The "index" api link is not an index.

[Pomax]

URL:

https://nodejs.org/api/index.html

Browser Name:

n/a

Browser Version:

n/a

Operating System:

n/a

How to reproduce the issue:

1. Visit https://nodejs.org/api/index.html
2. note that this is identical to the nav system on the left

That's it, that's the whole bug, the index page is not a documentation index page in the slightest. This should be a very long page (or many pages split on initial letter) with all terms across all of the documentation listed so that people can find the actual word they're looking for, and then click through to the thing they're looking for.

This is an index: https://docs.oracle.com/en/java/javase/18/docs/api/index-files/index-1.html

The nodejs.org page is not.

[ovflowd]

What you probably want to achieve is https://nodejs.org/api/all.html no?

[ovflowd]

Also, this is not really a bug, so I've removed the bug label :)

[ljharb]

I'm confused, there's nothing about "index.html" that requires it to actually BE an index - that's just the history of where the name comes from.

[Pomax]

@ljharb That doesn't make a lot of sense: when you're reading documentation, a link called "index" should absolutely be an actual index. And of course, "index" on the web can also mean the landing page, but that's usually called "top" or "home", and in this specific case the landing page for https://nodejs.org/api has no information on it, it's only a duplicated set of navigation links, so there is no reason for anyone to ever want to go back to that page (whether it's called "index.html" or not) as the only thing it offers is already found on every other page the user would already be on. It's not an index, or a landing page. At least not in the sense that any documentarian would agree with.

@ovflowd indeed, I would recommend that the link to "index" points to that page instead. But then with sensible pagination added (so really, separate pages that will work even with JS disabled) because that's an 8MB document right now, which browsers are not great at dealing with =D

[ovflowd]

I feel like the /index.html in the past was supposed to land to https://nodejs.org/api/documentation.html

Something might have changed, as on https://nodejs.dev/en/api we show the "documentation.html"

[ovflowd]

@richardlau did we accidentally change something on the API side of things with NGINX? (I don't recall such change, but just double-checking)

[richardlau]

Not that I'm aware of. You can check with https://web.archive.org/ what was being returned.

[ljharb]

@Pomax index.html is the default page on every site i'm aware of for decades; "home" or "top" aren't names i've ever really seen. I'm baffled anyone would have the expectation that a URL named "index.html" would be an actual index - that simply hasn't been the case since the 90s.

[Pomax]

index.html is the file that for the longest time most things automatically redirect to (but tons of things don't, it's still a manual thing you need to configure in plenty of servers), but the location it represents (when used in online documentation like API docs, which is the context here), is "top" or "home" or the like, not "index". The word "index", specifically in the context of documentation, online or on paper,still has the very specific meaning of being the index list of terms used in that documentation.

For regular websites, yes, "index" means the homepage, or landing page, or whatever else we call it. But this is not a regular website. This is API documentation. Its conventions and user expectations are measured against other API docs, not regular websites.

[ljharb]

API documentation is indeed a regular website. The only context where anyone should expect "index.html" to be an index is a filesystem directory listing.

[ovflowd]

I agree with @ljharb here, but I appreciate the healthy discussion about what the API docs index should or should not be (and what an index file even is).

I feel we might be derailing here.

Also, @richardlau seems I was wrong (https://web.archive.org/web/20190122151446/https://nodejs.org/api/index.html) index was just an "index".

I'm all-in onwards on changing this behaviour as I also would prefer the main "documentation.html" on serving as the "index".

The question now is, if other collaborators are happy with this. Also, this issue should be moved to nodejs/node as it is not (currently) on the realm of the Website Team regarding any changes on the living API docs.

If anybody from @nodejs/collaborators could transfer this issue to node that'd be great!

[cclauss]

cclauss: What is index.htlm?

ChatGPT: index.html is the default file name for the home page or the starting point of a website. It is a standard web page that is displayed when someone visits a website's main URL.

When a user types a website's domain name into their web browser's address bar, the web server will automatically look for a file named index.html in the website's root directory and display its content on the user's screen. If an index.html file is not present, the server may display a directory listing or a 404 error page instead.

The content of the index.html file typically contains HTML, CSS, and JavaScript code that defines the structure, layout, and interactive elements of the website's home page. It may also include links to other pages on the website, images, videos, and other media.

cclauss: My vote would be to close this issue.

[ovflowd]

I'd argue this is now a stale issue, and I doubt there's any short-term actionable here as the Docs are generated the way they are. Not to mention these docs are not generated on the Website repository. So I'd argue core folks should handle this.

I will move this to nodejs/core and see what the triggers and collaborators think about this.

[Ayase-252]

I think that the content of index.html is decided by their developers. It is not neceesarily a glossary page. It is worth to mention that the documents in whole is offered in all version.

So, I prefer no change on this.