Compiler Model

Here’s another post in my general theme of “using a technology/comp-sci metaphor for creativity, new media and productivity” … In the generous sense.

For the non technical: compilers are computer programs that take raw computer code and turn it into something binary that a computer can run. This makes it much easier to write programs, because compilers let programers abstract certain functionality¹, and because compilers can do some consolidation and tuning as they run.

While traditional compilers are used for the heavy duty code situations and applications written in C (and C++ and Objective-C, and so forth), when you think about it there are lots of situations where creators, in the digital world, are making something that is then processed by a computer script. In a way, web pages are like that, (though that happens every time their accessed rather than once.)

The wiki program that I use, ikiwiki, works in this way by taking a directory of files and turning into web pages that I can read in the browser. This is basically how the wonderful LaTeX typesetting engine works. Also, interestingly, I suspect this is how most pro-level video/audio editing apps like FinalCut work (make a lot of edits and then render something out. Edits are made as a series of instructions to XML file, which the program then renders or compiles.)

So what’s the benefit to this? Here’s what I’ve come up with:

• It makes it easier to write software that works with your data. Editing text is easy, particularly when that text is created either by you or by a program in a regular and expected way.
• Compiler modeled systems are less taxing generally on the system. Most of the time, what you’re doing with the computer is just editing a text file. That’s low power. Then when it comes time to compile it just has to crunch through some data. And then it’s done. In WYSIWYG and live editing, the computer has to be working constantly to get from you to what you see.
• Any repetitive task can be automated or template, and the chances are that there are a very limited repertoire of kinds of documents that you’re making. I write academic papers in APA format, (or did), formal letters, and full page knitting patterns. That’s about it. I have templates on hand that mean that I can push the same block of text into a template and get a perfectly formated document. I can push the same block of text between the wiki and the blog and it’ll compile (basically) the same way.
• It makes it more feasible to work in smaller files. I wrote a book in Word, and while I started writing the chapters in separate files, it quickly became apparent (and I think this holds true). If you’re using a program like Word, the best bet is to make really long files, and keep projects together in one file. On the other hand, if you’re working in plain text² it’s trivial to mash-up a bunch of text files, and this makes it easier to edit and organize a project.

So more than anything working in this way makes a lot of sense. Though I think to be fair there is a learning curve (but isn’t there always?). The question I’m asking myself (and you, dear reader) is other than the general improved efficiency and cleaner workflow, is there a bigger extrapolation or application of this that we (or I) could apply to another phase of the creative/productive process? Rather than just automating the presentation, is could someone take the model where some sort of raw file is “put together” later and have that assumption shape the way that they create content?

I’m not sure what that would look like, or even if it’s a good idea, but its something to think about.