I agree some of their examples aren’t great, but I think what they’re saying is that Astro’s benefits become less relevant for apps which are highly interactive in the client. You can certainly use Astro to build those, but in many cases you may be just as well served (if not better) by using a more established tool oriented toward whichever client framework/libraries you choose (eg Next.js or Nuxt or SvelteKit etc).
That said, there’s quite a lot of space between that extreme and the mostly-static “web site” extreme, and I agree Astro would be a better fit than they let on, for a large chunk of that space. My suspicion is that this is intentional, to keep their current focus and explicit use-case commitments narrower. And I wouldn’t expect it to remain that narrow in the future.
That said, there’s quite a lot of space between that extreme and the mostly-static “web site” extreme, and I agree Astro would be a better fit than they let on, for a large chunk of that space. My suspicion is that this is intentional, to keep their current focus and explicit use-case commitments narrower. And I wouldn’t expect it to remain that narrow in the future.