Nathan Andrew Cunningham (nac)

Documentation systems often assume reasoning will be obvious. It rarely is. Engagement teams understand the file while building it. Reviewers see it after context has faded. Inspectors see it without internal history at all. If documentation relies on shared memory, defensibility declines with each review layer. Design must anticipate distance. #AuditQuality #InspectionReadiness

5

1 Comment

Do you know if your site is ADA Compliant? In Today's episode, The Accessible Web: ADA Compliance Unveiled, I will go through the top accessibility requirements that can affect your site’s ADA Compliance. https://lnkd.in/dumDRj-r

1

Strong technical writing isn’t about writing, it’s about listening. • Product explains intent. • Engineering explains reality. • Support explains pain points. • Users explain confusion.

2

A lot of technical writing work is devoted to updating, expanding, and improving existing documentation. But the challenges are very different when a product doesn't have any customer-facing documentation at all. At this week’s Write the Docs East Coast Virtual meetup, we’ll explore best practices for new product documentation. Let’s break down the problems and share strategies to tackle them. RSVP below, and make sure your Zoom name matches your RSVP name. East Coast Virtual Meetup: Best Practices for New Product Docs Friday, February 20, 8:30am to 9:30am Eastern RSVP here:https://lnkd.in/enR2eWb8

15

1 Comment

QA was never really about writing scripts. It’s about protecting revenue. Every escaped defect means lost conversions, more support load, and potential churn. Yet most teams still spend the majority of their time maintaining test code. AI-first QA changes the focus. Define intent. Validate behavior. Measure risk reduction. When automation is generated at scale, the question shifts from: “How many scripts did we write?” to “How much business risk did we remove?” Read the full blog: https://bit.ly/3OK7zNn

3

This morning, while drinking my coffee, I was thinking about something that’s becoming a bit of a pet peeve in the QA world. For years the narrative has been: “We don’t have enough time.” “We don’t have enough people.” “We need more tests.” Now that AI makes it possible for engineers to generate hundreds of tests quickly, I’m seeing a common reaction from parts of the QA community: test volume is meaningless, this is garbage, AI-generated tests don’t count. Which is frankly becoming more and more annoying to me as a response. For a long time the constraint was supposedly resources. Now that we have tools that dramatically expand those resources, the conversation suddenly shifts to “well quantity doesn’t matter.” These posts seem to conviently ignore that good engineers know that more tests doesn’t automatically mean better testing. Bad tests exist. Low-signal tests exist. Maintenance cost is real. But there’s a difference between healthy skepticism and immediate dismissal. If a competent engineer says they used AI to generate a large set of tests, maybe the first reaction shouldn’t be “this is bullshit.” Maybe the first reaction should be curiosity: - What kinds of tests were generated? - What coverage did they add? - What signal do they produce? - What did we learn? AI doesn’t replace judgment. But it absolutely changes the economics of testing. Generating tests used to be expensive. Now analyzing and curating tests is the expensive part. That’s not a threat to QA. If anything, it highlights where the real expertise has always been: deciding what actually matters. Anyway. Just something I was thinking about this morning with my coffee.

43

24 Comments

Most documentation tells users what your product does. Few guide them through it. User personas aren’t just “who the user is.” They show where users are in their journey and what they need at each stage. Miss that, and users struggle to find what they need. Learn how to create documentation that actually guides users: https://lnkd.in/d5wmt-as #TechnicalWriting #UserExperience #DocsAsCode #Documentation #UX

3