Learn to read a binary font by reading a story • Hacker News Board

Re: Learn to read a binary font by reading a story

#111

Post by Cushman » Sat, Mar 31, 2012, 4:39 PM UTC

trogdoro wrote: post #65

Cushman wrote: post #50

Earlier quoted context omitted.

Yeah, it's been proposed before... It's pretty hard to convince everyone they should start using a new language nobody else can read. (More people now speak Klingon than Esperanto. Ouch.) I think computers change the equation, though. Machine translation would be almost impeccable. Using phonetics visually would make chording intuitive in a way it hasn't been until now... You can do a soft launch where you start out…

> The idea with chording is you press symbols simultaneously to write an entire word That's not typically the case - http://en.wikipedia.org/wiki/Chording_keyboard . > this is how stenographers are able to type >300wpm Stenography is very fast, but can't replace a normal keyboard (i.e. you wouldn't want to use MS Word with it or code with it). Also it has a large learning curve and can vary from person to person. > u…

You got me there, I was using "chording" when I really meant stenotype. Just being careless, I guess...

Re: Learn to read a binary font by reading a story

#112

Post by akavi » Sat, Mar 31, 2012, 5:07 PM UTC

trogdoro wrote: post #45

trothamel wrote: post #36

Earlier quoted context omitted.

Dotsies aren't all that space-efficient, though, when compared to regular old text. For example, in this comparison: http://i.imgur.com/Ldd70.png I'd argue that the text on the right, in the green font, is just as readable as the dotsies, while taking up less space.

That's probably largely because the circuits in your brain have become finely honed on latin characters, after reading letters many millions of times for many years. It's difficult to know how that image would appear to someone who was equally acclimated to both. At the moment no such person exists. One clue might be how they compare to each other after you scoot your chair back from your computer until both have bec…

I've put about five hours in to this (The magnificent motivating power of procrastinating on something-else-you-need-to-do), and already I feel like I can identify individual dotsie words from a greater distance than I can the Roman words.

So I suspect your thesis may not be entirely unfounded.

Re: Learn to read a binary font by reading a story

#113

Post by tennis » Sat, Mar 31, 2012, 6:39 PM UTC

This seems flawed in that it arbitrarily translates regular symbols into what looks like a graphic error soup. I feel like letter and di/tri-graph frequencies should have been considered, as well as conventional expectations (m looks thicker than i, f.i.).

Re: Learn to read a binary font by reading a story

#114

Post by Groxx » Sat, Mar 31, 2012, 6:56 PM UTC

suby wrote: post #88

Cushman wrote: post #33

This seems like it could actually be a real improvement on written Latin if it adopted a real phonetic alphabet. Just transposing all of my existing graphemes for new, smaller ones seems a little silly. (And since what you're really asking is not for people to learn a new alphabet, but to learn a new symbol for every word in their lexicon, it wouldn't be that much more of a leap. Think about it.) Thinking about it--…

I'm going to school right now to become a court reporter / stenographer. Basically, I type on a funny looking keyboard ( http://www.acculaw.com/images/products/display/stenograph%20... ) where you're not so much typing letters as much as you're typing sounds. So the word "talk", for example would be written "tauk", with each letter of the word pressed at the same time (if possible. Some words you need to come back fo…

As an outsider, I don't see why reading "TKAOEUPL" instead of "dime" would be an improvement (except for checking precisely what you wrote) for even professionals. Is there a reason why these machines don't convert the roughly-phonetic words into English words? I would think it wouldn't even be too hard to program, even with those 'little spin's, considering how far we've come with grammar checkers and gestural keyboards like Swype.

Re: Learn to read a binary font by reading a story

#115

Post by twfarland » Sat, Mar 31, 2012, 7:06 PM UTC

houshuang wrote: post #5

Not sure if this is worth the effort, but brilliant approach to learning. I've often had fun playing with inventing my own alphabets and stuff -- if you could actually learn to read this font at a decent speed, it would be fun to click a bookmarklet anytime you wanted some privacy :) I love reading, and I read pretty fast - I've often thought that the existing approach in most textbooks where they give you a short (5…

This is a really nice idea. Reminds me of Guy Steele's 'Growing a language' talk (http://video.google.de/videoplay?docid=-8860158196198824415) and generally of building a formal system by combining primitives.

Re: Learn to read a binary font by reading a story

#116

Post by abecedarius » Sat, Mar 31, 2012, 7:11 PM UTC

david_ar wrote: post #48

I really like this approach (even if I'm still not entirely convinced about the benefits of dotsies). For a while, I've had this idea of a book that starts in standard English, and gradually alters the grammar and introduces new words - so that halfway-through it's written in somewhat of a pidgin/creole language, and by the end you're reading a completely different language (e.g. French or Japanese). I'm not aware of…

http://www.cs.virginia.edu/~evans/cs655/readings/steele.pdf by Guy Steele is a little like that.

Re: Learn to read a binary font by reading a story

#117

Post by twfarland » Sat, Mar 31, 2012, 7:16 PM UTC

oskarth wrote: post #97

Cushman wrote: post #33

This seems like it could actually be a real improvement on written Latin if it adopted a real phonetic alphabet. Just transposing all of my existing graphemes for new, smaller ones seems a little silly. (And since what you're really asking is not for people to learn a new alphabet, but to learn a new symbol for every word in their lexicon, it wouldn't be that much more of a leap. Think about it.) Thinking about it--…

There's simpel-fonetik, which is extremely intuitive. It's basically a mapping from IPA to latin letters. Example: When you read Simpel-Fonetik words, you must pay attention to each letter. Remember: each letter has always the same sound, the sound given in the Simpel-Fonetik alphabet, regardless what letter is next to it. And here it is re-written in Simpel-Fonetik: Wen ju riid Simpel-Fonetik wörds, ju mast pei äten…

I love this kind of thing. I tried something similar a few years ago, after visiting Turkey and seeing their language. I'd had a single character for each sound, though (i.e. 'å' instead of 'ei' for the 'a' sound in 'late').

Re: Learn to read a binary font by reading a story

#118

Post by devrim » Sat, Mar 31, 2012, 7:18 PM UTC

baddox wrote: post #86

Cushman wrote: post #33

This seems like it could actually be a real improvement on written Latin if it adopted a real phonetic alphabet. Just transposing all of my existing graphemes for new, smaller ones seems a little silly. (And since what you're really asking is not for people to learn a new alphabet, but to learn a new symbol for every word in their lexicon, it wouldn't be that much more of a leap. Think about it.) Thinking about it--…

Unless you're a geek interested in language, computer science, or programming who doesn't want to sink a lot of time into a mostly useless hobby, you should read about Lojban. http://en.wikipedia.org/wiki/Lojban

i'm sure right after this carefully thought, polite and welcoming comment, he'll most definitely stop what he's doing, and do what you suggest.

(disclaimer: i'm just generally against comments like these, my apologies to other readers that didn't have to read this)

Re: Learn to read a binary font by reading a story

#119

Post by javajosh » Sat, Mar 31, 2012, 7:34 PM UTC

Groxx wrote: post #114

suby wrote: post #88

Earlier quoted context omitted.

I'm going to school right now to become a court reporter / stenographer. Basically, I type on a funny looking keyboard ( http://www.acculaw.com/images/products/display/stenograph%20... ) where you're not so much typing letters as much as you're typing sounds. So the word "talk", for example would be written "tauk", with each letter of the word pressed at the same time (if possible. Some words you need to come back fo…

As an outsider, I don't see why reading "TKAOEUPL" instead of "dime" would be an improvement (except for checking precisely what you wrote) for even professionals. Is there a reason why these machines don't convert the roughly-phonetic words into English words? I would think it wouldn't even be too hard to program, even with those 'little spin's, considering how far we've come with grammar checkers and gestural keybo…

It's funny how when someone on HN says, "My method works well, but there are some pathological cases, like X" someone else on HN will invariably reply with, "Yeah, but I don't see why X would be an improvement."

This is a bad practice not only because it misrepresents the original thought (the person was not claiming that X was an improvement, they clearly called it pathological), but more perniciously it subtly erodes the (altogether sensible and honorable) habit of good people to give full disclosure of their ideas and methods. To put it simply: attacking someone on the weaknesses that they themselves have revealed about their own process or idea is neither fair nor useful, and does not contribute to the discussion.

Re: Learn to read a binary font by reading a story

#120

Post by Groxx » Sat, Mar 31, 2012, 7:41 PM UTC

javajosh wrote: post #119

Groxx wrote: post #114

Earlier quoted context omitted.

As an outsider, I don't see why reading "TKAOEUPL" instead of "dime" would be an improvement (except for checking precisely what you wrote) for even professionals. Is there a reason why these machines don't convert the roughly-phonetic words into English words? I would think it wouldn't even be too hard to program, even with those 'little spin's, considering how far we've come with grammar checkers and gestural keybo…

It's funny how when someone on HN says, "My method works well, but there are some pathological cases, like X" someone else on HN will invariably reply with, "Yeah, but I don't see why X would be an improvement." This is a bad practice not only because it misrepresents the original thought (the person was not claiming that X was an improvement, they clearly called it pathological), but more perniciously it subtly erod…

If it came across that way, I'm sorry - I'm genuinely curious why such machines don't convert what was typed into English. In the past it was almost certainly a lack of computational power (or zero) - what about now? I was only asking about the reading portion, the writing is pretty clearly faster.

I'm not in any claiming that they should be typing on QWERTY. I'm wondering why the pathological case exists at all when it seems it doesn't have to, and if there's a reason for it that I'm not aware of. Maybe there's a market for machines like this to spit out English, maybe there isn't, but I'm not in the field so I really don't know.