Okay, imagine you wanted to make the perfect Autocorrect + Grammar-checker. You work for Apple, Microsoft, Google, or Facebook, so resources aren’t an issue, and you have access to as much data as you could possibly need.
Okay, so there a few things you want to look at, you want it to get full sentences, you’de rather it not-fire than misfire, and sometimes people are making mistakes on purpose. For the sake of this one, you want the program not to be too heavy, and you probably also want an external database to measure what is slang and what is human error. So first things first, what do we want the program to do?
We want the program to read text being written and near-instantly fill in a word that is missing, add punctuation and capitalization when needed, except on certain types of text messages, to adjust it’s setting to the user in question, to give the users options on what they want the program to do, to correct non-intentional spelling errors, and to not get in the way of the user. The end goal, the primary purpose of our Autocorrect program is to make communication easier and clearer without requiring additional user input. Now this is all well and good but some of our features may contradict that so let’s get them out of the way.
Intentional misspellings and slang are a thing, and while we want our program, or better yet the smart database (separate server) behind our program to recognize their usage. I’ll cover the database in the next paragraph but it can’t include everything; while some words are purposely misspelled more often than other we do some incredibly complex things with communication and sometimes the way we right things has to do more with delivery than with any form of consistency. So what I was getting to with this long ramble is this: If you don’t want your word to be autocorrected, swipe up after pressing the spacebar, or shift, or have the spacebar be split into 2 on mobile, really I don’t know what will be most comfortable so it needs product testing; really you should be able to chose the setting when setting up your phone or the app that uses it but if we need a default we should test what’s most comfortable, or more specifically what is most noticed by having the product testers only have one option and seeing how it impacts their experience. I believe that this one feature already puts us ahead of most Autocorrect software out there, and it means we can actually collect data by seeing what purposeful misspellings and acronyms are more common than others.
That was far too much rambling (I’ll work on it in future posts) but here’s the next feature. A long startup sequence where you get to select your features and all that.
You know what I’m done with this I can’t seem to get this in order, I haven’t gotten into the formatting and the pattern recognition, and I had my one good idea for today. Next time I make something of this sort I’ll do it in parts, in order, and publish one complete thing at the ends instead of trying to work on all the factors at the same time (hey I’m a one man team here) for now I’ll complete the subject.
A Few Possible Consequences Of A Near Perfect Autocorrect (I work better in lists I would have done ten but I’m done with this topic I’m not giving it any more extra time)
- Language Changing More Slowly – Through the history of the English language (others also but I know more about English), the spelling and pronunciation of many words have changed. While less so in modern times, a change I would predict happening in the next twenty years is that it will become common practice to spell though as tho and it will become so prominent it will be accepted into the language. The problem is we have a much more interconnected world nowadays, and enough people care about the right way to spell things, so these changes have slowed down somewhat. The existence of a perfect Autocorrect may slow down the shifts in language, even more, freezing spellings to the time it was made.
- Language May Change Faster – So I’m stating the exact opposite of my previous entry here it can go both way’s and here’s why. A popularly used Autocorrect won’t correct slang and will have an evolving (probably also dependent on region) library of slang terms it will leave-in place. Some words originally were slang but have become commonly used in most circles, i.e., cool, emoji, prioritize, etc. The use of slang terms in digital communications will bleed into language and slang spellings will be seen as correct by a large portion of the population.
- Language May Become More Localized – If there are sperate local databases for what is considered correct. Or even if the option to select between British, Australian, and American English remains the databases will both evolve separately and change separately causing the three versions to change, then it is only a matter of time before people start making their own personal regional versions.
- People Will Become Worse At Spelling- most of most people’s writing is done through a digital medium and we won’t need to bother with spelling the computer will do it for us.
- More People Will Write More- If there is a writer’s version, and if we don’t make one someone else will reverse engineer it. Then writing will become far less of an imposing task, aspiring writers won’t have to worry about making simple grammatical or spelling errors so they will spend far more time thinking onto the page for lack of a better term.
- People Will Read More Niech Content – The people writing that stuff aren’t always good at writing and more will be written anyway…