José Solé Proofing type is crucial in order to refine and ensure the fonts we design work as we want them to. In the last few years there’s been a plethora of tools that streamline the process of creating proofing documents. They tend to focus on facilitating the creation of a structure, while relying on the designer to provide the content; combinations of letters, words and sentences.
In this talk I will discuss how code can be used to generate this content, making proofing document content a dynamic element that adapts to each stage of the type design process.
I will start by discussing why proofing documents are important, and the content that can be found in good examples. Since many parts of a proofing document require words, I will review a few of the tools that can be used to obtain words based on requirements such as the inclusion or combination of specific letters, the inclusion of diacritics or the combination of multiple letters for creating patterns to check texture or spacing. I will briefly mention how current generative AI fails to provide accurate and useable results for this purpose.
Finally, I will explain how I ended up approaching my own proofing document generation using Python and tools like Drawbot, Wordsiv and a variety of scripts that filter through millions of words within text corpora from a variety of languages. I will demonstrate how to use these tools for both the Latin and Arabic scripts, and discuss how they could be expanded in the future.