Table of Contents

What is a Rich Text Editor?
The Modern Rich Text Editor and Emerging Challenges
How we Evaluated Rich Text Editors
Why Tag1 Selected ProseMirror
Conclusion

Among all of the components commonly found in content management systems (CMSs) and typical editorial workflows, the rich text editor is perhaps the one that occupies the least amount of space but presents the most headaches due to its unique place in content architectures. From humble beginnings in discussion forums and the early days of the web and word processing, the rich text editor has since evolved into a diverse range of technologies that support a lengthening list of features and increasingly rich integrations.

Recently, Tag1 embarked on an exploration of rich text editors to evaluate solutions for a Fortune 50 company with demanding requirements. In this blog post, we’ll take a look at what impact the choice of a rich text editor can have down the line, some characteristics of the modern rich text editor, and Tag1’s own evaluation process. In the end, we discuss some of the rationales behind Tag1’s choice of ProseMirror as a rich text editor and some of the requirements leading to a decision that can serve as inspiration for any organization.

What is a Rich Text Editor?

At its core, a rich text editor enables content editors not only to insert and modify content but also to format text and insert assets that add to the content in question. They are the toolbars that line every body field in CMSs, allowing for a rich array of functionality also found in word processors and services like Google Docs. Most content editors are deeply familiar with basic formatting features like boldfacing, italicization, underlining, strikethrough, text color, font selection, and bulleted and numbered lists.

There are other features that are considered table-stakes for rich text editors, especially for large organizations with a high threshold for formatting needs. These can include indentation (and outdent availability), codeblocks with syntax highlighting (particularly for knowledge bases and documentation websites for developers), quotations, collapsible sections of text, embeddable images, and last but not least, tables.

While these features comprise the most visible upper layer of rich text editors, the underlying architecture and data handling can be some of the most challenging elements to implement. All rich text editors have varying degrees of customizability and extensibility, and all editors similarly have different demands and expectations when it comes to how they manage the underlying data that ultimately permits rich formatting. In the case of Tag1’s top Fortune 50 customer, for example, the ability to insert React-controlled views and embedded videos into content ended up becoming an essential requirement.

The Modern Rich Text Editor and Emerging Challenges

Whereas many of the foregoing rich text editors available in the late 1990s and early 2000s trafficked primarily in basic formatting, larger editorial organizations have much higher expectations for the modern rich text editor. For instance, while many rich text editors historically focused solely on manipulation of HTML, the majority of new rich text editors emerging today manipulate structured data in the form of JSON, presenting unique migration challenges for those still relying on older rich text editors.

Today, there are few to no robust rich text editors available that support swappable document formats between HTML, WYSIWYG, Markdown, and other common formats. Any conversion between HTML, WYSIWYG, and Markdown formats will result in some information loss due to differences in available formatting options. As an illustrative example, a WYSIWYG document can include formatting features that are unsupported in Markdown, such as additional style information or even visually encoded traits such as the width of a table column. While converting a document format to another preserves the majority of information, there will inevitably be data loss due to unsupported features.

Moreover, as rich text editors become commonplace and the expectations of content editors evolve, there is a growing desire for these technologies to be accessible for users of assistive technologies. This is especially true in large companies such as Tag1’s Fortune 50 client, which must provide for content editors living with disabilities. Rich text editors today frequently lack baseline accessibility features such as ARIA attributes for buttons in editorial interfaces, presenting formidable challenges for many users.

How we Evaluated Rich Text Editors

Tag1 evaluated a range of rich text editors, including ProseMirror, Draft.js, CKEditor 5, Quill, Slate, and TipTap. Our mission was to find a solution that would be not only robust for content editors accustomed to word processors and Google Docs but also customizable and robust in handling the underlying data. But there were other requirements as well that were particularly meaningful to the client for whom Tag1 performed this evaluation.

An important first requirement was the ability for the chosen rich text editor to integrate seamlessly with collaborative editing solutions like Yjs and Collab out of the box. In addition, because of the wide-ranging use of open-source projects at the organization, a favorable license was of great importance to allow teams to leverage the project in various ways. In addition, other characteristics such as plugin availability, an active contributor community, and some support for accessibility were considered important during the evaluation.

As mentioned previously, other requirements were more unique to the customer in question, including native mobile app support, which would allow for mobile editing of rich text, a common feature otherwise found in many responsive-enabled CMSs; embedding of React view components, which would provide for small but rich dynamic components within the body of an item of content; and the ability to annotate content with comments and other notes of interest to content editors.

The table below displays the final results of the rich text editor evaluation and illustrates why Tag1 ultimately selected ProseMirror as their editor of choice for this project.

 

 

ProseMirror

Draft.js

CKEditor 5

Quill

Slate

Shared Editing

✔ (Yjs,Collab)

✖ *

✔ (Proprietary on-prem SaaS)

✔ 

(Yjs,ShareDB)

✖ *

License

MIT

MIT

GPL2+

BSD-3-Clause

MIT

iOS Mobile Support

✖ (issue)

Embed of custom views

Annotation of content

✖ **

✖ **

Plugin System Design

+++

+

+

++

++

Community

+++

++

+++

++

+

Accessibility⑅

?

?

✔ ***

?

?

Stability

✔ (though 5 is very young)

Demos

Table layout
Atlaskit

List of demos

Rich Text

Rich Text

Rich Text

* Doesn’t support feature yet, but could be implemented (additional effort & cost) ** Comments are part of the document model (requirements dictate they not be) *** Per CKEditor documentation -- needs to be verified (see review process below) ⑅ In Depth accessibility reviews must be completed before we can grade

Why Tag1 Selected ProseMirror

Ultimately, Tag1 opted to choose ProseMirror as the rich text editor of choice for their upcoming work with a top Fortune 50 company. Developed by Marijn Haverbeke, the author of CodeMirror, one of the most popular code editors for the web, ProseMirror is a richly customizable editor that also counts on an exhaustive and well-documented API. In addition, Haverbeke is known for his commitment to his open-source projects and responsiveness in the active and growing ProseMirror community. As those experienced in open-source projects know well, a robust and passionate contributor community does wonders to lower implementation and support costs.

Out of the box, ProseMirror as a tool is not particularly opinionated about the aesthetics of its editor or especially feature-rich. But this is in fact a boon for extensibility, as each additive feature of ProseMirror is provided by a distinct module encapsulating that functionality. For instance, while core features that are considered table-stakes among rich text editors today such as basic support for tables and lists are part of the core ProseMirror project, others, like those that provide improved table support and codeblock formatting, are only available through community-contributed ProseMirror modules.

ProseMirror also counts among its many advocates large organizations and publishers that demand considerable heft from its rich text editing solutions. Newspapers like The New York Times and The Guardian, as well as well-known companies like Atlassian and Evernote, are leveraging ProseMirror and its modules. In fact, Atlassian published the entirety of their ProseMirror modules under the highly permissive Apache 2.0 license.

Beyond the fact that many editors are based on ProseMirror as a foundation available through open source, such as Tiptap, the Fiduswriter editor, and CZI-ProseMirror, it was a logical choice for Tag1 and part of Tag1’s commitment to enabling innovation in editorial workflows with a strong and stable foundation at their center. Through an integration between ProseMirror and Yjs, the subject of a previous Tag1 blog post on collaborative editing, all requirements requested by the top Fortune 50 company will be satisfied.

Conclusion

Choosing a rich text editor for your editorial workflows is a decision fraught with small differences with large implications. While basic features such as simple list and table formatting are now ubiquitous across rich text editors like ProseMirror, Draft.js, CKEditor, Quill, and Slate, the growing demands of our customers obligate us to consider ever more difficult requirements than before. At the request of a top Fortune 50 company, Tag1 embarked on a robust evaluation of rich text editors that satisfied some of their unique requirements such as React component embeds, accessibility, and codeblocks with syntax highlighting.

In the end, teams opted to leverage ProseMirror due to its highly active and passionate community, the availability of unique features such as content annotations, native mobile support, and accessibility support. Thanks to its large community and extensive plugin ecosystem, Tag1 and its client can work with a variety of available tools to craft a truly futuristic rich text editing experience for their content editors. As this evaluation indicates, it is always of utmost importance to focus not only on the use cases for required features, but also on the users themselves who will use the product—the content editors and engineers who need to write, format, and manipulate rich text, day in and day out.

Special thanks to Fabian Franz, Kevin Jahns, Jeffrey Gilbert, and Michael Meyers for their feedback during the writing process