We can do better than wysiwyg editors.
Standing for “what you see is what you get”, this style of editorial interface has become so ubiquitous that many people consider it a must-have for a writing environment.
I take the opposite view. Wysiwyg should be a must-not-have. They should not even be an option. They should just not exist in a digital publishing environment.
Wysiwyg is based on a promise that cannot be met. Moreover, it is based on a promise that should not be met.
As such, the obsession with wysiwyg is leaving the integrity of content in tatters.
The false premise of wysiwyg
The idea of wysiwyg is that, when you are writing and editing content, you will see, in real time, a realistic impression of what the content will look like as its final output.
For example, when you are editing a webpage, it is supposed to show how the content will look on the webpage. When you are editing a document in Microsoft Word, it shows up as it would if you were to print it out.
This is appealing to many content editors and organisations. It gives them a sense of control — particularly over how the output will look.
The problem is, we cannot — and should not — control how content looks.
What you see in digital is inherently uncontrollable by the publisher
One of the main reasons for this is accessibility. A great deal of the power of digital technologies like the web is that it gives our users control over how they consume information. This is necessary to meet accessibility needs.
For example, your user might need to increase the font size. Or they may want to view it in high contrast mode. Or dark mode. Or any number of highly-configurable display options.
They might be reading your information on an uncommon technology of their choosing, like a TV-based web browser, or a low-resolution e-ink screen.
They might not even see the output at all. They may hear it through a screen reader. Or they may touch it using a Braille terminal.
Moreover, even leaving aside assistive technologies, and user preferences like dark mode, the notion that you can know how anything will look on another monitor is an illusion. Different screen types produce pixels in different ways. This makes pixel perfection an impossibility.
In short, just because something likes a certain way on your device, it almost certainly presents differently on someone else’s device.
What your users get is not what you see in your wysiwyg editor.
Why it’s a problem
You might not think it is a big deal, as long as your website’s design has passed an accessibility audit.
But the problem is, if you are focusing on what your content looks like when you are using a wysiwyg editor, by definition you risk creating confusion between what it looks like and what it means. And this creates accessibility problems.
For example, countless times I have seen people misuse a heading just because they wanted the text to be bigger. Or use a blockquote because they wanted to indent the text. Or hard-code a different font colour because they just wanted to use a different colour for whatever vague reason.
This all causes accessibility errors that would not be caught by a generalised accessibility audit. In short, your users can become confused unless the meaning of your visual choices is clear — in both a human-readable and a machine-readable way.
Wysiwyg risks losing meaning from your content
Part of the problem is that content authors who do things like change font colours presumably do so because they mean something by it. They may not consciously know what they mean, beyond something like: “this is important and/or different”. But any meaning like that cannot be conveyed using visual cues alone.
An important concept is the separation of concerns. This holds that different aspects of a computing problem must be managed separately. In this case, we are specifically talking about separation of presentation and content.
Beyond the accessibility reasons described above, we need to do this for a major practical reason. If the visual design needs to change one day, we need to be able to make that change in itself, without needing to touch the content. This is why CSS was invented — in order to separate stylistic elements away from HTML.
Therefore, when an editor tries to use visual cues to convey meaning, they risk that meaning being lost. It can be lost when the visual design changes. It can also be lost if the interface uses an assistive technology — like a high-contrast mode that reinterprets visual cues, or a screen reader that can’t even present purely visual cues.
By binding content with presentation too tightly, wywiwyg violates the separation of concerns principle.
Instead, we need to ensure that the meaning intended by the author is conveyed in a machine-readable way. If necessary, that machine-readable code can then be used to achieve any required visual change.
Wysiwyg is a symptom of the print-based mindset
Wysiwyg came to prominence in the 1980s through desktop publishing. In that context, wysiwyg made some sense. The output of a print product could be controlled.
Indeed, it is hard to imagine how any desktop publishing tool would have succeeded without giving the author a sense of how the final output would look. The amount saved in time, ink and paper alone is enough to make the benefits of wysiwyg clear.
But even in this era, flaws in the approach were felt. In 1996, Conrad Taylor — an information designer with a background in print — wrote a lengthy essay about the drawbacks of wysiwyg.
Wysiwyg makes well-paid people get bogged down in visual minutiae
Conrad Taylor noted that wysiwyg merely provides the illusion of control. In reality, it forces non-designers to get “bogged down by minutiae”. He noted that wysiwyg essentially assumes that the content editor is almost impossibly multiskilled:
- Subject expertise
- Writing and editing ability
- Visual sense and design skills
- Computer and keyboard skills
More recent research has confirmed this to be a problem. A 2019 study suggested that well-paid professionals are spending 8 hours a month just on formatting PowerPoint documents, not actually writing them. Yet, despite the huge amounts of time sunk into presentation visuals, 70% of documents fail to adhere to corporate design guidelines.
Wysiwyg creates a distraction from the actual writing
iA.net noted the problem with this focus on formatting:
This preoccupation mirrors a form of fetishism, where the superficial attributes of a document are revered, often at the expense of the substance within. Users find themselves ensnared in a cycle of selecting fonts, adjusting margins, and tweaking layouts, believing that these elements inherently enhance the quality of their work. This fixation on form over function diverts attention from the core purpose of writing: to communicate ideas clearly and effectively.
In another article, they noted that focusing on presentation when you are supposed to be writing creates a cognitive burden. That’s because you are essentially making yourself multitask.
As Michael Tsai wrote in 1998:
When I’m composing a document, I don’t want to be concerned with the way it looks. Just because the actual document uses a small, hard-to-read-on-screen script font, doesn’t mean I should not be able to edit with a large font designed for on-screen use. When doing layout, it’s nice to see where page breaks fall and how columns and margins look. When I’m writing, they just take up screen real estate would be better-used for letting me see more of the document.
Rachel Andrew put it even more strongly when she said: “your wysiwyg editor sucks“.
Wysiwyg keeps us shackled to the print-based mindset
One of the most notorious wysiwyg editing environments is Microsoft Word. When you open it, the first thing you see is a blank piece of paper. Even today, in 2026. The baked-in assumption is that you are writing in order to print it on an actual piece of actual paper.
And we wonder why so many people remain shackled to the print mindset?
As noted by iA.net, Microsoft Word is not a writing app. It is “a second-class layout app”.
And this is the core problem of the wysiwyg ideology. It was never about writing at all. It was always about layout. And layout should be separated from content.
Wysiwyg is a barrier to multichannel publishing
As corollary to this, wysiwyg’s print-based metaphors are an active barrier to multichannel publishing.
This was already being observed way back in 1996. Conrad Taylor noted that if you are in the business of publishing on the web:
…it’s impossible not to be a multiformat publisher when you publish this way. You have no control over whether your publication is being viewed in color or in black and white, you do not know what fonts your page is being viewed in, and since you do not know how wide the browser window has been set, you cannot predict where line breaks occur. In the future, Web pages may also be viewed more on tv screens and hand-held “personal digital assistants,” strengthening the arguments for a properly generic markup.
Wysiwyg makes our writing superficial and our markup bloated
Despite all the problems with wysiwyg, we still see editing interfaces that are no more than glorified layout-tweaking systems. They assume that you want to make a layout first.
“Drag and drop! You’ll love it!” Place this widget here. Drag that block there. (Just hope you don’t rely on a keyboard to make all these changes.)
This literally makes the actual content an afterthought.
WordPress’s Gutenberg editor is a case in point. You can spend endless hours dragging widgets around a screen trying to get it to do the thing you want it to do. You might not even succeed at getting the layout you want. Yet the result is a worse, less informative piece of writing.
As for the code wysiwyg editors produce, it becomes a bloated vat of div soup. HTML tags are stripped of their semantics, and abused in support of superficial looks.
The wysiwym movement — What you see is what you mean
Of course, I am not the first to consider the need for an alternative to wysiwyg. In the course of researching this article, I have discovered an entire approach that describes semantic document editing as “wysiwym” — what you see is what you mean.
But this does not go far enough. It still implies that you will see the information. But in our rich multichannel world, we could be hearing it spoken out loud by a smart speaker. Or users of assistive technology may perceive your information in a completely different format like a Braille terminal.
Going further, we need to face up to the reality that our information is often not even being consumed by users from start to finish. I’m not just talking about the way people scan a page to find the information they are looking for.
Large language models are doing their own sort of cherry picking. Generative AI systems take our information, and send it into a giant cauldron containing every second-hand book known to humanity. They then synthesise it into something different, in a completely different context from where the information was originally published.
Towards wygiwym — What you get is what you mean
We need to completely move on from the idea that we should control how our information is presented. We need to stop trying to control the fonts and the colours. We need to stop thinking of our information as pixelated print documents. We need to stop even expecting our words to be represented faithfully in the order we placed them.
We must do this if we are going to create a dynamic information environment that is fit for purpose for today’s multichannel experiences, and users’ multifaceted expectations.
Don’t create layouts.
Don’t create a page.
Encode the meaning.
That way, what you get is what you mean.