slowest learners. There is a reason for every line you put into a program, and those reasons get lost because the process flow gets lost. After awhile, it just becomes the finished product and you loose the sense for how you got there.
Wow, this post about though-process on the babystep tutorial system was going to go in the internal system, but it provides such insight to the HitTail site, and the type of content that’s going to be found here that I think I’ll add it to the HitTail blog. I am also thinking about actually putting out the tutorial of the birth of the babystep tutorial system. I like the way that it is so self-referential. I will use the babystep documentation system to show the evolution of the babystep documentation system. It’s all very circular. Some of the best systems are circular, and self-referential in this way.