Tuesday, October 18, 2005

Why read?

Out of the mouth of ...

Well he's not a "babe", but his diction is lacking and he can sound like that sometimes. Anyway, back at the blog:

In order to read a "word" you will need to be able to take four things, group them (reasonably accurately) together into one thing, and remember that one thing.

In order to write the word you will need to create one thing from four things, in the correct case and order. And you need to have remembered the four things.

Of course reading is easier.

For him in school, maybe. But not for me at work.

For source code it is true that writing is easier. Probably because we get more practice at it, but also becuase code is not intended to be read the way we read. Not without a code-browser that is good at linking the bits together, because the code is not sequential. There is a story, but not a narrative.

We read on the stack - read some lines, find something interesting in it, push context, read the interesting stuff (probably with more stack pushes), pop context and resume some lines of code. Except that we are not, well, I'm not a good stack, and after 7 pushes and 3 pops I tend to start from the top again.

It is easier to write on the stack - becuase if your pops don't exactly match your pushes, some fucking tool is gonna stand up and whack you in the head just to remind you. It may only look like an error message to you but it can be just as sore as a whack in the head to a coder who is flowing.

And that feedback has trained us into holding as little syntax-checking as we can get away with in our heads. We have machines to hold the rest of it for us, and would much rather fly along spewing semantics.

0 Comments:

Post a Comment

<< Home