I got interested with these symbols when my son was born and I was doing all of our laundry what seemed like all hours of the day. So it was a good opportunity to play with machine learning and vision edge by gcp. Ended up building a pair of apps for iOS and android that can identify the symbols
I noticed a huge difference in how long my clothing lasts once I switched to hang-drying most things. It's not that huge of a hassle and saves energy and I don't have to buy clothes as often. I started doing a lot more hang drying after I bought a bunch of merino wool base layers I use for skiing (which can't really be machine dried).
If you look near the laundry baskets at any home or department store, you'll find comically large 'lingeree bags'. Turns out running anything with a fine weave through these - satin, rayon, exercise clothing, high TPI pillow cases - not only makes them last longer but also prevents pilling.
Always button and zip your jeans, and if you're not in a hurry, cotton clothing seems to be less worn by friction in the dryer than by the high heat. I run a lot of my cotton knits through twice on permanent press instead of once on cotton. And I don't use dryer sheets. Dryer sheets keep your clothes from getting static cling when you have over-dried them, but over-drying them damages them. The static cling is a symptom that you shouldn't ignore.
What you want to do is pull your clothes out when there is just a hint of moisture in them. The air and the latent heat should be more than enough to suck out that last hint of dampness. And if one towel or pair of pants is still damp, nothing stops you from running them by themselves for a couple minutes while you fold the rest.
I do this for cycling clothing. (Almost all synthetic, some wool, fair amount of spandex-y stuff.)
This stuff lasts forever when washed on cool/warm and then hung.
Friends of mine have complained about one brand or another not lasting very long, but they've been tossing the stuff in the drier.
We're fortunate to have a basement with a nice beam I can place hangers on (for winter drying), or a hanging bar I fitted in the garage (for summer). Lately I've been getting rid of 8-10 year old stuff that I no longer like or no longer fits, and it's sellable, as opposed to just worn out.
(Doesn't sell for much, but folks will happily pay $20 - $30 for special print cycling jerseys that are still in good shape and cost $80-130 new. Way better than tossing them in the trash.)
If hang-drying outside, turn your stuff inside out.
UV from the sun kills bacteria, and you want that on the side near your skin, and as a bonus the inside fades (UV again) but the outside doesn't. I have some t-shirts that are quite faded on the inside but still reasonable on the outside.
I assume you mean 40 C, not 40 F? I started washing everything on cold (my washer actually has a 'Tap Cold' setting - just tap water) and it works just as well. I encourage everyone to just try it once - it won't hurt anything and you can always re-run the load - and you will never go back. Also, you don't have to sort clothes.
I read in some credible, non-technical publication, I think the NYT or WSJ, an interview with a engineer in that field (something like detergents or washing machines) who said that detergents used to need heat to enhance the chemical reaction, but that it's no longer true and cold water works just as well.
EDIT: Does anyone know a good technical, authoritative resource on laundry? Consumer Reports has well-researched info, but not in the depth I'd like.
> When a family member is sick, use hot water mixed with chlorine bleach to reduce bacteria in the bed linens and towels. The same goes for cleaning dirty cloth diapers, or other messes.
But also,
> Heating water accounts for about 90 percent of the energy needed to run a washer
So when different clothes prescribe washing at different temperatures, it's because the ‘activation’ temperature of detergent changes depending on whether you use it with jeans or underwear?
As far as I know the clothes don’t prescribe a temperature to be washed at, but rather a maximum temperature threshold upto which the material can withstand without risking damage to itself.
So a 40°C cloth can be washed at any lower temp but might deform or loose color or even breakdown if washed warmer than that.
Instructions haven't kept up w/ washing machines and detergents.
> "Front-loaders and high-efficiency top-loaders run normal cycles 10 percent cooler than agitator washers, and the 'warm' wash temperature in the U.S. has declined by 15 degrees over the past 15 years," says Tracey Long, communications manager for P&G's fabric care products in North America. “Traditional detergent enzymes can be sluggish in cold water so we worked to create a mix of surfactants and enzymes that deliver cleaning performance in cold water across all product lines," says Long.
> Consumer Reports’ past tests found detergents have gotten much better at putting enzymes to work in removing dirt and stains at lower water temperatures, and are less effective at higher temperatures.
If you want to sanitize anything, just use some bleach. It's harsh on fabric, but so is hot water, and the bleach will do a much better job. (FYI, I don't have any whites at all.)
My understanding that was for the special fabrics and/or the dyes used. For example raw indigo bleeds a whole lot more with higher heat. I know there are some fabrics and blends that are fragile compared to something like a cotton tshirt and recommend colder water.
A tip we rediscovered a few years ago(it's well over a century old) is that adding a bit of borax to the washer not only helps get your laundry cleaner, but it will make all your laundry smell better, too, since borax/boron is an extremely powerful agent for killing bacteria, molds, and fungus. My wife is sold, and almost refuses to do the wash w/o borax anymore!
(This might even help deal with the continually scummy front-loader problem, but I can't speak to that as we prefer our 33-year-old Kenmore top loaders that can still actually be repaired rather than replaced with expensive new Chinese/Korean crap every few years. Mechanical timer controls and durable design and mfg FTW! For what it's worth, our total cost of purchase and repairs over 33 years is maybe $1200.00 for the washer/dryer set.)
Or maybe a broken thermostat heater. Apparently this is a separate component sometimes.
What on earth is thermostat heater, you ask? Surprisingly, many dryers apparently have a simple fixed-temperature thermostat, and in order to make lower settings work, a heating element tricks that thermostat into perceiving a higher temperature.
If that heating element doesn't do its job, then the dryer acts like it's always on the highest setting.
> The thermostat heater is often located within the cycling thermostat. However, it may sometimes be a separate component mounted to the dryer's cycling thermostat. Depending on the dryer's temperature setting, more or less voltage is supplied to this heater. Low settings supply more voltage and create more heat, while medium settings supply slightly less voltage, generating less heat. High heat settings will not energize the thermostat heater at all. In this way the thermostat is tricked into thinking that the dryer is hotter than it actually is, so it opens at a lower drum temperature.
That's not true though. The additional energy to create that heat doesn't have to equal the time saved.
You'll notice this in heat pump dryers. They cannot generate the same amount of heat. They take way longer to dry the clothes. But they're way more energy efficient than other forms of dryers.
Edit: I thought of another example. Heating your home with hot water running through radiators. It's significantly more energy efficient to reduce the temperature of the water. This outweighs the additional time it takes to heat up your home. There are various drawbacks and considerations though, e.g. if the house has terrible insulation (noticeable draft) then it'll not be beneficial. There's various other things that'll significantly reduce energy usage, this while anyone would assume that generating heat is already very efficient.
To some extent drying clothes is generating heat (evaporation heat). If you're clever about it you might be able to avoid heating the (wet) clothes and rest of the contents of the dryer (or the outside!) too much. However evaporating water requires an incredible amount of energy, even if you just boil water away then most of the energy is still spent evaporating the water rather than heating the water, so it's not really too clear-cut that running a dryer hot is massively inefficient.
Edit: Also it's not that using lower-temperature water to convey heat is somehow more efficient, the thing with heat pumps is that they are more efficient at heating things to a lower temperature. If you're burning gas it doesn't really matter either way, you just get the energy out you put in.
> so it's not really too clear-cut that running a dryer hot is massively inefficient.
My heat pump dryer came with an energy estimate for various functions and loads. The various functions which shorten the time, or the functions which increase the heat (often related) are specified to use way more energy. To me, it's pretty clear, plus the manufacturer specifies it.
> Also it's not that using lower-temperature water to convey heat is somehow more efficient. [..] If you're burning gas it doesn't really matter either way, you just get the energy out you put in.
That's what I used to assume as well. It isn't accurate though. If the water that comes back to the heating element is too hot it'll not be as efficient as when the temperate is lower. Similarly, the additional energy that's needed to heat the water to e.g. 75+ degrees Celsius is wasteful. You can save around 30% of the energy by reducing the temperature of the water that's used to heat your home (though might not work due to various considerations). There are loads of other things that are possible which also significantly reduce the energy usage.
Regarding how to save energy when using a boiler there's a huge Dutch topic about it with loads of tips: https://gathering.tweakers.net/forum/list_messages/2027810. I assume similar information can be found in other languages, though heating using gas and water is really popular in NL (more so than any other country I assume).
> energy needed to heat water to 75+℃ is wasteful. You can save around 30% by reducing water temperature to heat your home
That depends. For resistive electric it shouldn't make a difference, pretty much all heat is transported.
For non-condensing gas (or wood etc.), if your heater is going full blast and a lot of the heat goes up your chimney and lowering the temperature makes a smaller, slower flame, that gets absorbed better, I think you could get 10-30% difference. The heating of water itself to 20℃, 75℃ or 110℃ shouldn't make much of difference, as you're not supposed to cool the effluents too much, or you get condensation, acids, rust ... which will likely kill you equipment.
Condensing gas is cool, extracting so much heat, that water condenses, but the gas must be clean enough and the condenser resistant to corrosion. Here, lowering water temperature can safely lower the effluent gas temperature for more heat extraction (even in optimal power range), and condensation of resulting water vapor from burning gas is about 10% extra energy that would otherwise go up the chimney. I'd expect about 15-30% more heat than non-condensing, especially if run on lower temperatures.
Heat pumps are quite efficient at moving heat, where 1W of electricity can move 3W of heat for a 4W heating yield. A steeper gradient means more work, so pumping heat from 20℃ to 75℃, 1W may only move 0.5W of heat for 1.5W yield (numbers not accurate). Lowering the temperature can make a 2x difference, or even more in extreme cases.
I wouldn't be too trusting of the claims of a manufacturer who's main selling point is the savings in energy...
They might still be true though, but if you keep in mind that it takes about 5 times more energy to evaporate water than to heat it to 100C, and that heating water is more difficult than most other substances it is really not clear why using more heat would be (far) less efficient. Sure it would consume heat at a higher rate, but also less long.
I'm not sure if heating with cooler water is more efficient.
Some places you pay for the joules delivered into your home. You have flow meter and temperature meters on he input and the output of the radiators and the price for joule is constant regardless of input and output temperatures.
What saves you money is keeping your interior cooler because heat loss is propotional to the temaperature difference.
Dutch energy companies by law have to advise their customers how to save money. The app I use give exactly this advice (lower the temperature), plus various other advices.
> Some places you pay for the joules delivered into your home
That's something different than what I said, no? I'm talking about when you generate the heat in your home. I'm aware of that solution as well, they're efficient because of volume plus part of the heat (energy required) is waste-heat from some industry.
There's still various ways to save energy despite exactly measuring the temperature out and in. E.g. radiator fans.
I know this all seems entirely illogical. Energy in (or required) should stay the same. Practically though, it's probably energy losses that somehow occur and are avoided.
E.g. for the radiator fans people measured if they save energy. They do, though the cost of buying them might outweigh the savings. DIY is cheap though.
> The additional energy to create that heat doesn't have to equal the time saved.
Right, it doesn't have to, but it's also possible that more heat makes it take proportionally less time (or close enough, with negligible decrease in efficiency).
Obviously, yes, using a heat pump will use less energy than a resistive heating element. But the question is more about how much and how quickly heat is input (regardless of how it was generated) and how that affects drying times.
Yes, but it's not a simple linear use of energy. For example, it might use 10x energy to dry twice as fast. That's a gain if you're in a hurry, but not so much if you're relaxing at home, on a tight budget, and/or have unusually high cost of electricity.
It might, but does it actually? Or does it use 2.05x energy to dry twice as fast, making the energy use difference negligible?
Edit: Consider also that the shorter you run the dryer for, the shorter you are running the (substantial) motor and fan, as well as less time spent heating the shell of the dryer and the air surrounding it.
Depends, for most dryers the temperature is limited because water evaporation is taking all the energy. The motor takes the same energy per time, so twice as fast actually uses less energy. However there is a limit to this, eventually (the end of the cycle) you reach the point where water isn't evaporating fast enough to use up all the input energy and temperatures go up to heating clothing fibers to no useful purpose and this is wasteful.
Also, is the dryer located in a climate-controlled part of your home? If so, the air that it exhausts will be made up in equal volume by outdoor air pulled into your living space. How much extra energy does that make your heater or AC use?
From an industry pdf I stumbled across, it looked like moisture sensing improvements were the best bet to save the most energy. Though, I didn't see anything about comparing heat settings in that doc, which may be telling.
Until near the end of the cycle your dryer is putting all the energy into evaporating water, so the temperature inside the dryer is actually fairly cool. Right at the end things change as the remaining water isn't enough to counteract all the energy being put in and so you heat the clothing to no purpose. So at the end off the cycle you should either shut off with a little moisture in the clothing, or regulate the temperature so that the heat input is balanced by the water evaporation.
Sounds like we need ultrasonic no-heat clothes dryers to be commercialised. The technology is there - just play music to the water molecules, and they dance!
The article compares the energy savings over an already inefficient dryer. The article mentions that an existing dryer takes 50min (average). Such dryers are not what anyone should buy, they waste too much energy. Over time it's cheaper to buy a heat pump dryer. Those easily take 2.5 hours to dry. They're significantly cheaper over an e.g. 5-10 year period than buying a cheaper and way more inefficient dryer.
The links to more detail with:
> The goal of this project is to develop a clothes dryer prototype, using ultrasonic transducers, with an EF above 10 lb/kWh.
But also:
> DOE’s Building Technologies Office is seeking new clothes dryer technologies that can increase the energy factor (EF) from 3.7 to 5.43 lb/kWh
The link in the article and my link shows that the intend is to go way over 10 lb/kWh. The link I found showed it could be around 20 or even 44 lb/kWh (seems to depend on the frequency used). This while being way quicker than anything else, especially heat pump dryers.
Yup, I’ve never once in my life paid any attention to any of this, and I only ruined one sweater once. My wife was not happy, it was a brand new cashmere sweater from some brand name. Still though, if that’s my only screw up and amortized over a lifetime of not caring about this, still positive ROI for me. Twist ending: we saved the severely shrunken fancy sweater and now it fits my kids, so not a total loss.
Well, my approach is to limit my day-to-day clothes buying to just those washable in 40°C and machine dryable. Makes both the shopping and life overall so much easier.
Any women who still have to wear business attire interested in a similar approach should check out MM LaFleur. Well-designed, machine washable staple pieces at a (mostly) reasonable price. IIRC the company was founded by a young French woman who used to work in consulting and knows the pain of constantly needing to dry clean your clothes.
Yes, if you only buy things that you wash in 40 degree water and dry in a machine, you can just wash everything in 40 degree water and throw it in a machine.
My rule of thumb is basically just to exclude from the dryer anything stretchy, slippery, knitted, or lacy. With that stuff hung to dry, what’s left is all the plain cotton shirts and jeans that can take whatever you throw at them.
If you have dryer it nearly strilizes everything you just washed with hot air.
I'm washing t-shirts in 40 deg and drying them in my washing machine with built-in dryer.
They come out a bit damp to avoid creasing too much. I never had them smell even though I was just unloading dryer into a huge pile of damp clothes and leaving them like that for a day or two to dry out completely. I even forgot to take them out of the washing mashine and found out few days later. They were still damp but didn't smell. I washed and dried them again though to be on the safe side.
Most washers (especially top loading) in Japan don't support warming water so people wash with cold water, use hot bath water, or hopefully the house has hot water faucet for washer.
I've bought an expensive front loading washer-dryer with water heater (upto 60C) and heat pump dryer recently. Now I always wash with at least 15C water, even 15C, it's significant difference in cold winter situation.
It doesn't sterilize, but it does bind to viruses and bacteria so they can be washed away. It's the reason why you don't need antibacterial hand soap: you don't need to kill, washing off is enough.
The exact opposite is true. These symbols require (but do not have) translation on every piece of clothing, because almost no-one understands what they mean, and it's not currently possible to type them in to Google to identify their meaning.
On the other hand, if the text was just in a language that I did not speak, I can either type this into a translation app myself, or use the automatic translation camera on my phone to identify them.
And sure, I can apparently now do that for those washing symbols too, which is nice... but why is that information being conveyed in a "language" that almost no-one speaks? I'd be just as confused if every item of clothing I purchased had washing instructions in Lojban.
I fully understand why an attempt at standardised symbols was made, being able to communicate that information more compactly would be useful, but it very clearly failed. They're not intuitive enough to learn without third-party resources, and a very very tiny fraction of the population is willing to put in the effort to do so. Just use text.
Current state, I would have to figure out what to type into google to see what some symbol like ⧇ or △ means in laundry terms. Either better symbols I don't have to google, or text in a language that I don't speak...but can figure out. The ones with temperatures don't even have a degree mark.
It's strange that on a site like Hacker News people are so averse to learning new things. We seem to have no problem learning other kinds of symbol languages. These are standard symbols - you can search Google for "laundry symbols" to find a key
The audience for the symbols are people who do laundry for a living.
Symbols that you never use aren't very useful when all you really need to know is "do not dry", "do not bleach", "wash with like colors". It's really not that big of a deal to put a tag written in English and Spanish to the US, or in French to France, etc.
My girlfriend works in the fashion industry in Taiwan. Designers in the US make an artistic drawing of a dress. She converts it to a Gerber file using CAD software. The Gerbers get sent to factories in China or Vietnam, who make prototypes and send those to Taiwan for testing. When approved, thousands are ordered and shipped directly from the factories to Macy's or JCPenny in the US.
When artists use English, CAD designers use Traditional Chinese, and factory staff use Simplified Chinese or Vietnamese, and consumers use every language - there's a need for a standard symbol. You have a good point about dry cleaning services too!
The difficult part is that the people through this international supply chain can't type these symbols in emails, or search for them in databases. That's why a Unicode code point should be assigned.
That clothing design and production process sounds fascinating. A search on some of the terms doesn't seem to return anything useful... do you have any good reads or videos for an outsider to understand a typical workflow?
I just learned it from her directly, and was surprised when she mentioned Gerber files. I used those for PCB design, and it's the same file format that they use for clothes!
That general summary was great, and plenty of jumping-off points. And the knowing the terminology helps. Thanks.
I'm with you, it bears uncanny parallels to the PCB design workflow. Not just the Gerbers, but the component/fabric selection, simulation, design drawings, etc.
I would agree. The issue isn't that they are symbols, it is that the symbols are really hard to understand, other than hand wash where a hand is in a bucket of water the rest have little connection with the outside world.
> a pair of apps for iOS and android that can identify the symbols
Sounds like a failure of the point of those symbols. How about using words instead? Then you don't need an app to decode them. You can use google translate if you must.
The ultimate -parent- hack is to only buy colored clothes (no whites... And if you happen to get any just wash them with the colored clothes anyway) non shrinkable clothes. Dry on low - "damp dry" so they don't all get fried. One load every 5+ days. The thicker clothes hang around the edge of the clothes basket (sweat pants in our case).
Oh and I guess just by normal kid clothes (cotton/polyester) so they don't have special instructions.
I've been doing that for my own clothes. White t-shirts turn pink, but pink is a good look too so I don't really care. Maybe that means I'm still a kid :)
We do that but how on earth do you manage with only one load every 5 days? We are a family of 4, one teenager, one under 10, and we’re doing one or two loads of laundry every day.
pants can usually be reworn 2-3 times. occasionally a shirt can be reworn if it was only used for part of a day. this might require an unusually clean child. as a single adult, I do laundry on a two week cycle. and that's driven more by my inventory of season-appropriate clothing than the capacity of my washing machine.
Would you please clarify how doing laundry constantly allowed time for ML research? Is it just because you had maternity/paternity, or was there something about the laundry workflow that let you squeeze in this activity?
Standard Textile Care Symbols? There is an actual standard for that?
It's a serious question. As far as I can tell, going by my wardrobe, there isn't. It's always been one of the minor annoyances in my life - clothing and clothing-adjacent vendors put these icons on their products, and I have no first clue what they mean, because which family of signs is used seems to depend on some combination of country of origin, target market, and the position of Saturn relative to Jupiter.
In case someone still finds this thread: standards my ass. Guess what, in Poland we have different ISO 3758:2012 symbols. Compare [0] with [1]. As it turns out, some of the symbols have "A" and "B" variant, and we use the "B" one for some reason.
Also I couldn't figure out WTF is "drip drying", and how "drip line drying" and "drip flat drying" differ from "line drying" and "flat drying". The Internet is very unhelpful here - every single article I saw assumed you already know what it means. Fortunately, [1] contained a note that the methods of drying are defined in ISO 6330:2000, and as luck would have it, I saw a copy of ISO 6330:2000 fall out of the back of a truck. Taking a quick peek, I now know that "drip line/flat drying" means...
...just "line/flat drying", except you set your washing machine to not do the final spinning cycle.
> In case someone still finds this thread: standards my ass.
The standard is there. If people decide not to use it that's on them. Just because the US decides not to use the SI units, that says less about the metric system and more about the US.
Most countries in the EU (AFAICT) do use the ISO standard.
The fact that one country does not use it is not an argument that that standard is not a standard.
> The fact that one country does not use it is not an argument that that standard is not a standard.
My point was slightly different. Poland does use the standard, the same standard. It just so happens that the standard has "B" variants for several icons.
I guess it's sort of an improvement that I only need to know 1.5 sets of icons instead of 3-5 different ones.
Having different displays for different markets is exactly what Unicode allows for: you specify the intended symbol, and then it's localised in the font you choose. An example from the proposal is "flat dry when wet, European variant has two bars, US and Canadian variant has three bars".
A glyph in one font may look confusingly similar to a completely different glyph in another font (or the same font). If you don't have a digital version you then have to rely on context clues. You can quite reliably tell if a given C is LATIN CAPITAL LETTER C or CYRILLIC CAPITAL LETTER ES, by assessing the surrounding text and the medium to see if the author was most likely writing for a Russian or an English audience.
Of course, it's possible that manufacturers currently mix contexts confusingly. In that case, having each one draw from a single Unicode palette should help.
This whole "is it Cyrillic or Latin" confusion sometimes does happen IRL. There's an insurance company that has its logo in all capitals: PECO. For most of my life I've always read it as Latin. But at some point I heard it in a TV commercial and they said it as "reso". Yikes.
Also РОСНАНО. It's not "pochaho", it's "rosnano".
The reverse is also true: there's a "Bona Capona" restaurant but its logo uses a typeface that makes it tempting to read it as "Vopa Saropa".
Well thankfully the rest of the world aren't subjected to the Cyrillic-Latin confusion, because it's only inside the country that Russian companies use the mix of Cyrillic and Latin... And specifically English, while a lot of people over twenty-five can't pronounce or read English.
Ah, well, I should've written ‘Russian-Latin confusion’—forgot about the alphabet changeovers. But at least I'm told that:
> It can be written in Serbian Cyrillic or Gaj's Latin alphabet, whose thirty letters mutually map one-to-one, and the orthography is highly phonemic in all standards.
Though digraphs spoil the clarity somewhat.
Which reminded me that I should check what Kazakhstan is doing to its language. They seem to at least have avoided the eldritch monstrosity of Polish with ‘Grzegorz Brzęczyszczykiewicz’ and the like (https://www.youtube.com/watch?v=AfKZclMWS1U). But they are still setting themselves up for English-language people pretending they can pronounce anything Latin-looking—despite English itself demanding you memorize the dictionary twice, for speaking and writing. So Kazakhstan will soon have its own great writer Karel Kaypek.
Are fonts consistently localised? What happens if the pool of people who use a font (eg Arial) contain multiple washing symbol regions? Do you spin up a new, localised font for every country?
No, you simply rely on your tooling, which should allow you to set the locale (or "script", which is the OpenType term for it) for your content.
Thankfully, a lot of tooling designed to present text to a human (From Word to LaTeX to HTML) already have ways to do this, so folks writing textile documentation will now be able to take advantage of this, too.
Where are you from? Serious question, since Europe do tend to have standard symbols for nearly everything due to cross-border (and cross-cultural) differences, while this is meh (why? we have English!) to North America.
I believe this regional symbols vs. text difference can also be seen in many washing machines - it is quite common to have symbols for the knobs and settings here (e.g. https://imgur.com/a/7yRzT), while I've understood that this isn't common in e.g. US.
Though I thought textile care symbols are common in US as well, or at least I think all my US-bought clothes have them too. So this is probably not one of those US-Europe differences.
I remember doing some Internet searches for this over a decade ago, back when I started doing my own laundry :). Maybe I didn't look thoroughly enough - but between that search and the actual labels on the clothes I had, I saw no indication of a standard.
Might be because my clothes were probably a mix of branded things sold on EU market, and (mostly) unbranded ones I bought on a bazaar, which could've been imported from Turkey or East Asia.
Makes sense since it was over a decade ago. US and Asian textile care symbols were a bad mess (differs from brand to brand) back then (while ISO sensibly implemented the then-EN standard because they were the same across brands).
I think I have three different sets of textile care symbols in my closet depending on whether the piece of clothing was ordered from the US, Europe or Japan. It's a handful.
Not a bad idea, but there’s dozens of other ISO symbols that I think should also be considered[0].
Some are quite important; like safety symbols.
One of my pet peeves has always been the difference between US emergency egress (“EXIT” signs), and everyone else in the world. In the US, they tend to be red, and the word “EXIT”. Everyone else does green, with the man running through the door.
In Japan, medical facilities were denoted by a green cross. In the US, it tends to be blue or red.
Color can be an important factor, but that’s tangential to the Unicode set.
The US takes English for granted so often, despite our actual diversity and no "official" language. Although I do love the National Park standardized symbols.
But to totally wander down your tangent, color can be dangerous to rely on for accessibility reasons.
Green in particular, seems like a poor color choice for accessibility, given the prominence of red/green colorblindness.
I worked on an internal system with a Web UI once with someone much less sensitive to this than I was. He wasn't a jerk, just would do things and didn't like revisiting to change. I'm generally not a UI person and got tired of fussing about it, so just rolled with his "it's just internal" for his parts.
One of the things he made would highlight rows of text by changing the color from grey to red without any other visual indicator (bold, etc.). The grey text already annoyed me, but that's a different story.
Our very first user when we piloted just happened to be red/green colorblind and was completely confused. It didn't occur to him that color was the issue, and my coworker didn't understand why he couldn't differentiate. It dawned on me to tell him that the selections were red, and ask if he was colorblind. We walked through the rest of it with him and found another place where red and green text were used to indicate "good" and "bad".
A week later, my coworker took colorblind-support as a 'requirement' and had reworked everything he did to ensure multiple visual cues. When he did decide to fix something, he was very thorough. And I'm sure he carried this forward to everything he's done since.
I'm not an expert, but I think the red/green issue is the most common (especially for men), but I know there are other varieties, and other vision problems can make color difficult.
I think that the figure running through the door is a great example of a good symbol, though. No reliance on language and an easily identifiable shape. But I'm definitely not an expert in either accessibility or design. I just try my best at both when I have to do that stuff.
Veterinarians don't seem to have any color code, that I know of. Many don't use crosses (they tend to have cutsey animals).
The hospital near where I live has a blue cross at its emergency entrance, and a nearby urgent care clinic also has a blue cross, however, another nearby urgent care facility has a red cross.
I don't know of any pharmacies that have any prominent cross displays. They tend to have their own branding (which may or may not have a cross integrated).
Interesting, as here (France and Belgium) only pharmacies consistently display a green cross. All of them do, and I wouldn't be surprised if that was a legal requirement. It's not as consistent for the other medical facilities, but I'm pretty sure that at the minimum they never use the "wrong" colour.
I don't know where this usage comes from (and can't really look it up now) but now that I think of it "the Red Cross" is a doctors organisation, so it fits the scheme I told about.
For instance, while messing around with setting up the i3/Sway bar, I couldn't find symbols for:
* WiFi. Really. There's U+1F4F6, but it's not exactly the right thing.
* Anything to indicate CPU usage, like some sort of microprocessor chip.
* Anything to indicate RAM usage.
This is curious because for instance U+1F50A and U+1F507 exist, so you'd expect to have a set of such generally useful icons. Plus, it's computer tech. The lack of computer related stuff is odd, given that there's no lack of many things I have no idea what they're good for, such as U+1F574 ("Man in business suit levitating")
Edit: would be nice to have unicode support on here. Seems it gets stripped.
You are misunderstanding the reasoning behind symbol inclusion in Unicode.
It’s not about “what could be handy” but about what is needed to cover character encoding already in broad usage.
Japan was early adopter of cellphone internet. One way to save data usage was to compose graphics as characters in a proprietary encoding (specific to the carrier!)
Basically, in order to move Japanese handsets of custom encodings and onto Unicode compatible systems, Unicode needs to be able to display not just all the Japanese characters but also the weird graphics that carriers decided to come up with in order to decorate their online services. This is also why there is so much Japanese food characters.
Emoji existed since late 90s[1], just it was completely domestic to Japanese carrier intranets and gatekeeped hard that there aren't a lot of records remaining in the public Internet.
Japanese feature phone market was weaponizing lack of emoji support in modern smartphones to chase them out, but iPhone happened anyway and Apple started incorporating their own fragmenting SoftBank-iPhone-specific implementation into iOS so Google pushed it all to Unicode to fix them all up. They wanted to capture user emails into Gmail, and all emojis sent from featurephones being replaced with = by gateways was problematic.
>Emoji existed since late 90s[1], just it was completely domestic to Japanese carrier intranets and gatekeeped hard that there aren't a lot of records remaining in the public Internet.
Emoji yes, the new emoji we get in annual Unicode updates, not...
The process for inclusion of most modern emoji is the exception, not the rule. That exceptional process isn’t one you can get in on as an individual, either. It goes like this:
1. Some large-userbase OS+messaging services provider, like Apple or Microsoft, sends a “proposal” for some new codepoints to the Unicode Consortium. The explicit message is “do you think these are worth standardizing?” But the implicit message is more unilateral: “we’re giving you advance notice, that we’re planning on including this emoji in our fonts + supporting it in our messengers, whether you like it or not, and whether it’s inter-compatible with anyone else’s systems or not. If you don’t standardize it, we’ll encode it using the Private-Use Area.”
2. Given the Unicode Consortium’s goal of never having proprietary text bits flying around the Internet — and given these big service providers’ histories of running messaging services (e.g. MSN Messenger) that did use proprietary encodings for emoji, resulting in some lasting problems with digital archaeology — the Unicode Consortium feel pressured by this “proposal” to standardize the proposed emoji codepoints, whether they really think they’re “worth” being standardized or not.
3. The Consortium also then feel pressured to get a new standard revision out quickly, to get ahead of the planned usage by these service-providers (because that usage would have to be encoded somehow, and if they don’t give the service-provider a codepoint to use by the time they ship their new font version that includes the relevant emoji, they’ll just have to make up their own.) In most cases, the Unicode Consortium is reactive to existing usage, allowing them to judge whether there is real existing adoption of a symbol; but here, they have to be proactive, pushing out a standard that includes a codepoint before knowing whether anyone will use it.
As a result, you’ll see these large dumps of new emoji where Apple/Microsoft/Google/etc. just decided autocratically that the world needed some more emoji, and Unicode begrudgingly followed along. It’s like what happened when Unicode first absorbed the emoji codepoints of Japanese feature-phones; but happening just-in-time, one at a time. (If those Japanese feature-phone manufacturers kept introducing new emoji, they’d have had to absorb those just-in-time too.)
There’s a separate, much more sensible and deliberate process the Unicode Consortium goes through you’re a regular-sized actor who’s not attempting to strong-arm them, where you do tend to have to prove existing use in either analogue documents, or as part of some proprietary digital text format.
You have summarized the problem of the Unicode emoji process so well, but I want to say that it is not the one-directional pressure. Vendors do not want to implement too many new emojis, so the number of new emojis per year is essentially limited (50--100 [1]). Vendors also do not want that other vendors don't implement their new emojis, so they do cooperate with other vendors and obey the rules set by the Consortium (e.g. no trademarks). It's not comparable to individuals, but it is not exactly possible for vendors to put any emoji to the standard as well.
Emoji is the sole exception to this rule and then you still need to convince the committee that separate WiFi, CPU or RAM emojis are needed to communicate the intent (say, for example, given there are other computer-related emojis). In fact I do think you can, given they were probably never suggested before [1] except for "wireless" which status is "Prioritization Pending".
Emoji are for communicating to other people. I can’t imagine why you would ever send someone a message with a ram stick in it. OP sounds like they want the icon for a task bar. In that case they can simply import any SVG they want.
> I can’t imagine why you would ever send someone a message with a ram stick in it.
Would you consider this to be a very different thing from, say, a MiniDisc? Yet there is an Emoji for that.
Obviously we cannot include every picture possible in Unicode, but there certainly seems to be demand for a lot of them. Besides really enjoying their use in texting, I've also come across a lot of professional uses as well.
Yes, this means that the Unicode Consortium probably has needed to adapt their original mission a bit, but I don't see the harm in that.
Would you consider this to be a very different thing from, say, a MiniDisc? Yet there is an Emoji for that.
MiniDisc makes sense for two reasons. First, because it's a legacy inclusion. Second, because absolutely, you might use it in a text message. "Don't forget to bring over your bitchin collection of [MiniDisc] on Thursday!"
Yeah the OP's arguments wouldn't work: the typical answer would be "use PUA and a custom font" as Powerline symbols do (incidentally, they have been proposed to Unicode [1]). But as a general emoji CPU and RAM can be actually substantially different from any other existing emoji, which is the main concern of the Unicode emoji process.
This is an important point that most technical people miss. They want to turn Unicode into a cross-platform FontAwesome. That's not what it's for.
There are a lot of specific-use symbols in Unicode (chess pieces, for example), but those are legacy inclusions because many of those symbols were included in computers before the internet.
Chess pieces aren't included just because of legacy computer software, chess pieces and many related symbols in Unicode come from "there are hundreds of years of books that use these symbols and encoding those books needs these symbols".
I think that's also a distinction often missing in technical people's assumptions about Unicode and why it isn't just a cross-platform "FontAwesome" even just of legacy proto-FontAwesomes like Wingdings (which is also included and is its own different story). Unicode Consortium likes proposals to include things such as scanned documents of "here's how this 1850s book used chess symbols in the flow of text to communicate how the game is played". Not as adornments or images or separate figures, but directly as a part of the text.
That was one of the things that the Power Symbol Proposal [1] that was heavily discussed on HN in the past (and sort of spun out of HN comments in the first place) wound up learning and realizing how big that was to Unicode Consortium's needs in a proposal. "How was this used in real examples in the flow of text?" Finding and being able to cite and scan real world examples from text books and help documentation is an important part of the process. Even the chess pieces were about communicating to other people, in the flow of text in historic books and magazine articles and discussions.
Yeah, I’m willing to concede that there are multiple reasons to include things in Unicode and some things are included for utility rather than because it exist.
Things like half stars was added by petition similar to TFA, so my post should have been more nuanced.
The point is that the reason why there is a seemingly useless emoji and not the [insert character you really want] is usually because one is preexisting and the other isn’t.
I believe unicodes mission is to include all characters expressable in all writing systems “in the real world”, and being a replacement for major text encoding is seen as part of accomplishing that mission.
The alternative is to just not have Egyptian hieroglyphs in Unicode. The phallus is a frequently used symbol in Middle Egyptian. Then again, what's there is half-assed. There are no joining characters, so nothing displays right. And while you'll probably have the characters you need most of the time if you're dealing with Old Egyptian or Late Egyptian texts, Serapis help you if you're working with Ptolemaic-period texts.
At least the situation is better than with cuneiform, where Unicode alone can't represent a text because the characters drastically vary over time and you need to make sure you're using the right font.
The fact, however, is that Unicode is not sufficient for scholars who use Egyptian hieroglyphs or cuneiform. For the former, most people seem to use JSesh. For the latter, line drawings + transliterations.
I wonder if academical papers of today won't be taken seriously in the future because some of the symbols we use will have turned into a cultural faux pas? I hope not.
> such as U+1F574 ("Man in business suit levitating")
It looks like this character made it into Unicode due to the fact that it was part of the Webdings typeface. Given how widespread that was, many of its characters made their way into Unicode.
> This character was originally introduced into the Webdings font as an “exclamation mark in the style of the rude boy logo found on records by The Specials". This levitating man was known as Walt Jabsco.
Emojipedia is not just an encyclopedia, it is a major contributor to the Unicode emoji process (its editor, Jeremy Burge, is the vice-chair of the Unicode Emoji Subcommittee).
> According to Jen Sorenson ... the Man in Business Suit Levitating glyph in the Webdings font was intended to be an exclamation mark in the style of the rude boy logo found on records by The Specials published under the 2 Tone Records label. So perhaps the Unicode character would have been better named Rude Boy Exclamation Mark.
The reason seems to be that links are not a feature of a text document, they are part of a markup language, and therefore, it should be a feature of that language instead.
For HN, AFAIK, they deliberately strip emoji, but most Unicode work fine. ("私はひどい下痢です" should show up correctly)
I believe Cirth and Tengwar are just less prioritized, unlike Klingon which was explicitly rejected (due to the non-usage of Klingon speakers at that time). Maybe it just needs a re-submission of the 1997 proposal by Michael Everson.
Klingon has a much more complicated history than this. The usage of Klingon is higher than many of the obscure things encoded in Unicode, so that's not the full reason. The real reason is most likely that Paramount considers the Klingon language its intellectual property and has attempted to enforce it[1]. Thus, I believe the official position of the Unicode consortium is that Klingon is written in Latin script but with a Klingon font (ie is basically a simple substitution cipher with a handful of ligatures), which satisfies everybody except for those who want Klingon included in Unicode.
ETA: The proposal[2] has been revived recently, and in the committee meeting[3], they're clear that the blocker is legal.
Ah, thank you for pointers to the revived 2020 proposal (my radar to the L2 register was partially off at that time). I'm not sure that Klingon the script was in broader use at the time of the initial proposal though---there are several constructed scripts that enjoy uses comparable to Klingon's and I don't think they have been well received. Or alternatively, they all actually have a potential to be encoded but only Klingon speakers are enthusiastic enough to pursue.
I hit exactly the same issue (no wifi symbol in Unicode) when working on my hobby project of train timetables. Some trains offer on-board wifi and it is indicated in the timetable (along with other icons like bicycle transport available, accessibility, etc.). Sure, I can use SVG or custom icon font, but I am definitely doing "communication", not UI.
i'm not well-versed in the politics of unicode, but it seems like you're looking for use-specific iconography, not symbols that are part of language. there's no standard symbol for "bicycle transport available", or "wifi access available". the three curved lines radiating from a circle image adds some visual flair to your project and can be used as a shorthand if you make it clear what it means, but any symbol you could use there doesn't necessarily convey the same meaning if i use it in my project.
contrast that with the symbols in this proposal - the TEXTILE CARE SYMBOL WASHING TUB WITH SEVENTY DEGREES CELSIUS glyph means the same thing no matter where it's used.
I also discovered 11 Chinese characters for Hakka and Taiwanese, and wrote a proposal for getting those added. They were accepted, but are still waiting for the next batch of CJK Unified Ideographs to be released.
There are various Unicode blocks dedicated to symbols. There are cartography symbols, domino tiles, alchemical symbols... There is even a block for "legacy computing".
And Unicode in particular has loads of non-letter symbols in it. It's only natural, given that a lot of these often show up in line with regular letters. Think of various paper documents you read, like instruction manuals, which make heavy use of non-letter glyphs in text.
Yes, that's a main reason why Korea and Japan (used to) have their own laundry symbols (KS K 0021 and JIS L 0217 respectively). Japan recently switched to ISO 3758 to match with China, which had switched from its own symbol much earlier though.
The document you linked to says they are "protected under trademark law", not copyright.
I would guess they are "certification marks", which are a bit different from a normal trade mark like "Apple" or whatever.
Their web site says: "GINETEX IS AN INDEPENDENT NON-PROFIT ASSOCIATION UNDER FRENCH LAW". So I would hope the fees are not exorbitant. But who knows? It's an independent organisation. If it did turn evil, would the French government intervene?
At least in the USA: Symbols can’t be copyrighted, only their encoding (eg the svg or otf) may be. Visual styles and their likeness may not be copyrighted.
Symbols may be trademarked but that is specifically for logos and is a different concern (eg fair use rules differ, intention is to avoid confusion not preserve IP, etc).
(You can make an exact riff off any font but if you code it from scratch it is ok.)
They are not copyrighted, they are trademarked by a nonprofit called GINETEX that licenses them freely, but with agreements that mandate their correct use.
They are not likely to want to give an open license to the public via unicode that might result in incorrect labeling.
It seems like the Unicode Symposium has engaged in a massive employment program for itself with a mission that will continue into perpetuity.
Obviously ASCII was insufficient as it only really had Latin characters. Providing a standard interchange for the world's languages is a noble goal.
But then (IMHO) Unicode went completely off the rails by trying to create a code point for every imagine symbol, including emoji. And now I guess textile care symbols.
All the while, it's missing (at least as of 2015) key characters in living languages [1], engaged in the highly controversial Han Unification [2] (interestingly, there are Latin/Cyrilic and other duplicates that there is no attempt to "unify" [3]) and implemented Unicode code point modifiers.
Emoji, in particular, seems to be a huge mistake (IMHO). Like... it's going to be constantly changing. What's wrong with hierarchical approaches that we've used for things like DNS? Create a Unicode code point for "emoji" and then the next code point is a completely different standard.
The goal of creating a code point for everything just seems... wrong. We already have >1M code points. In 100 years at this rate we're going to have 1B+ code points where 99.99% of them are never used.
Why can't we just solve the problem of expressing written languages and keep the rest to a different space?
I'd argue that textile care symbols are part of "written language", just as mathematical symbols or typographical marks for example, which exist in Unicode.
Emoji were certainly written language prior to unicode, but their expansion is unquestionably political since there is no well-defined "natural expression" of emoji that don't exist yet. It doesn't help that their standard [0] on what qualifies for inclusion and what doesn't is very poor, especially this: "Already representable. Can the concept be represented by another emoji or sequence, even if the image is not exactly the same?"
Let's say I want an emoji for glass of water for example. There is already one for glass of milk: . Well, I could use or or maybe even by itself is good enough, so my proposal would not be accepted.
It is not true that every emoji is a single codepoint - for instance the flags are country code strings and there are many other cases. The original set were separate characters because they were imported from a different standard.
Nobody really cares about Han unification either. It's fine, especially since you can use variation selectors or specify the text language to get appropriate fonts.
Now I'm curious because Wikipedia mentions the Han group has "experts from North Korea"…
I guess this poses the question: Is there a clear boundary between text and icons? And in which cases should icons be handled like characters?
Plain symbols within text clearly belong in a font. Logos, on the other hand, are images, not characters. Icon fonts, which are somewhere in between, became popular for webdesign but seem to be falling out of favor again. What is the current "best practice" here?
No, there is no clear boundary between text and icons. What is included in Unicode and in what way is decided case by case. See for example the regional indicator symbols.[1] Always two of them indicate a particular country code according to ISO 3166-1.[2] They resolve into a sort of a logo: the flag of the country. (Guess what happens with rendering a legacy Unicode document, when a country changes its flag ...)
I'll just add a meta comment that `understood (or not) irregardless` is the most value-ambiguous utterance I've read in a long while; quite clear to me what you meant, but probably would be quite a 'head scratcher' for an AI.
There is no entity that can decree something an international language or otherwise, even if an entity were arrogant enough to claim that it could do so.
Many writing systems evolved from what we might consider icons. The letter A comes from the head of an ox, which becomes much more obvious if you put it upside down like ∀. Many Chinese characters retain a similarity to the objects they represent: "tree" is 木 and "fire" is 火 (imagine a bonfire). The Korean consonants ㄱ and ㄴ represent the general position and shape of your tongue when you make the "k" and "n" sounds, respectively. All of these symbols began their life as icons, but were later recognized as text as people converged on a certain usage.
I always thought that unicode opened pandora's box with emojis. It makes sense to represent those that already existed in previous encodings for compatibility (it is meant to be universal, after all), but beyond that it's just a minefield IMO. In particular a big annoyance with emojis is that their representation varies a lot from device to device or even from application to application. You can expect that a B or a Ф will be recognizable in any non-windings font, but many emojis look different enough to the point of changing pretty drastically their semantics.
I'm starting to think it would've made more sense to allow unicode to allow to embed small vector icons directly in the text format. Some kind of restricted SVG-like dialect that would be handled correctly by all unicode parsers and look the same everywhere.
This way anybody could design their own application specific icons and emojis, you could either allow free-form emojis or only whitelist site-approved variants. Think something like fontawesome, but it'd work anywhere you can put unicode text.
Sure, it would significantly increase the size of the text but is it usually a big problem? It would compress very well, and in my experience text storage is pretty negligible these days in most applications compared to images and videos.
Instead it seems that we need to allocate codepoints and ask people to create fonts for textile care symbols, and national dishes, and all sorts of sports, and...
I have a site that makes heavy use of Emojis and to have the same rendering on every device I use the Twemoji (A simple library that provides standard Unicode emoji support across all platforms) made by Twitter.
> I'm starting to think it would've made more sense to allow unicode to allow to embed small vector icons directly in the text format.
Unicode folks are also feeling this pain and there have been multiple attempts to decouple the emoji business from Unicode: the closest to what you've described is the Coded Hashes of Arbitrary Images [1]. (The proposed encoding is wasteful but of course would have been improved if it were accepted.) Considering that the Unicode consortium is a consortium of vendors, this emoji business is suboptimal even for vendors. But there seems no particularly good way to efficiently and securely implement such mechanisms.
Interesting. Apparently TR-52 allows for arbitrarily long strings of tag values for emoji variants. So the standard could, theoretically anyway, allow an emoji variant that includes an entire PNG or SVG.
Unicode is supposed to encode writing. These are not part of any writing system I'm aware of. If Unicode adds these, that steps over the line from being descriptive (ie. this is writing actually in use), to prescriptive (ie. these should be treated as writing).
Maybe that boat already sailed with emojis, though.
I think the main motivation for stuff like this is for people to feel good about having an impact on something like Unicode. I fail to see how this adds any value to Unicode.
This is more useful than some other recent additions, I'd say. It's somewhat surprising that this hasn't been included before, but maybe that's because this was standardized well before Unicode came to be.
There was an attempt to add these to unicode in the mid 00s, and it failed for the same reason this will: GINETEX won't grant the trademark license, because then they'd no longer be able to ensure that whoever has the rights to use the symbols is using them correctly.
For me it was always annoying that Unicode doesn't allow to code full set of subscripts/superscripts. It's so obvious and more necessary than many emojis they added recently.
Unicode subscripts and superscripts mainly exist because of IPA. They are not general ways to encode arbitrary sub/superscripts as it would require too many letters to be encoded.
Well that's actully what U+008B (PLD, Partial Line Down) and U+008C (PLU, Partial Line Up (though officially it's inexplicably Partial Line Backward)) were originally for:
PLD text PLU would produce a subscript and PLU text PLD would produce a superscript.
No they’re not. Superscripts and subscripts in English as used in literature may be a typesetting concern but that’s not their only purpose. Eg in chemical notations or in mathematical text it is not a matter of typesetting but an actual separate symbol with distinct semantics.
Italics change the semantic meaning of text too, but they're still rich text, not plaintext. Unicode has been generally taken the attitude that Unicode is for plaintext, not rich text, even if it's stretched the meaning of that a bit. (There's also the time they goofed up and added interlinear annotation, which is pretty clearly not plaintext, but they've since discouraged the use of that and I don't think they want to make any more mistakes like that.)
Sure, I could have chosen a more watertight example to demonstrate the point. (And I think that "although" is an important "although" -- those characters are not meant to be used for ordinary italics; if you do this, you're going to cause significant problems for anyone trying to process your text, such as by, say, searching it.) It doesn't have, I don't know, characters that are blue-colored bold italic underlined superscripts, you know? Whatever exceptions or seeming-exceptions Unicode may have made, they're not looking to make more of them.
Little related but saw a nice documentary(1) called "Have you ever heard of the "Emoji Commission"? | DW Documentary" where a lady became a passive member by buying a $20K ticket or something to get a dumpling emoji added (IIRC it was sometime ago).
If you have time to kill, it's super interesting on how they decide which emojis to add.
Of course that limitation is artificial, due to compatibility with UTF-16 and UCS-2. So Microsoft would likely block such a proposal, since they'd have to change how Windows does character encoding.
I wonder what the purpose is of adding more and more symbols to a standard that was originally intended to represent texts in written languages. It almost feels like it is the easy part of 'improving' the standard: Adding more symbols, while at the same time there are still many problems with missing characters for existing character sets. https://www.w3.org/International/articles/missing-char-glyph...
The document was last updated in 2003. Nowadays Unicode provides a solution to the gaiji problem: the Ideographic Variation Database [1]. For unencoded Han characters using Ideographic Description Sequences as ligatures is now quite viable.
Related : I've started to draw new emojis depicting the environment and energy subjects.
As I don't think they would make it in the unicode standards soon, I've found the openmoji project very nice, it lets developers and designers use meanwhile.
> Here's a test: how many of those icons can you correctly identify without reading the accompanying explanation?
Well, funny thing --- the explanations on TFA are all in a language I can't read, so the words that "work just fine" mean less to me than a lot of those symbols do. But, the answer to your question is "35".
> Good luck googling a scribble.
Reverse image search? Google in fact provides just such a service. Realistically though, this is how you google a laundry symbol:
You seem quite upset about this subject, but I'm not really seeing what's got you up in arms. These symbols have been used for ages now. Check any laundry garment you own. Having them included in Unicode (rather than forcing the use of some specialized font) isn't going to make them suddenly show up more places in physical print (and, conversely, keeping them out of Unicode will not make them go away). On the other hand, if you're encountering them digitally (as Unicode), you can just copy/paste the glyph to search for it. It's literally implemented as text at that point.
> Words can be looked up in a dictionary
Why on earth is is acceptable to pull a dictionary off the shelf and page through it (or load up dictionary software, etc.), but not okay to consult a laundry reference (digital or otherwise; if digital, almost certainly found through Google)?
> You seem quite upset about this subject, but I'm not really seeing what's got you up in arms.
Because icons make things pointlessly harder to use. Instead of teaching children 26 letters, do you want to burden them with learning thousands of "standard" icons in order to read?
Are you really helping tribes that have had no contact with the rest of the world?
As for Unicode, a boatload of these things gets invented and added to Unicode every year. Where does this stop? Do we really need an "alphabet" with a million symbols in it?
Good luck googling Chinese text? Symbols don't require localization, and they're a very small and simple subset compared to a whole language. Even illiterate can learn them, and literacy rates can go down as low as 30% in some countries.
So use English. Even illiterates can learn 5 or 6 English words as well or better than they can learn icons. These words also transfer to learning English in general, while laundry icons are worthless knowledge.
Unicode is a secret moat for the large tech companies. The more crap they add, the harder it is for anyone who starts fresh to catch up. For every new emoticon, that teenagers expect to have, the chance they have more platform choices diminishes. Standards like these should not exist, unless the standard body also provides a free conformant reference implementation.
Unicode is a vast standard and have many separate parts. Unicode the character set is a significant but miniscule part of it. You don't need much to support Unicode if you don't do display (or delegate the display to someone else, like web browsers). The conformance requirements are also loosely worded that you can declare many parts of them as unsupported and move on [1].
Unicode also packages an essential but (in hindsight) complex concept and process like "what is the word boundary" or "what are characters usable as a part of identifiers" into a neat algorithm and data table. The algorithm is more complex than the data, which is readily available for you. Assigning a new character to Unicode mostly means a change to that data, not to the algorithm. If you consider one should be able to display every assigned character as expected to be fully conforming to Unicode---I stress it's not true, but if we assume so---then there would be no conforming implementation of Unicode at all. All Unicode implementations just implement what they need to support.
[1] For example, the requirement C12 (https://www.unicode.org/versions/Unicode13.0.0/ch03.pdf#page...) specifies that you need to implement the Unicode Bidirectional algorithm only if you support right-to-left characters. In the other words you don't have to implement the Bidi algorithm for conformance.
(Not knowing your startup ill make loads of assumptions here, sorry if I get it wrong)
If you are drawing Unicode characters that include a full range of Emoticons, you are most likely using a platform API provided by Apple/Google/Microsoft or similar. You, like most developers, most likely live in "Big-tech land", so you get this for "free".
If you want to port your application to an open source platform or a new upstart platform, where this isn't available, they may not have the resources to draw thousands of Emoticons, so its harder for them to attract you as a developer.
Lets say someone is writing a SMS app for something like a PinePhone. SMS is open enough protocol that this is possible. The UI for a chat app is something you can build in fairly short order. There are plenty of open fonts, you can use, but there are no open fonts that have thousands of detailed and sometimes animated Emoticons. Without Emoticons very few people are going to want to use a PinePhone as their main communications device. The amount of effort needed to implement a SMS chat vs a SMS chat + draw all emoticons people use is massive.
Unicode, and SMS are both open, but given the size of the Unicode Spec, its effectively shutting out any small player form implementing it. Big tech gets to claim they use open standards, but at the same time make sure no one can threaten them.
Its a bit like when Microsoft tried to make a standard for documents so complicated that only word could implement it.
You seems unaware of why emoji got accepted to Unicode in the first place. Emoji was originated from Japan's three big telcos, which had different sets and encodings of emojis (e.g. some used Shift_JIS extension, some used embedded images via HTML, some used SI/SO sequences). It was already impossible to write a SMS app without accounting for this peculiarity. Unicode thus made this situation much easier, not harder.
> If you want to port your application to an open source platform or a new upstart platform, where this isn't available, they may not have the resources to draw thousands of Emoticons, so its harder for them to attract you as a developer.
Are there not any open-source fonts that support the full Unicode spec?
You only need to implement glyphs people are actually using. If people are using [some glyphs] from the Unicode spec in your app, yes of course you need to use a font that implements them. How else could it possibly work? What would your alternative be?
I think a better solution would have been to only have glyph in Unicode, and then any emoticons would have been implemented by sending embedded images in various chat protocols. Its much easier for a receiver to parse a common image format then to make a font that includes all emoticons. It would also allow users to add their own emoticons, and you wouldn't have any ambiguity of what you are sending. Some facial expression could be argued dont express quite the same thing in all fonts, so there may be something lost in translation between users of different platforms.
Glad that the language is something I cannot read. It was fun trying to interpret the icons and coming to a realization that either I or the images are too obtuse.
What does an extra line added to an already existing icon mean?
A font is just[0] a table that maps numbers to pictures. A typical font will map a number 65 to a picture of what we know as letter 'A' in Latin alphabet. But there's nothing preventing you from making a font that will contain, at index 65, a picture of a dog. Or a "hand washing" symbol. And this is what people did - both to support non-English alphabets in the era where "printable character" meant "7 bits", and to support mixing graphics with text (if you add that missing bit to make a full byte, you get 128 more codes to play with).
Ultimately, screens and printers don't understand letters - they understand arrays of color values (or, occasionally, colors attached to parameters of a curve). Fonts are the intermediary, and you can draw anything with text if you supply the right font. The problem is though, if you had a custom mapping of characters to pictures, you had to ship it with your text, and make sure users know how to configure their software to use your specific font for your text.
Unicode exists to allow every number describing a character to have a fixed, standardized meaning, so that you can send text around independently of the fonts. Instead of fighting for the same 256 codes, there are now over a million possible codes, and fonts declare which code ranges they know how to draw. As a result, you don't have to worry about mixing alphabets and graphics in text anymore - worst that can happen is that something will get rendered as a placeholder. But it won't get rendered as a wrong symbol.
--
[0] - Well, strongly simplifying. In reality, the "picture" can be anything from a bitmap, a set of bitmaps, a vector image, or executable code; the mapping part can also be executable code. A prime example of "code is data" :). Simple devices like thermal printers might accept only a trivial byte-to-bitmap mapping (aka. "bitmap fonts"). More complex ones will work with modern fonts, which are essentially programs that interpret text and produce instructions to draw it. Pushed to extreme, you get things like a videogame inside a font[1], or a font that automatically translates startup marketing copy into plain English[2].
What problem is alchemical symbols(https://unicode-table.com/en/blocks/alchemical-symbols/) solving ? Not every character is solving an issue, but they represent something that exists today and is already a standard icon.
Unless you can be sure the client has the typeface to render the unicode symbol, you need to supply the font, so you might as well just send the needed PNG or SVG.
Is it time to admit that simply storing everything as an image is the future?
Sure, it isn't technically as neat, but it gives far more flexibility. It happens already in memes (where text is usually part of a gif)
Text and text encoding dates back from the days when every byte of storage space mattered. Now lossless high resolution images are the norm, and wouldn't constrain what symbols could be included. If you want to make a word be slightly bendy for style, you can!
Accessibility, relayout, etc. can all be done with images too - you simply have a tool which extracts the text, transforms it however you please, and then puts it back as an image.
- You can copy a text symbol and put it into a search engine to figure out what it means.
- Screen readers can recognize text symbols without needing an alt text.
- Text is more flexible on dark or light backgrounds and has better contrast.
- Text is better suited for virtual keyboards and auto-correct or auto-replacement.
- Text can be used in input fields.
- Text has native rendering hints built-in, such as baseline and kerning, which improves inline rendering.
- Text is more flexible when choosing a glyph. For example, you can render it larger for the visually impaired, or change fonts and put a special glyph into the font.
- Text is more flexible when choosing color vs. black-and-white output. Try printing an emoji to see what I mean.
Tooling could be built to do all of that for images too.
For example you could take part of an image and put it into a search engine too... That search engine would take images as input, and where the image represents writing if some kind, it would return relevant results.
Screen readers could look into an image and read out any writing found, just like some screen readers can describe an image ("person in boat holding flag"). Deep nets that can do this reasonably well have been around years now.
Fonts are a standard and cross-platform way to bundle vector images.
Now if we only had some way to figure out which image in the bundle corresponds to what... Maybe some encoding scheme where we assign a numeric ID to each image?
This is the most absurd thing I've heard this month. The page you are looking at right now is in fact an image that your computer put on screen.
However, that fact does not help you when you want to reply to my comment. Or more generally, if you want to edit, transform, store, compress, search.
I can easily search 1GB of text on my computer for a specific pattern. I could not even store all that information directly encoded as (say) raster or vector graphics.
you simply have a tool which extracts the text, transforms it however you please, and then puts it back as an image.
Then you still need some way to handle text, and image to text is not reliable.
Text simply has so much more distilled information. Images are nice for humans, but I can't imagine them as a storage format for programs.
How would the software that lets me add text to images work? Presumably it would use unicode.
And how would the software that extracts text from an image store the extracted text? If it comes accross a number of symbols from different domains how would it save them, and what table of contents would it use as a reference? Unicode.
But that's an implementation detail - some software might let you hand write text into an image with a brush or pen, and that might use as its internal representation the X,Y coordinates of the brush strokes.
So then how will the software convert the X,Y coordinates into text that is read out to a screen reader? I think many of the other comments put it more eloquently than I can.
https://jameshrisho.com/2020/10/making-laundry-less-terrible...