Keyboards, UX, and “Fuck”
blogmask.herokuapp.com
blogmask.herokuapp.com
There's a vast swath of people who don't want to use curse words for various reasons, and how terrible it would be to accidentally insert "fuck" in your work-only phone!
Besides, we aren't all super coordinated with phones. I've been typing on these keyboards for two years now, and I wouldn't be able to live without auto-correct. I'm guessing that, since you are such an expert of using the keyboard that you apparently never misfire, perhaps you ought to roll without auto-correct? I know a few people that do...
From the keyboard, hold comma to get to the settings menu, pick Google Keyboard Settings, scroll down under "Text Correction", and uncheck "Block offensive words".
It still stays shy about suggesting naughty words when you swipe ("gesture type"), though. I got curious, and it looks like manually adding the ones you care about in Settings > Personal Dictionary gets around that. Just for completeness, voice recognition has its own "block offensive words" option in its settings (phone settings -> language and input -> settings icon next to "Google voice typing").
I don't see a way around it in SwiftKey. I wouldn't mind a similar checkbox (hey, always good to match Google Keyboard's features, right?) but it hasn't been a huge deal--it's like typing names or other things not in the dictionary (tap the left completion when you're done), which works for me.
Fucking annoying, but relatively easy to get the little fucker to output the shit I want.
Actually, the most mistakes I've ever made when sending blurbs of texts was the several months before I figured out how to disable autocorrect on Tweetbot on my Mac. Man that was terrible. Kept having to delete tweets because I hit "Send" before fixing the autocorrected text.
Alas, if you frequently use two different languages and often mix them both in the same sentence (common among bilinguals) autocorrect is your absolute bane because it will always do the wrong thing.
The sticking point for me, so far, has been when I use an English word but declense it like it was a Slovenian word. Every time I do that, autocorrect thinks I can't type even though I typed exactly what I wanted to.
And with 18 different declension suffixes for nouns and ~15 for verbs and proverbs, it would take autocorrect ages to learn how to spell. Especially since the correct spelling is context dependant and usually only a letter different than the wrong spellings.
Scratches an itch I had with posting to HN, sometimes I just want to write a quick rant that's somewhat off the record.
Currently in beta, feedback is very, very welcome.
"Technology" is misspelled here: https://blogmask.herokuapp.com/browse
Also, why copy the Kudos feature from Svbtle?
As for the kudos mechanism, people seem to be happy to engage with it. I know on HN it's something that everyone recognizes but the average web user doesn't notice it (kind of like Bootstrap).
I was recently setting up my daughter's cheap Android phone, when I decided to send a test Viber message to my wife. I started typing "Hello mom", and "mom" got autocorrected to a declination of "mother" which is 99% of the time used in "f-ck your mom" phrases (мамата, for compatriots). My daughter is old enough for a smartphone, but too young for this grade of swearing; I considered it highly inappropriate, and just one more appalling thing in this first contact with Android.
Crowd-training obviously has its limits. And you should err on the side of caution in this sort of things.
(The other appalling thing was the obviously crowd-translation of the base OS, with glaring typos permeating the interface; some were occasional, but some of the most offending ones were systemic, e.g. the word Audio was wrong everywhere. You'd think between them Google, Alcatel and a mobile network operator working in half of Europe could afford to have an editor read through a few dozen thousand strings.)
I see your point entirely, and you're right to complain, but I doubt I would've seen that coming if I were making an autocorrecting keyboard.
If I type 'fick', it'll try to correct to 'duck', but 'fuck' doesn't get corrected.
That seems terrible. I only want to add words to a dictionary explicitly. Make a typo often enough now you got teh in your dictionary. I turn off autocorrect anyway.
To bypass the five times, just type a word into safari's omni bar and it will add it to the dictionary.
it would be an awful ux if words needed to be added explicitly to the dictionary.
What I'm learning today is that there is no such thing as best UX practices.
Some people are clearly adamant that having to add words explicitly is awful. After all, the phone should be able to figure out what is wanted, right?
Other people are just as certain that having the phone implicitly add words using any specific rule will fail often enough that it's obviously worse UX to try.
I find myself in sympathy with both camps. I would like my devices and environment to just know what I want and do it without my having to explicitly direct actions. Unfortunately, it doesn't seem to be working out that way for a lot of areas. Instead, we have a mishmash situation where devices and apps try to do what most people want, and fail fairly often for a given user (unless I'm a major outlier). Ten years ago, when my computer failed to do what I expected, I understood that I'd failed to use it properly. I read documentation until I understood the failure or, at least, had some idea of what I might have done wrong, and tried again, and eventually I would understand the system.
Now that the system is supposed to understand me, I've noticed two effects:
First, detailed documentation of user interfaces is incredibly scarce. I wonder if there's a perception among other developers that if you need documentation, the UX has failed, and therefore the most direct route to apparently better UX is to have less documentation?
Second, I find that partly due to the apparent complexity of my phone (or perhaps personification of it), I'm more likely to just swear at it and complain about it than to actually try to understand what I'm doing wrong.
Of course you can obfuscate by escaping to hex, if you want.
I looked at the filters for Windows once. The usual profanity, but a few surprises, including many mentions of keywords involving the consent decree, FBI, Janet Reno and so forth.
Sent from my iPhone
Seriously, words like "shit" or "penis" aren't offensive at all these days. Other are just collegial and unprofessional. For someone from Europe this really looks like needless uptightness. It's like a kid isn't even allowed to see her naked mother.
Also, from some educative point of view, I think it's likely the worst solution to simply conceal them from your kids. As a parent you could easily explain their meaning, that it might be rude for some etc. Actually, the important point is to be reflective when learning those words. Those rather simple words are not like showing hardcode porn to children. I don't wanna say that you couldn't be verbally offensive, but that's not the point here.
If it "just works", of course it won't give you only some people want.
What makes this a bit more interesting than usual UI stupidity issue for me is that the walled garden can get away with certain personality just because it fits the marketing. And certain personalities get bought along with the hardware.
Text messages and Facebook posts/replies are mostly German, Twitter is 50/50 and random stuff is likely to be English as well.
e.x. http://d.pr/i/tQgZ