In this lesson you
- Answered one question per picture: describe it, it's text, or skip it
- Started from an AI draft, and added what only you know
- Applied every answer at once, in one new version
Transcript
One question per picture
Describing pictures. A screen reader can't see a picture. It can only say what someone has written for it. That written stand-in is called alt text, short for alternative text. Only a person can say what a picture means on its page, so Proof hands this work to you. It's on the Pictures and tables step, labelled Objects along the top. For every picture, the step asks one question. Describe it, it's text, or skip it? Describe it is for a picture that carries meaning, like a photo or a chart. You write what a reader needs to know. It's text is for a picture that is only words, like a banner or a letterhead. A screen reader says those words. Skip it is for decoration, like a border or a divider. A screen reader passes over it. The cards on the left list every picture in the document, and a tag on each says where it stands. To fix, in red, means the picture has no text alternative yet. Staged, in green, means you've answered it, but the answer isn't in the file yet. This newsletter has three pictures, and each one needs a different answer.
It's text
The first is a banner. It looks like a picture, but all it does is say something. So choose It's text. A button appears to transcribe it. Choose it, and Proof reads the words in the image. Check the transcription against the banner. A screen reader will say exactly these words. Where you can, real text is still better than a picture of text. It can be resized, and it reflows on a small screen.
Describe it
The second picture is a photograph of the new playground. It shows what the article is about, so it needs a description. Choose Describe it. You don't have to start from a blank field. To get a first draft, choose AI Suggest.
Starting from an AI draft
In a few seconds, there's a draft. Read it against the picture before you use it. AI can miss things, and it can describe things that aren't there. Here it's accurate, so choose Use suggestion.
Adding what only you know
The draft describes what it can see. Only you know where this is, and why it's in the newsletter. So add that. Keep it short. One or two sentences is enough, and the field holds up to two hundred and fifty characters. Charts and diagrams need more than a sentence or two. Charts, diagrams and images of text have their own lesson.
Skip it
To move on, choose Next to answer. It takes you to the next picture still waiting. The last picture is a row of leaves between two articles. It's there for decoration and carries no meaning, so skip it. Proof stages it as an artifact, the PDF's word for decoration. Screen readers pass over it.
Apply
Nothing has been written to the file yet. Each answer is staged, and the count at the bottom shows three. When you're happy with all three, choose Apply 3. Proof writes them together, in one new version. Then Proof checks the document again. When it finishes, the Objects step shows a check mark, which means nothing is left to do there. The step confirms it. Every picture has a text alternative, or is marked decorative. Now you can decide what each picture is, start from an AI draft without trusting it blindly, and write every answer in one go. Tables share this step, and have a lesson of their own.
Recap
That's the end of this lesson.
Maple Ridge Community Health, its documents and its people are fictional, made for these lessons.