It must sound like an odd notion that it would be so much faster to program in one language over another. Clearly, an algorithm that adds two and two will be easy to write in any language at all. In any high level language, 2+2 is managed in a single line of code and for this purpose all high level languages are equivalent.
Yet today’s programs are big. They do a lot and must work seamlessly with third-party black boxes; there’s a lot more to think about than will fit in the human brain at one time. We need a way to understand the algorithm at the right level of detail. I am talking about a form of algorithmic mip-mapping, a way to see just the algorithm in a clear obvious context. In other words, we have to be able to quickly find our way to the right place in the program, and once there, we have to be able to understand what we see.
Source is typically distributed across a number of files, each of which is a list of lines. With ASL, we slice those big lists into small swatches. The swatches relate to each other, and those relationships form a tree. It is much faster to traverse a tree than an ordered list. The actual factor is dependent on the size of your source code, but in a commercial application of average size, your navigation speed should improve by at least an order of magnitude.
The second part of speed is being able to quickly understand what we see. ASL is finer grained than the source file level. A typical ASL swatch fits easily on a screen. It is easy to understand 10 or 15 lines of code, especially if it is named well and formatted properly. This example shows how big complex code turns into small simple swatches.
