SpellCheck 2.0: Bringing the benefits of Web 2.0 back to the desktop

by Greg Swann

Suburban Phoenix Real Estate Broker

Author website

I wrote this when BloodhoundBlog was very young, and I don’t think anyone got the joke at the time. There have been plenty of amazingly stoopid Web 2.0 product launches since then, but that doesn’t mean this is not still humor-for-one.

I was writing today, and I realized that spell checking, for all its added efficiencies, isn’t terribly smarter than it was on the dedicated text-processing systems of the 1980s. It made the jump to desktop machines, of course, a trusty sidekick of word-processing, the first true “killer app” of micro-computing. But both were quickly eclipsed by spreadsheet software, and text management tools have been a red-headed step-child on desktop systems ever since. Everyone needs them, and everyone hates them when they don’t work properly, but no one lays awake at night wondering what new computing paradigms might be expressed in future versions of their favorite word processor.

And spell checking has had it even worse. It’s the sidekick to the step-child, after all. If it had a more tangible form, it might be stuffed into a junk drawer, handy to have around but usually just in the way. Spell checking has missed virtually all of the internet revolution, of course. Many web development tools incorporate spell checking, as do some on-line web sites. But there was no formal presence for spell checking in the Web 1.0 paradigm. No spelling look-up servers. No advertiser-supported spelling portals. No spelling IPOs. In fact, not one single wide-eyed investor pissed away his retirement savings on a Web 1.0 spelling start-up.

Worse yet, it seems almost certain that spell checking will be passed by in the forth-coming Web 2.0 revolution. This would be unfortunate, since spell checking is in fact the perfect Web 2.0 application — er, platform. Note these criteria from Tim O’Reilly’s seminal paper on the characteristics of a Web 2.0 platform:

• The Long Tail
• Data is the Next Intel Inside
• Users Add Value
• Network Effects by Default
• Some Rights Reserved
• The Perpetual Beta
• Cooperate, Don’t Control
• Software Above the Level of a Single Device

If we envision a product — er, application — er, platform called SpellCheck 2.0, we can incorporate all that stuff and then some.

It may occur to you to ask, “Why?” That’s really a pre-Web 1.0 question. Web 2.0 is not about getting things done, it’s about getting people together. It’s not about what we as a club-composed-of-people-who-have-never-met have accomplished, it’s about how we feel about it. A Web 2.0 platform is a phenomenon first. It is only secondarily phenomenological.

So to start with, SpellCheck 2.0 should be heuristic — if for no other reason than that you can’t even write about heuristics without a really good spell checker. But even at the isolated, atomistic, disconnected, non Web 2.0 level, SpellCheck 2.0 should do more than just run a data-base look-up of potentially-acceptable substitutes for suspected spelling errors. It should learn from your writing which words you habitually misspell, which typos you persistently make. By iterative analysis of the error/suggestion/selection/replacement process, SpellCheck 2.0 should learn, over time, how to suggest the most likely corrections for you as a unique user.

Now let’s scale that up to the internet. If, like every good Web 2.0 platform, SpellCheck 2.0 is in constant contact with a centralized file server, it can heuristically analyze and learn from the errors of millions of users. The platform becomes inherently and intrinsically social: Users Add Value; Network Effects by Default; Data is the Next Intel Inside. We even find room for The Long Tail: The least frequently used words are apt to be the hardest to spell correctly. An heuristic Web 2.0 spell checker can turn the dross of collective ignorance into the gold of collective intelligence.

Obviously, this is Software Above the Level of a Single Device. It requires the interaction of millions of distinct micro-computers.

And, of course, this is a Cooperate, Don’t Control epistemological model: If a misspelled word goes persistently uncorrected by the majority of users, it can no longer be considered to be misspelled. Vox populi, vox dei. The ensuing work product might be considered to be Some Rights Reserved, but any rights to anything would be difficult to enforce.

And obviously SpellCheck 2.0 is a Perpetual Beta, since the product is essentially always in a state of self-generation — self-improvement if you will.

Judging by these standards, SpellCheck 2.0 may be the perfect Web 2.0 platform. After all, it exhibits every one of O’Reilly’s eight criteria. But we’re not done yet. We haven’t even considered the true Web 2.0 benefits of SpellCheck 2.0.

Consider a user interface that permits you to rate particular spelling errors. Or particularly bad spellers. Imagine a screen that lets you see the most-misspelled, least-misspelled or most-uncommon misspellings. Picture a community of nationally-known spelling webloggers, haughty and self-satisfied, justifiably so, all linked together though SpellCheck 2.0. Stretch your mind and envision a mash-up of SpellCheck 2.0 with Google Maps that permits you to identify geographical misspelling hot zones. The social networking possibilities of SpellCheck 2.0 are limitless.

All that, plus it will even check your spelling!

I’d love to hear from interested venture capitalists, preferably with their pockets bulging with cash, but I wouldn’t absolutely hate it is some Seattle billionaire just bought me out, lock, stock and barrel. I’ve heard that it’s possible to be so rich that you don’t have to give a damn that your text is misspelled…