Thursday, April 12, 2007

My Cloud Library

(or No, it will not be meta-ironic if I call it Spellr”)
The inline spell-checker in FireFox 2 is one of its most useful features. Particularly now that I am writing much more text within the browser. The days when I really only typed little more than passwords and occasional email addresses into browsers are well past. Today I am typing long blog posts, short comments, and all sorts of data into web applications.
I habitually work on more than one computer. A practice which has revealed an (perhaps unforeseen) opportunity within this dictionary/spelling space.
The pre-loaded core dictionary that the spell-checker begins with is of limited utility without all of the custom words: names, surnames, brand names, and industry-specific terms that I am constantly adding. Extrapolating from the individual to the collective, each of us has our own unique, arguably valuable, collection of words
FireFox caches my unique wordset onto my hard drive. When I write a first draft blog post on my work Mac, I correct all of my mis-spellings and add all of the unrecognised, yet correctly-spelt, words into my custom dictionary. Later, when I complete that post on our home laptop, the screen is once again littered with red underlines as that second instance of FireFox is checking my text against its own local custom dictionary. I then have to go though those words again, adding each of them to this custom dictionary. (Once you have published a few blog posts with screaming typos in them, you get pretty careful about copy-checking before hitting the Publish button.)
Therefore, why not have my custom wordset out in the cloud, and not stored locally at all? Think of doing for spell-checking what Del.icio.us does for bookmarking. Actually best not to think in terms of spell-checking at all. That is really only the task supported by, and enhanced by, your personal wordset. It is more helpful to think a lot broader than the specific example I gave above. Think of your custom wordset as data that you continually add more value to every day. There have to be great benefits to having that dataset be accessible across multiple platforms and also to-be portable. For example, say a FireFox-killer arrives in a few years time, do I want to have to start the whole process of generating a complete new custom dictionary again if I migrate to a new browser? Or if I change job and am issued with a new laptop? Do I want to have to teach my friend’s names to every new mobile phone that I buy?
I did a quick search and the existing model for online services in this space are basic spell-checking sites: paste in a block of text and have it checked. A closer model to what I envisage is Google’s inline spell-checker (which seems to override FireFox’s, although I am unsure how that pecking order works). What I cannot divine is whether theirs is fully integrated or not. I want to teach GoogleSpell a word by adding it while writing a Gmail today and have it recognise the same word in Google Docs tomorrow*. If I include someone’s nickname in their contact details in Gmail, then I want Gspell to not flag that word as misspelled in my Gmail. Making this happen within a suite of products yoked together with a common user profile and log-in has to be more manageable than aiming for the Internet at large. Building this within a related suite of applications could provide short term lock-in.
Back to creating this as a stand-alone web service then, technically speaking there would be some a lot of non-trivial issues to overcome to achieve this. Would you have to log-in to your browser to activate your wordset? Or, if the service is disassociated from the browser, would it need an open tab at all times? Interoperability with all my devices and platforms would be ideal: I imagine writing something in my smartphone on the morning commute; giving it a polish in Google Docs over lunch, and then publishing via Blogger that evening. Would it be better for my smartphone to simply cache the most current wordset when I sync it, or to pull it live directly out of the cloud?
Another issue is that a service like this adds another component to your online data shadow. If someone (say a potential employer) could gain access to your personal wordset and run statistical analysis or personality profiling over it, what are the implications there? As usual the Faustian trade-off appears to be increased functionality versus having your private data residing out of your absolute control.
Batting this one around with David, he pointed out that, apart from giving the Google Conspiracy Faction even more to fret about, this class of solution will probably evolve into a component of a master personal dataset stored online. This ultimate dataset facilitating everything from spell-check, drag-and-drop files/data/ in and out, to whatever your having yourself. Therefore not being computer/browser/Google-specific. All of which gives me a few more ideas...

*It does not at the moment. I used the made-up word “Cloodliberairiewooord” to check.

Delicious Tags: | | |
Technorati Tags: | |