Friday, September 12, 2008

To Chunk or Not to Chunk

As the poor, poor readers of this blog know, I'm working on a masters degree in technical writing. The subject of the week: When do we know when a web text needs to be chunked or not. My initial thoughts:

I came up with a few instances when you might not want to chunk. Some are better than others:

1) The text naturally does not chunk.
2) The text is a narrative, essay, article or other block that readers will read/need to read to understand/enjoy to the fullest.

1) The text is used by audiences with vast differences in experience and knowledge, with some chunks directing the experienced user to different areas.
2) The text is used to explain how to do something, or describes something that is done in steps, or, in "real life" is done in steps. I'm thinking here of DIY instructions, or chunking text to guide a person through the process of buying something online.

These are pretty shaky at the moment. I'll keep thinking.

Here's another one that just struck me: Chunking works for texts we're used to seeing chunked. It would be pretty weird, for example, to read a Bible chapter not broken into verses.

Then, my thoughts start to break down. The question is much more complicated than I'd thought:

Just realized as I was sitting here, contemplating my bellybutton, that one of my favorite books on the shelf above my head is chunked and breaks one of the rules I set for non-chunking. So we can chuck the "narrative" rule out the window.

The book is "The Final Days," which focuses on the Nixon White House during the Watergate scandal. It is very nicely chunked, first with a "cast of characters," a lengthy list of people mentioned in the book. Very helpful, because the story is complicated.

Then the text itself is chunked, as the authors switch from one scene to the next, one viewpoint to the next, rather like cutting scenes in a movie. So I need to modify the narrative no-chunk rule:

If a narrative is complex, chunking is natural in separating events, identifying characters, settings and such.

How would this apply to the web -- well, if this book were presented as a web page, I'd definitely want it chunked, perhaps with cross-referenced links between the individual threads that the authors weave together. That might be one of those new web texts that Prof. Hailey mentioned -- a book where one can read one thread in its entirety, without interruption, but then get back into the linear flow of events.

So in dealing with multiple narrative threads, a cast of thousands, chunking would be beneficial.

And more exceptions:

I think whether chunking would or would not work in this situation depends on the purpose of the writing. I can see in some cases -- such as a report we'll have coming out next week at work detailing our performance during a management self-assessment on a new waste process -- where chunking is going to be valuable in helping us identify our strengths and weaknesses in a hurry. It's going to be a complex matter, needing a lot of background and analysis, and it will lead itself naturally to chunking because of that need for analysis.

It's a tough call. It's like whenever people ask me how to write something -- I don't have the vocabulary to tell them, aside from the boring grammar mechanics that makes their eyes glaze. Part of me thinks we can come up with a comprehensive philosophy on chunking versus non-chunking, part of me wonders if we'll be stuck with I'll know the chunking situation when I see it. But that's a cop out, because we have to have some kind of ground rule in order to see it. Argh.

So it's a tough question. Looking to the Internet for help is, of course, problematic, because you either run into the “chunk it or else camp” or the “chunk it because you're supposed to chunk it crowd” or the “chunking?” crowd. Must admit I was in the third category for a long time. Now I'm in that netherworld where I'm at the back of the temple where the idol of chunking is worshipped, wondering if there's some benefit here or if I should go back out the door. I can see the benefits, but I know it's not a blanket approach. Must keep thinking.

No comments: