Msg#: 4558109 posted 11:52 pm on Mar 24, 2013 (gmt 0)
Now that I'm about as blind as a bat I'm relying heavily on those little red underlines and I'm finding out what I'm not seeing nearly as many as I would like because the dictionary has a lot of words, many that I don't even know, but it turns a typo into technical correctness and if I miss it reviewing my text makes me look like a moron.
What they need is an option to let commonly used words, particularly it should learn your most commonly used words, and flag anything outside of the norm.
Why should my spell checker let old words hardly used or some vague scientific stuff fly thru without maybe highlighting it in yellow as this might not be the word you intended opposed to red which is obviously not a word at all.
It's nice the spell checker knows everything but most of us use less than a few thousand words on a regular basis, not 60K words, and it would help those of us that can't see what the heck is going on if anything outside the scope of that limited set, which we could obviously add to, would show up as a potential error.
For instance being one letter off when typing and "good" becomes "hoof" and how many times do you discuss a hoof? I'd put that in the "uncommon" list for me.
Accidentally added a 'g' to the end of "within" and got "withing" which is something to do about weaving that I've never heard of and I've learned a bunch of words I didn't know existed in the same manner of accidental discovery lately.
Not that we can do something about this directly, but I wish I knew who to contact that does the spell checkers for the browsers and suggest this change because it would help a lot of us struggling to cope from looking so stupid, especially with auto-correct trying to fix everything.
The other day I was replying to a text message and thought I sent SURE and sent SITE instead which made no sense at all and caused even more messages which were equally as hard to read to correct that typo but I digress...
Anyone got any ideas on this as I think it could be easily accomplished and help the visually impaired by helping us not send the wrong words.
requires a closer examination. (That was a simple cut-and-paste.) Q: What are words like "Prom" and "fax" and "modem" doing on the list? A: They are extremely rare in pre-1922 publications.
* Words that are mis-read by the OCR in a way that results in a perfectly legitimate word. The he:be pair gets a set of tests all to itself. I've got a further list especially for long-s books, though I don't suppose that is exactly what you require ;)
Msg#: 4558109 posted 11:21 pm on Mar 25, 2013 (gmt 0)
I think where you're going to see progress is in phone/tablet spell predictors. There's Swype and it's latest big competitor (forget the name) that basically study your writing and predict based on the words *you* use the most much like you train a voice recognition app.
Msg#: 4558109 posted 7:38 pm on Mar 26, 2013 (gmt 0)
I don't know as current versions of Swype have the Predictive Text Input technology I'm thinking of - I believe it's still in Beta.
Swiftkey and Swye are duking it out and Nuance recently acquired Swype so they are pushing things along too. This is a domain where there's actual competition unlike the desktop, so that's why I was thinking you'd see progress here, but I'm not sure you're seeing it yet.