6 ms·
Dotsies: Font using dots instead of letters - Optimized for reading, not writing
- pavel_lishin 15y agoError rate seems like it would be incredibly high from a human perspective.
- dfc 15y agoI was surprised by the lack of research justifying the claims about the latin alphabet and/or supporting the concept of dots. Does anyone know of any research in this field?
- breadbox 15y agoI'm not; I'd be surprised by any research that upheld any of those claims. They sound 100% rectally extruded.
- trogdoro 15y agoNever mind rectums. Try it yourself before you judge. If you're so smart, you should be able to memorize 26 letters without much of a problem.
- yew 15y agoThe ease of memorizing 26 letters is as irrelevant as the age of the Latin alphabet, really. This thing doesn't appear to have the least bit of intersection with any research. I suspect it's an art installation or something though. I've seen much weirder.
- trogdoro 15y agoWhat constitutes research? Standing in a lab in a white jacket? The best research for testing the readability of something is to read it. Help out by giving it a try yourself. How would Arabic look if you hadn't encountered it before?
- dfc 15y agoI'd like to hope that the design was influenced and informed by research in cogniotive science, HCI and so forth. And yes real research is in a lab with a white jacket and controlled tests, with hypothesis and broad sample size and statistical signifigance. 20+ years of progress with unicode and thid happens...
- Terretta 15y agoArabic, Sanskrit, and Idu are distinctly non Latin but incredibly easy to discern even to a first time viewer. Like Latin characters, they evolved. Your sample paragraph of simplistic un-anchored dots is to the human visual system as line noise is to speech. Which is to say, just because some geeks can sync over the phone with a 300 baud modem doesn't mean a whole lot of useful communication will result.
- trogdoro 15y ago> Arabic, Sanskrit, and Idu are distinctly non Latin but incredibly easy to discern even to a first time viewer Most of these look like spaghetti to me: http://en.wikipedia.org/wiki/File:Phrase_sanskrit.png http://en.wikipedia.org/wiki/File:Phrase_sanskrit.png If I scoot my chair back 4 feet from the screen discerning anything is hopeless, despite the font size being somewhat large. Whereas at the same distance much smaller English text is discernible to me.
- yew 15y ago> What constitutes research? Gathering experimental evidence by having many different people from many different demographics read many different things, preferably under controlled conditions. Comparing the resulting statistics to comparable statistics regarding other writing systems. Analyzing the interaction between the human vision system and reading. Probably a dozen other things besides, some of which do indeed involve lab coats. > How would Arabic look if you hadn't encountered it before? Probably weird. Distinguishable as a writing script, though. Your point? Anyway, assuming you're serious: This script is certainly more compact than the Latin alphabet. It's also completely lacking in things like contrast, visual cues, and optimization for letter frequency. Furthermore, it's factually incorrect that the Latin alphabet was optimized for writing, and totally irrelevant that it was invented thousands of years ago. Given that we have the Latin alphabet as a well-established and imminently functional existing standard, your argument simply aren't very convincing. Sorry.
- freehunter 15y agoYou're replying to the creator (or so he is leading me to believe through his comments), so you can straight up ask him if it's art for the sake of art.
- yew 15y agoAh, I wasn't aware. I wouldn't necessarily trust him to give an honest answer if it were art, though. As I said, I've seen stranger.
- switz 15y agoI always thought we could create a better language based on more efficient grammar rules and language habits. I just don't think the rest of the world can follow. I'm intrigued to try this out though. I'll see how it goes over the next couple of days.
- jimminy 15y agoIs this a piece of satire? I might be able to use it, but I still feel that I would need space between characters otherwise there is a risk of error, without needing to slow down and make sure things are aligned properly. Also the examples make it apparent that the font size would have to be nearly doubled for accuracy purposes.
- trogdoro 15y agoRemember when you learned the ABC's and you had to use really large letters? It's not because kid's don't have good eyesight. It's because your brain's recognition of patterns improves with exposure. Try the example on the bottom-left.
- jimminy 15y agoYes I remember that, but my eyes are actually adding elements to it and creating more noise around them. If you want to fit more information into a smaller space a script without a 1:1 relation to Latin alphabet might be better. It's almost like being asked to learn another language, and there are already other languages with higher information density in their character symbols.
- trogdoro 15y agoWhy high error rate? Do your own research by giving it a try. It'll only take you a few minutes.
- barrkel 15y agoLower redundancy means higher error rate. How would you distinguish between AAA, BBB, CCC, DDD, EEE without a baseline of text (for example, in a headline)? Without a reference for the top and bottom of the text, here's no way to know whether the dots are high or low.
- trogdoro 15y agoA good point, but try it in practice and you'll encounter these situations very infrequently. The adjacent words serve as a sufficient context. If the text is by itself, a small line will suffice for a baseline.
- dfc 15y agoAnecdote does not equal evidence.
- dsr_ 15y agoIt's optimized for nothing, as far as I can tell. Efficiency: letter frequency is not taken into account. Internationalization: no capital letter/lowercase distinction. No final letter forms. No circumflexes, accents, gravures, hats, dots, umlauts... punctuation is not specified. Physiology: no attempt to take into account how human eyes work. If this were going to work, Braille would be a lot more popular among sighted folk.
- csytan 15y agoNot to mention another big flaw. It's nearly impossible to differentiate between A, B, C, D, and E as individual letters since they are all represented by one dot.
- trogdoro 15y agoDon't miss the forest for the trees. If the letters are next to other words, it's easy enough. If they're on their own it is more difficult, but this is rare and there are simple ways to compensate.
- trogdoro 15y ago> letter frequency is not taken into account. False. > no capital letter/lowercase distinction False. > No circumflexes, accents, gravures, hats, dots, umlauts... True, but not impossible to overcome though if you'd use your imagination.
- preinheimer 15y agoCare to elaborate? I was thinking the same thing, all the vowels have a logical symbol in the pattern, as opposed to something unique or more striking.
- trogdoro 15y agoIf you look at dotsies.org you'll see capital letters have smaller dots above them to distinguish them from lower case. Using a pattern was done to make it easier to learn, but many options (with patters and without) were tried out and rejected. Notice that the vowels all have dots at the top row or bottom, which is desirable as it tends to give words hollow shapes with negative space in them. Optimizing for space efficiency has many dimensions, too dense is not good, and too sparse is not good. There were a couple versions where the frequent letters had fewer dots and were concentrated toward the bottom, but that increased possible ambiguity between words one might mistake for being shifted up or down one dot.
- csytan 15y agoIf the creator happens to read this, I'd suggest upping the font size significantly. There's a reason children's books have such a large font. When we learn how to read, it's much easier to focus on large single characters. Only after we become accustomed to reading individual characters can we then start to recognize words.
- trogdoro 15y agoVery true. I recommend using Command + (or Control +) to blow up the size. Didn't want to make it unnecessarily large by default.
- onli 15y agoBut you change the default to be smaller than the default (13px fixed size font instead of the 16px the typical desktop-browser has as default). You shouldn't mess with that even on a normal site, but especially not in this case. One can't recognize the dot-positions anyway, but making it smaller makes it only harder.
- jrockway 15y agoI might be convinced if I could see a list of words. The symbols are too difficult to distinguish individually, but common words could eventually become recognizable, perhaps. My opinion is: if you're going to learn something enough to be fluent in it, why not make it a real language spoken (or in this case, written) by actual people? And finally, I'll give the author twenty bucks if he can read arbitrary text that I supply. (This would have to be an in-person thing.)
- trogdoro 15y agoI'll take your twenty dollars any time. You in the bay area? I can read it about %75 percent as fast as normal text, and speeding up over time.
- subtenante 15y agoHow long have you been training? I'm not sure what 75% as fast means. Faster or slower? (If slower, what the point?) I would have chosen not to go the alphabetical order way to attributing dotsies. I'd have chosen single points for vowels, for example. And the consonants grouped by phonetic similarity. You're bound to Ctrl+ to be able to read anyway. So if you're making the dotsies bigger vertically to be able to distinguish things, why use only 5 vertical slots? 6 or 7 would allow more distinctive patterns. 2 dots at the bottom for plosives, 3 dots for fricatives, etc. Seems interesting as a personal thing, but I'd venture it's hard to teach to someone, way more than you seem to believe. Have you already taught someone in a few 20 minute trainings?
- jrockway 15y agoI'm in NYC, but I'll probably be in Mountain View in the near future. I'll let you know; I'm good on the bet :)
- trogdoro 15y agoExcellent:) I'm also trogdoro on twitter. Doing it over skype could work as well. You could use the bookmarklet on a local html page and share your screen so I'd have no way to know the original text. There are only 26 letters. It's much easier than it seems. It just takes time to build up your speed and be able to recognize smaller typefaces.
- nhebb 15y agoI can't tell whether this is supposed to be serious. Here's a simple test, show someone the dot => alphabet mapping for A to E. Next, write a dot in the middle of a plain white piece of paper and have your test subject tell you which letter it is.
- freehunter 15y agoYeah I tried the first example (A-E) on their memorize site and was blown away by how poorly thought out the whole thing was. It asked for the answer in dotsies and all that came to mind was "... uh, a dot somewhere around this region? -click- oh nope, it was up here further. How mind-numbingly interesting."
- trogdoro 15y agoIf you've got a better mapping in mind, post it. It would be interesting to see. Many things must be taken into account beyond just where the dots seem like they should go on the letters. However, there's a decent enough correlation on enough of the letters that I suggest it would be pretty easy for you to remember them. For example, A, C, F, H, I, K, O, P, R, S.
- freehunter 15y agoI'm just not convinced that single dots work well without gridlines showing their relative position on the page. Maybe it needs a better demonstration, or maybe I'm just not getting it. I don't want to take away from your accomplishment, I can tell a lot of thought went into it. But, like Dvorak, it might be better, but is it better enough to put in the effort to switch?
- dsr_ 15y ago" The latin alphabet (abc...) was created thousands of years ago, and is optimized for writing, not reading. " ... that's demonstrably false. Cuneiform is optimized for writing (with a stylus). Gregg shorthand is optimized for writing with a pencil. The Latin alphabet forms have lots of redundancy (though not quite enough) in 2D visual space in order to be clearly recognized and distinguished. Lower-case is better than upper, in most fonts.
- trogdoro 15y agoConsider that before the printing press was invented, the way most words were produced was by writing them by hand. The next time you're using a pen, ask yourself whether the letters you're writing seem A. pretty convenient to write, or B. difficult to write for logistical or other reasons. If you're leaning toward A, it's no big surprise since the people who developed and evolved it were also using a writing utensil, as with almost every other http://en.wikipedia.org/wiki/Writing_System http://en.wikipedia.org/wiki/Writing_System. I make no claim about whether it is more or less optimized than other writing systems. People used chisel's here and there, but that has little bearing.
- tiddchristopher 15y agoTypographically, this has horrible legibility: the letters are difficult to distinguish and recognize on an individual basis. This contributes to poor readability when the letters are grouped to form passages of text. This seems much more about compacting information in a human-decipherable format than creating a more readable alphabet. There are no ascenders or descenders in a traditional sense, no contours or apertures, and there is no stroke contrast. Also, the baseline is thrown off in many words. Put this all together, and you have text that's miserable to read. This isn't even to mention the difficulty of getting readers to abandon tradition. In all, I'd consider this an interesting experiment, but a practical failure.
- trogdoro 15y ago> letters are difficult to distinguish and recognize on an individual basis The important part is that the words are not difficult to distinguish. > Put this all together, and you have text that's miserable to read. Interesting how you can come to that conclusion in a few minutes. It would be tempting to write off Kanji if you encountered it for the first time, no? Give it a try before you shoot it down out of hand.
- yaakov34 15y agoPlease explain the advantages of this over reading Braille dots (which lots of sighted people can do already), and please explain the lack of reports by sighted people saying "hey, I noticed that reading in Braille is faster and easier than with the regular alphabet."
- trogdoro 15y agoLook at some text in braille and look at some text in dotsies. It looks pretty different. Dotsies letters are smashed together so the words look like shapes of their own. If you did that with braille it would likely be confusing due to each letter having 2 rows of dots, and it would be stretched out to about the same proportion as normal text.
- yaakov34 15y agoThat makes about no sense (except as really deadpan satire), the strength of the human visual system is in "chunking" shapes, and making letters into clouds of disconnected dots is about the last thing you want to do to it, if you don't have to. If you do have to, there are already well-established systems that do this (Morse code and Braille). Teletypes switched to printing actual letters as soon as they could, even though there were people very highly trained in reading the dots and dashes.
- trogdoro 15y ago> the strength of the human visual system is in "chunking" shapes Which is the main strong-point of dotsies. Words are more chunked together, and appear as recognizable shapes. > and making letters into clouds of disconnected dots They're not disconnected. That can be easily seen by a cursory glance at dotsies.org.
- yaakov34 15y agoThey sure are disconnected; you've made clouds of dots instead of continuous shapes. Ponder the fact that a dumbbell shape is easier for a human to process and recognize than two disconnected dots. On the other hand, if you're on some kind of a business-school inhibition-reduction assignment (you know, like "get up in public and say the stupidest thing you can think of"), then you're doing well.
- trogdoro 15y ago> They sure are disconnected; you've made clouds of dots instead of continuous shapes. You may be looking at the table of letters, which is just for illustration, where the letters aren't connected. But if you look at the examples you'll see they are very much connected, and appear to form shapes.
- the-cakeboss 15y agoI won't dismiss this outright. I do find it to be fairly interesting. If a system like this were to gain wide spread usage, I think many invaluable insights on reading ( and all those behaviors associated with it ) would be realized regardless of how effective a system this proves to be.
- marshray 15y agoI guess as a bonus anyone shoulder-surfing your screen will think you're jacked-in to The Matrix.
- nwatson 15y agoif this idea has any legs at all, the first improvement will be to distinguish vowels from consonants ... and then to make sure common letter sequences that have particular meaning (English), -ing, -ly, -tion, un-, in-, etc., stand out visually. I can live without upper/lower-case distinction for most reading. and common 1- and 2- letter words need to look unambiguous when floating independent of a readily visible base- and top-line you'd see in most longer words.
- chromedude 15y agoNot sure why there are so many upvotes - I don't know anybody in their right mind who would use this.
- sheraz 15y agoChrist people! This is so stupid I have to think we have been trolled with this one. "Optimized for writing, not reading?" You have to be kidding me, right? Yes -- and the Romans had indoor plumbing, aquaducts, and crosswalks. All just as antiquated as the serif. See there? I've fed the trolls. -- I would love to see the brouhaha this would elicit from typophile.com.
- mwhooker 15y agohmm, z is indistinguishable from py. (at least on my screen).
- maytc 15y agothere's a problem with your font.... take 'a' with twice the pixel size with 'ff' the letters will look exactly the same.....
- gojomo 15y agoMerely black and white? Sure, writing was developed when there was just black ink and parchment, but now even free (with service plan) handheld devices have rich color displays. Convert each letter into the trinary representation of its 0-based ordinal position – 26 is conveniently close to 3^3, isn't it? Use each digit as a scaled R/G/B value. So for example, 'a' is letter 0, trinary 000, HTML color #000. 'g' is letter 6, trinary 020, HTML color #0F0. And so forth. Every one of Dotsie's dots can now be a whole letter! I can fit the full text of 'Huckleberry Finn', 'Pride and Prejudice', and 'The Adventures of Sherlock Holmes' on my 1920x1200 monitor – and still have room for the 'Kama Sutra'! (Free Gutenberg Project plain-text editions, of course.) Yes, it may require a jeweler's monocle to read, but think of the space optimization. The color-blind can join the blind in relying on braille.
- Geee 15y agoReally interesting, I don't get it why everyone is so eager to shoot this down. Of course you can't know the benefits before you learn it. Also the pure dots present just the base system, typography and different typefaces can evolve from there with more aesthetics, redundancy, readability etc. I'm definitely up for it for science/entertainment!
- trogdoro 15y ago> Also the pure dots present just the base system, typography and different typefaces can evolve from there with more aesthetics, redundancy, readability Absolutely. There could be several improvements in that way. Maybe along the lines of turning a square of 4 dots into a circle, a line of diagonal dots into a smooth line, etc. Some words can be drawn out pretty simply with a pen, as if the line were passing through the positions of where the dots would be. This is even more pie-in-the-sky than using it as a font but is fun to speculate about, though I fear it may ruin the afternoons of a few HN commenters to see it mentioned.
- dacoinminster 15y agotrogdoro, I love this idea, but I'd suggest choosing dot combinations that visually approximate the shape of letters where possible, For instance: - lower case L (l) would be a long vertical bar - i would be the lower two dots filled and with a dot in the second highest position You could then make some other common letters look like a vertical cross section: - o could be the bottom dot and the middle dot - s could be the top, middle, and bottom dots - m could just be the bottom dot - w could just be the top dot More complex letters could represent crossing lines as gaps. For instance: - t could be a vertical bar with the second highest dot empty - k could be a vertical bar with the middle dot empty You could add a sixth dot underneath the normal line for letters which commonly drop down below. For instance: - j could be an i with an extra dot underneath - p could be an o with an extra dot underneath - g could be an un-dotted j - y could be a v with an extra dot underneath You could take advantage of certain letters which are the inverse of each other by inverting the dots (white dots become black, and black dots become white. For instance: - z could be the inverse of s - q could be the inverse of p (but with the lower dot still filled in) For letters which are upside down versions of each other, just make the symbol upside down - w would be an upside down m (as already shown above) - n would be an upside down u You could also add an extra dot above the normal line of reading to denote capitalization (kinda like a representation of the shift key). Doing this would make for 7 dots total. I'm sure that not every letter would be representable using rules like this, but the less common letters could get the less obvious patterns. The nice thing about doing it this way is that some words will kind of look like the shapes we have already memorized for those words. Kind of. If you squint. This is a fun experiment, and I commend you for trying it. I look forward to seeing vs 2.0 :)
- trogdoro 15y agoGreat ideas. > This is a fun experiment, and I commend you for trying it. I look forward to seeing vs 2.0 :) Thanks! I may code up something to let people pick different mappings and preview how it would look on some text. If I do I'll post it on HN.
- pg_bot 15y agoThis seems like a solution in search of a problem...
- trogdoro 15y agoI don't propose that Dotsies is without flaws or destined for any amount of adoption. Having said that, here is my parody of the reaction the poor sap (let's call him Marcus Librus) got who originally posted the latin alphabet on the ancient Roman version of Hacker News (let's call it Papyrus Tidings): Brutus Quidus: I've looked at your alleged alphabet for 30 seconds and have discovered all of the flaws. It is clear it will never work. First off, your O and Q are nearly impossible to distinguish. What if there's a piece of dirt on my scroll where I put the O? It's obvious you didn't do any research. Claudius Acerbus: Is this serious? This will never ever work. Besides the achilles heel that if you have an N next to an I people will think it's an M, that circle thing looks just like a D. And if you have 2 D's on top of each other people will think it's a B! Christ! Err, wait, not born yet. You'll see!
- jimminy 15y agoEven your parody is flawed. If you go back to when the Latin Alphabet was introduced there were approximately 100 million people in the entire world, and only a subset of a subset would have been required to adopt this. Over nearly 3 millennia, the Latin alphabet has been improved upon and become widely adopted by even common-folk, to the point several billion people are familiar with it. You're task of adoption is several magnitudes harder. Your alternative provides no significant value over the Latin Alphabet, unless you want to use it for a simple obfuscational encoding. If you want to reduce the size and amount of visual information provided, look at Kanji, mathematical syntax, or shorthand notations (although these tend to vary by person). Don't get so defensive. As it is it's not very good for tons of practical purposes, many that have been pointed out here. You even said that you can't interpret it as quickly as you can the Latin Alphabet, yet. Find a way to fix the flaws that have been pointed out, and come back. It's a novel approach and idea, but not practical for any use, currently.
- gavinpc 15y agoWait, what problem does this solve? I have to agree with @nhebb in doubting whether this can be serious. The idea is provocative, but contains the seeds of its own destruction. If the deficiency in Latin script is that it is (putatively) "optimized" for one function and not another, why make the same (or rather opposite) mistake on purpose?
- trogdoro 15y ago> why make the same (or rather opposite) mistake on purpose? The thought is that it's low effort to click a button on a computer to change the font of something you're reading (vs high effort to do the analogous thing to text you've written). The intention is to augment latin text in a small way (you can use it when you feel like it) more so than to replace it. > Wait, what problem does this solve? It's helpful to consider the extremes. Let's say the convention was to represent the "A" sound you were supposed to draw a tropical island, including palm trees and a peaceful lagoon and a volcano. (And for the "B" sound, you draw a herd of buffalo, etc.) We would probably all agree that the palm trees are superfluous and should be removed. We'd probably further agree that maybe something even simpler like just the letter "A" would be an even bigger improvement. Desert islands take up a lot of space on paper, are a lot for the eye to look at, and when viewed en masse are less visually recognizable than simpler shapes. Along that line of thinking, why not simplify even further? Whatever your objections, you'll probably concede 2 or 3 dots is simpler than most letters. A valid objection would be that there are some detractions that outweigh the benefits of increased simplicity and efficiency, or that the benefits aren't significant enough to bother with. But, I think the jury is still out on that.
- gavinpc 15y agoI agree of course that it's simpler. This reasoning reminds me of the "nutritionist" argument (as I understand it from Michael Pollan's book In Defense of Food), in which individual nutrients are presumed to contain in toto the benefits of the foods in which they are found. Sure, a leaf of spinach may contain some inert matter that is irrelevant to your body from a chemical point of view. But it has not been convincingly demonstrated (he argues) that "the good parts" can be extracted or understood separately from the whole. My gut reaction, and I'm sticking with it, is that this is the reductio ad absurdum of a beautiful, human mystery. But take that with a grain of sea salt, as I was a student of "letters" before I was a hacker.
- drobilla 15y agoMaking a modern condensed character set is a nice idea, but this is a really, really bad execution. Many letters are the exact same shape, shifted vertically 1/5th of a character height? A-E are completely identical but for position. This is terrible enough to be a show-stopper on its own. Braille has all the advantages of this, but is FAR more readable, since each letter has its own distinct shape. You say in a different thread that "Braille has the same proportional shortcomings as normal text". This is false. Braille characters are a grid of a fixed number of dots (6 or 8), and every character is the same size. It also has numbers, punctuation...
- Arun2009 15y agoWhen I read something (mentally), I can see that there are two parallel processes that take place: 1. The "physical" aspect of mental reading - mentally sounding out the words that are in front of me. This is quite fast already and is pretty mechanical. 2. Converting the words into meaning and registering it in my mind. This is the much harder task, and you actively need to focus to do this. To see that these two are indeed parallel processes, notice that it's possible to "read" something in your mind without any of it actually registering while you think of something else, say a narrative about just how awesome your life would be if X or Y were true. I often have to actively break the internal narrative and pull myself back to focusing on what's read. The focusing part is the real bottleneck, not the mechanical reading part. This is painfully obvious especially when you are trying to learn information-dense material such as Mathematics, where you have to pause to recall definitions, follow argumentation, and so on.
- trogdoro 15y agoI added a link to an interface for trying your own mappings. It didn't get any notice, so adding it here as well: http://dotsies.org/design-your-own/