[askaboutit]

Wikipedia is a static delivery website. Of course it’s easy if all you’re delivering is static content. Try that with an ecommerce site where dynamic information is sent from sometimes multiple sources. It’s better to load it once and have a reactive site call just the necessary apis after. Calling everything over and over again in eCommerce is a pain for users. Unless you can afford a database and servers to be distributed globally to counter the added latency of downloading everything again and again. PWA is the next step for that. But, yeah, ok, for Wikipedia you can be static as you want.

[viraptor]

Unless you have a very active community commenting and interacting around the items, you can definitely go almost entirely static, even in ecommerce.

Regenerate pages after changes happen and you'll be fine. You'll still need dynamic search most likely, but even those can be pre-generated for many terms. (Your admin / content management part needs to be dynamic of course, but that's not customer-visible)

If your catalogue is so large that constant regeneration is impractical, you can generate on-demand and cache long-term a few layers above for anything not requested recently.

[briandear]

> You'll still need dynamic search most likely, but even those can be pre-generated for many terms.

Have you ever worked on a high scale e-commerce site? I have and what you are talking about is impractical and pretty much impossible.

Products have multiple variants, photos for each variant. Various companion products that depend on what you already selected. Pricing options that can depend on quantities or packages. And search? Spend 5 minutes on any serious e-commerce search system and there is no such thing as “common search terms.” Of course there are common searches, but on any non-trivial e-commerce system, you have potentially thousands of distinct common searches.

I was one of the original engineers for https://www.matalan.co.uk and you can’t just “regenerate” pages after changes. You can regenerate the cache for images or product descriptions, but e-commerce isn’t like a printed catalog. We put exceptional engineering into that application and to trivialize that sort of application like it was some kind of blog site kind of demonstrates a lack of experience in building something that serves millions of visitors per month — visitors that all have different paths based on what they want to buy.

[viraptor]

There are various types of ecommerce. Some will be harder to do, some easier, done impossible. Some are search heavy, some aren't.

I agree with the issues you're raising, but even then, you can make a lot of the site bypass dynamic rendering.

[askaboutit]

Extremely performance store. Mind giving a brief overview of the stack?

[jmull]

> Wikipedia is a static delivery website.

What do you think all those little "Edit" buttons all over wikipedia do?

[adventured]

> What do you think all those little "Edit" buttons all over wikipedia do?

They take you into the CMS where you can produce more static content.

[marmshallow]

Surely you must agree that Wikipedia is a more static site than, say, Facebook.