What about an entirely hieroglyphic / semiotic user-interface?

Translating Haiku into multiple languages is labor-intensive. Here’s an idea for discussion: What about a user-interface for Haiku where there are no words?

Instead, a universal set of well-designed symbols — neo-‘hieroglyphs’, if you like — are used to represent all operating-system actions and menu options?

The obvious critique of this method is that there are many menu-options in Haiku; The ‘vocabulary’ of necessary-symbols would become substantial; Learning the full-set would be like learning a small subset of Kanji (the more iconographic-form of written Japanese).

However, if the symbols were sufficiently clear and well-designed, and they gained traction, we would end up with an operating system that transcended international-boundaries and language. The first of its kind.


I wondered what the community here thought about the idea of an operating-system that uses its own custom semiotic-system throughout, rather than an existing phonetic-syllabary-language-system?

Really, universal-semiotics for computing systems should have been defined decades ago. We still haven’t, for example, agreed globally on a symbol to save a document. There are still applications using the floppy disk icon for this, while others use a range of different signs.

It’s also worth considering, that to the ‘average’ person, most of the menu options in any OS, currently, are indecipherable even in ‘English’ etc. Is it time to strip out all those legacy ‘words’ and go full symbolic?


Would use a lot of space in menus, and also not be very clear, so one would need to learn by heart a lot of meanings. Something that people do not seem much inclined to do.

1 Like

And then landing in a endless discussion about what is the best icon for each menu option? If we can not agree on a name for a mere non critical tool like PoorMan… :rofl:


A separate hieroglyphic language should be created for that purpose. This means that the language should be learned separately. So translations will be needed.

Anyway, even now we have contractual signs, program icons. And for a person who doesn’t know English, the inscriptions next to the icons are just letter patterns like hieroglyphs.

In my opinion, what Haiku lacks most now is tooltips with explanations in the application menu.

1 Like

That’s an interesting idea. I doubt it will go anywhere, but, still, it’s interesting to imagine how things would look if we went this route.

Ah yes, I’ve been wishing for something like this to make Icon-O-Matic more intuitive.

The BToolTip class is available if you need it.

But, if I remember correctly, the UI guidelines says it should be used sparingly, only when there is no other way to make the user interface work.

Inventing a new language (and writing system) is more work than using the already existing languages, even if there are many of them. And the translation work is done once, and benefits all users, while the learning of this new language is a task that each user must accomplish. Not really the most efficient way. Even if we only have 1000 users, and Haiku is translated in 25 languages, that means 40 times less work for the translations than for the 1000 users to each learn the new thing :slight_smile:


Ideally this computer-centric ‘hieroglyph’ set would be so well-designed that most of the symbols would not require much active-learning.

But I see what you mean: It would have to be demonstrated that a pure-symbol-based OS could be learned easily, and that such an OS was provably better-liked by users across cultures. Otherwise, as you point out, it won’t be an upgrade, just an obstacle.

I was curious about the general idea of an all-symbolic OS. Not proposing a UI overhaul just yet :slight_smile:

1 Like

I hadn’t considered that.
A mad dictator would have to mandate the icons and censor all dissent :slight_smile:

Radial-menus could free up space for a symbol-only OS user-interface. Below are two implementations of this type of system. However these two examples are not purely iconographic, and so not precisely what I would propose for Haiku:


radial 2

I agree simply swapping-out the text in current menus would not be effective. As you mention, there’s not so much space.

1 Like

Yes, it turns out, text is incredibly efficient at packing a lot of information in a very small space. After all, there may be a good reason hieroglyphs falled out of fashion 5000 years ago in favour of alphabetical and syllabic scripts?

There is also another problem: even if you decide to use pictograms for everything, there is still a cultural barrier on how certain things look. Even more so when talking about objects and actions in a computer, which are very abstract. For example, english has a notion of “download” and “upload”, with a natural direction (one is up, the other is down). French has “télécharger”, which would translate as something like “remote-load” and doesn’t have any implication of direction. Surely, depending on who you ask to design the pictograms, such differences would appear.

This is going very fast into a “let’s change everything” discussion, so I have to ask the question, if you want to change everything, why start from Haiku as a base? I mean, if you want to experiment with it, sure, have fun, that’s great! But if the goal is to integrate the changes in upstream, it seems a bit unlikely that that would happen.

I hope this doesn’t discourage you from experimenting, and, who knows, maybe some ideas that can be integrated into Haiku will still emerge from the experiment?


It would be possible to simply use the already existing and developed system of hieroglyphs. Hieroglyphic signs can be read by everyone in their own language. So you take Chinese hieroglyphs and go…
All you have to do is create dictionaries of these hieroglyphs for each language…
Basically, everyone will have to learn the Chinese writing system.
Now the question is, do I want to learn it?

1 Like

I mostly agree; probably. Japanese Kanji disagrees however, it can still be traced back to meaningful pictograms and is in common use.


We could view hieroglyphs as a ‘failed’ system, but there is always the (admittedly remote) possibility that the Egyptian and Mayan cultures who created advanced hieroglyphic-systems were destroyed by extra-terrestrial-interests. The collapse of those languages, therefore, may not have been on account of any innate flaw in them. Clearly this is off-topic for the Haiku forum, so I’ll move on very, very quickly…

This is a significant concern and, oui, je suis conscient des risques. That said, we are growing closer culturally (although it may not feel like it these last few years) and we already have the beginning of globally-recognized hieroglyphs that work internationally.

Haiku is the best operating system I’ve used in years. My intention is not to change it yet. If anything, I would attempt to code a demonstration of the UI and then it could, if it developed to such a point, be used for a custom-build where I’d evaluate its usefulness and popularity.

Realistically, the pictographic system would exist in-addition to the current Haiku UI, and be radial-menu based. A set of training-wheels for those who were ready to exit the legacy tech of ‘words’ and ‘sentences’.

I joke, of course.


I take your point. However, I’m proposing a smaller subset of symbols for key tasks related only to the operating system. The symbols would be designed to be easily-read by a contemporary human.

The problem with the Chinese writing-system and other quasi-pictographic scripts is that the writing-system became massively-abstracted from the original thing it represented. Most pictograms were obfuscated to the point where the intended meaning can no longer be guessed at.

Both Mayan, Egyptian and Japanese use a mix of logograms (signs associated directly to a specific meaning) and syllabic glyphs. This allows them to write words that are not easy to represent as a logogram, and is very useful to bring (written) language into abstract concepts that are difficult to represent only with logograms.

So, the one that remains in the “purely logographic” category is Chinese. Which indeed does not seem like the easiest thing to learn, at least to me.


In any case, it will still be necessary to have additional verbal records (even if they are optional) for those hieroglyphic signs for beginners.
The second point is that the simple system has such a disadvantage that it is limited in expressiveness. This is precisely why human languages are so complex, because they are comprehensive and even redundant.

Such signs can and are (in modern computing and elsewhere) only additional signs, such as logos, trademarks and the like.

1 Like

One might argue that this idea has existed for a while: toolbars. Toolbars often have the most-used functions in pure icon format with no text. Usually they work fine, but they do take up more space, and many people still prefer to use the textual menus, or turn on text below each icon.

There have also been attempts to solve this textually via “universal” languages such as Esperanto.

It’s an interesting thing to think about, though.


Absolutely. But I feel the problem with toolbars is that because there are also menus as backup, and hover-over text, toolbar designers rarely take the time to make the icons unmistakable in meaning.

Yes, I was thinking of Esperanto. Again, the problem is that Esperanto is not self-teaching. I am speculating on the idea of a UI symbol-set that is relatively small, and quickly teaches itself to the user through simplicity, clarity and elegance.

There have been numerous attempts to create “languages” that transcend languages. I feel motivated to give this hieroglyphic-OS idea a try, since my specific implementation of it will prove, or disprove, the idea’s worth.

Broadly agreed. But although some languages, like Chinese, have identifiable ‘logos’ in their scripts, these are not as clear as what I have planned.

In other words, less of this obfuscation of meaning, as in the Chinese system:

And more of this clarification of meaning:

Obviously these examples aren’t strictly relevant to typical operating-system-related scenarios, but you get the idea.

TL;DR: I will now attempt to improve on Chinese as the world’s leading logographic language and implement the outcome in an experimental-build of Haiku.

This sounds nice, I’m interested.

I’d probably start with a unicode private use area for simplicity, package the icons as fonts and make a “iconlang” translation.

Do you want to develop this open initially?

1 Like

Is it me that computer-usage iconography is coming full circle in this discussion? The original purpose of the GUI icon is to “recognize” something to use or to do without reading a label.

Prior to the advent of any GUI, in the early days of computing, CLI did not require “icons” to represent activities or applications. My first GUI was from Apple IIGS (which was superior to the early Macs) and my last Apple product. Then the Amiga GUI, which for me were the best computers during their era.

So any “mouse-driven” interface uses icons to interact with the OS and other activities since icons by the very nature are language agnostic. An example is a 3.5-inch floppy disk icon in the GUI that represents a physical floppy disc and, by extension, its drive in the PC. If a label exists, that would be language focused.

So back to the beginning of this discourse, if we, as a community, agree to the system and application iconography–which I believe we have–then system language/translations should have a less impact to the user because the user can interact with the system not by “reading” but by “recognizing” the icons to use.

Do I believe localization efforts should be minimized–no. If Haiku plans to be R1 ready, that means we should get better at correcting current translations and add others to meet other international language needs. Sounds like we need to hire polyglots who can assist localization efforts.

1 Like