Improve Signal And Readability

This page is dedicated to AlistairCockburn, a high quality contributor and refactorer of Wiki in its "formative years", and to PeterMerel.

Alistair said by email to a few of us in the first quarter of 2000:

There have been a few pages where the thread was refactored and signatures left in place. The pages that have become very long and were refactored anonymously have such a high content density that I can't even read them at all.

Peter replied:

This is interesting. High content density isn't something I've been thinking of as a drawback. I guess the question is whether a page is easier to read when it's a concise definition, or when it's a PassionateDrama?.

This exchange really struck me and led me to suggest that the twin goals of Wiki refactoring should be to improve signal and readability.

Alistair also said, on looking at RefactorFasterDeleteMore:

Not a problem refactoring live threads. The participants are visiting daily, and can respond to the changes.

That doesn't mean he's right. But I do respect his opinion, based on his experience of refactoring some pretty massive and tricky multi-author pages, a whole lot.


Where are the requirements for a WikiPage?

High signal and readability?

That's not sufficient. You don't have enough requirements to create a WikiPage. If those were the only two requirements, Wiki would be barren.

I'm not sure I followed that. ExplainOrRefactor?

You still need an idea to create a page in the first place. Refactoring messy pages is a good thing; just be careful not to destroy signal in order to achieve high signal. You could achieve high signal by deleting everything on a page except for one solid point but that would be very, very bad because you'd remove all other good points.

For me that might increase signal density but would not improve signal. I agree that it would be a bad thing.

A more interesting issue is the continuing mix-up between "refactor" and "rewrite." Refactoring doesn't change the semantics of the target, it just reorganizes it. Rewriting is a more "violent" act. Proper refactoring only deletes content-empty expressions, in either a programming language or a natural language.

I increasingly doubt the possibility of refactoring without changing semantics at all. Human language is much more subtle than this. But for me the risk is worth taking and Wiki is much the richer for ongoing refactoring to improve signal and readability.

Plus, there is an unfortunate trend in Wiki toward content-free sentences, paragraphs, sections, and whole pages. Unfortunately, RefactorFasterDeleteMore doesn't work because people form an attachment to their 'contributions', irrespective of how little (or how much) value they have added. What other solutions are there ?

Maybe islands of compatible authors who feel trusting of each other and who will allow the killing of darlings (see KillYourDarlings). Not to mandate this, but simply to recognize that it has emerged, and perhaps to leverage off that a little by suggesting it gently. I liked working with TomAyerst on one of the Quality pages, and appreciate RichardDrake's efforts to coagulate authorship on ChristianIntellectual.


See also CommunityLifeCycle. We need more patterns!

View edit of July 10, 2010 or FindPage with title or text search