← Back to context

Comment by wooque

4 years ago

I, as a user, hate them with passion.

I regularly stumble upon stale data in SPAs, even in big names' like Linkedin, Jira, Github, etc. I wonder how the hell corporations with thousands of engineers, lot of them being brightest engineers in the world, can't make proper SPA.

Can we just go back to server side rendered pages, with a bit of JS sprinkled in, please.

Github is not an SPA. They use the middle road of dynamically loading HTML snippets, using webcomponents, and lightweight JavaScript. Any problems you see with most of Github's interface should be chalked up to server-rendered pages with sprinkled JS.

  • I dread using GitHub's fake client-side navigation. For me it's always faster to just load a link "for real" rather than wait for their terrible JS to pretend to load the link for me.

Interesting that you use Github as an example, when Github is not an SPA. So basically just admitting these problems aren't necessarily inherent to SPA's, but to the developers/teams building them. They may just be a bit harder to get right in SPA's.

> Github

GitHub is a great example of a site that would benefit from actually being an SPA, imo. It is server rendered pages with JS sprinkled in, and you can easily end up in the situation where parts that live update fall out of sync with parts that don't. For example, being able to see that an issue is closed while the badge on the Issues tab still says you have one open.

I share your general sentiment over the state of them though. It's not actually that hard to avoid breaking navigation buttons, having links open in new tabs correctly etc, so I don't know how it gets screwed up so much.