Don't Skip HTML Validation

You’re probably already using a few tools to automatically check your software — like linters, automated tests, code formatters, or a type-checking compiler.

These are all great, and we use them at Supercede too. The more checks we can run, the better. And the earlier in the development process, the better.

What I’ve noticed over the years though, is that we tend to focus on more advanced, often esoteric forms of analysis. With training and practice, we can express our application’s logical constraints in the type system, or verify properties of functions with property-based testing.

Yet with all this sophistication, we often forget to check whether the HTML we’re generating is even valid.

Producing invalid HTML has bitten us in the past. Fixing the HTML is usually easy, but until it’s fixed you might have a broken page. If you’re lucky, it’ll be broken in an obvious way. Sometimes it might be broken in a subtle way, which is harder to debug (which means it costs your team more money).

We’ve had online HTML validators like Nu HTML Checker for ages. You can paste your HTML in there and immediately validate your HTML for free.

This does get tedious though. You (or your colleagues) won’t remember to do this every time you update a page. And having a human do this manually just doesn’t scale.

I’m writing a web application which validates pages across entire websites, automatically. It uses both a crawl-based and a push-based approach, which means it can validate pages that are behind a login screen.

Here’s a screenshot of how it looks currently.

I’m planning to launch this as a public tool soon.

Interested in trying it? Drop your email below and I’ll let you know when it’s ready.