Why can’t we have our own spell checker as a personal assistant, which understands which keys are next to which on the keyboard, and therefore most likely to produce the same errors, which in most cases are easy to see, and what’s more in all the languages I usually write in? In short, by now, surely a machine learning algorithm can be developed to quickly scan our previous texts and check the vocabulary we use and the most frequent mistakes we make so as not torment us with incorrect suggestions or red underlining? And yes, I know that I can add terms and specific words to the dictionary, and that I can even configure substitutions on my computer, but nobody in their right mind would waste their time doing so, and besides, it only works for specific programs. For some reason, I often write buisness instead of business (which is interesting, considering I teach in a business school!).
Is it really so hard for developers to understand the keyboard layout, and that when a word incorporates the letter next to it instead of the corresponding one, it should be corrected? As a Spanish and English speaker, my most common typographical errors are, for example, hitting the “ñ” key when I want to put an accent, or the “k” key when I intend to write an “l”, meaning it would be extremely easy to understand by inspecting a few other texts: we all have our own, and will see them come up time and again. Most of my spelling mistakes are accidental typos, which are particularly annoying, because they seem even easier to correct. Given that most of us have a large file of texts written over the years in word processor documents, e-mails and on social networks, why we continue to use generic proofreaders that are only capable, at best, of incorporating some of the words we generally use? Why don’t those proofreaders automatically eliminate, without even bothering us, our typos? How it is possible that such a seemingly simple, but essential function as a spell checker is still based on a standard dictionary, a practice that we know produces so many undesired results and leaves little opportunity for personalization?