Accessibility Solutions

Continuum of accessibility solutions ranging from literal through to free.

Introduction

Conversion of content into accessible formats involves a variety of translation activities and solutions. Whilst the world of accessibility has yet to formally categorize these types of activities, taxonomies of language translation solutions have existed for some time. The translation solutions described by these taxonomies are very useful when considering the difference between automated and human accessible document formatting.

These translations solutions can be expressed as a continuum between literal and free solutions.

Automated translation solutions have traditionally focussed on literal translation solution types but are getting better at the sort of translation that humans do. For example, Facebook now uses AI to add alt text to photos.


Lexical translation

Before computers can start doing anything with a document, they need to convert characters into tokens. In the case of a scanned PDF, OCR software recognizes characters in raw input and then converts them into tokens, such as words.

PDF extraction flowchart

Source: Data Wrangling Handbook


Literal translation

Literal translation takes place on a word for word basis. An example of this is converting audio into captions.


Explicitation

Explicitation takes implicit or general information and makes it explicit or specific.

Example – Use of color

In the following example, red and green indications of no and yes are made accessible by adding the words 'None' and '64'.

DateIEChromeSafariFirefox
Feb 2018     

In the following example, text information has been added to make the values clear.

DateIEChromeSafariFirefox
Feb 2018 None 64 10 or 11 58  

Example – Input field labels

In this example, users are assumed to know what they are searching for.

In this example, the purpose of an input field is made explicit.


Resegmentation

Resegmentation involves joining sentences, cutting sentences or re-paragraphing information. Assistive technologies often resegment information for users based on the semantic structure of documents.

Example - reparagraphing

An example of reparagraphing would be taking a paragraph in a PDF document, that is split over two pages, and combining it into one paragraph.


Semantic translation

Semantic translation involves describing the structure, function and relationship of elements in a document. 65% of screen reader users use headings as their primary form of navigation.

Example – Data table

For example, when navigating a data table, assistive technologies can combine the column heading, row heading and cell data.

Calendar

TimeMondayTuesdayWednesdayThursdayFriday
8:00-9:00 Meet with Sam     
9:00-10:00   Doctor Williams Sam again Leave for San Antonio

Screen reader: ‘Wednesday 9:00 -10:00 Doctor Williams’.

Example - headings and lists

An example would be identifying different levels of headings.

Inaccessible Content

- enable students to develop contacts to assist their entry into employment

- allow students to gain overseas referees who can comment on their work skills

Accessible Content

  • enable students to develop contacts to assist their entry into employment
  • allow students to gain overseas referees who can comment on their work skills

Compensation

Compensation involves expressing material in a way that will make sense in a target document. For example, sighted users can read horizontally across a line of text, encounter a footnote symbol, and then jump vertically to an explanation at the bottom of the page. Users of assistive technologies move in a much more linear fashion and may need to have footnotes relocated to the end of the document.


Intersemiotic Translation

Intersemiotic translation involves taking transferring meaning between two different sign systems. An example would be taking a visual sign, such as an image and describing it in text or taking silent actions in a film and describing them in words, e.g. 'door opens to reveal dark hallway with a light at the end'.