Replace www links with /

[eksperimental]

Speaking of links, one things that I have had on my todo list is to change any links that point to www.erlang.org to instead point to / if they are not to be redirected in _redirects. This would allow the broken link checker to check those links as well just to make sure that they remain.

Originally posted by @garazdawi in https://github.com/erlang/erlang-org/pull/61#pullrequestreview-855608153

[eksperimental]

These are the contents of _redirects

## Redirects to erlang.org
/download/*         https://erlang.org/download/:splat
/~*                 https://erlang.org/~:splat
/course/*           https://erlang.org/course/:splat
/documentation/*    https://erlang.org/documentation/:splat
/mailman/*          https://erlang.org/mailman/:splat
/mailman-icons/*    https://erlang.org/mailman-icons/:splat
/pipermail/*        https://erlang.org/pipermail/:splat

## Old removed URLs
/doc/search         /doc
/docs/versions      /documentation
/rss/news/*           /news.xml
/erlang-enhancement-proposals/home.html /eeps/
/erlang-enhancement-proposals/* /eeps/:splat

## Redirects from old sites
https://blog.erlang.org/feed.xml https://www.erlang.org/blog.xml 301!
https://blog.erlang.org/*        https://www.erlang.org/blog/:splat 301!
https://bugs.erlang.org/browse/* https://www.erlang.org/bugs/:splat 301!
https://bugs.erlang.org/*        https://github.com/erlang/otp/issues 301!

So what should we do?

[garazdawi]

This is an example of such a link, as it does not do any redirect: https://github.com/erlang/erlang-org/blob/master/_news/149.md?plain=1#L24

However, we do not want this type of link to be changed:

https://github.com/erlang/erlang-org/blob/master/_news/145.md?plain=1#L22-L26

as they are caught by the redirect.

I would expect that almost all of the links we want to rewrite point to https://www.erlang.org/doc/, but there might be other as well.

[eksperimental]

What I will do is replace now that the links are standardized , replace https://www.erlang.org/ with /

when go over the redirect rules and find (/download|/course) etc and prefix them with https://erlang.org that way we are explicit as when we redirect out of the site. WDYT?

So in other words, do no rely on _redirect rules for links within the site, and only leave those rules from incoming links from the outside

[eksperimental]

Once done we can run a link checker throughout the site and find broken links and fix them manually.