[laumars]

That depends on the popularity of your site and how the content is generated. If it's a relatively static site (eg the pages are generated into flat HTML files before uploading) and/or you only get a few visitors a day then you can probably get away without any server side caching.

However if you're running a relatively heavy site for content that doesn't change regularly (eg many CMS solutions like WordPress) then you'd want to do caching to lesson the burden on your web servers. And if you run a popular site then caching will reduce your bandwidth and thus allow you to serve more traffic on existing (virtual) hardware. Both those reason will reduce your hosting costs in the long run plus make the site feel snappier (thus improving its user experience)

However browser caching is something I'd recommend even on quiet and static sites because that will have a direct impact on the speed the site loads and can also potentially save your visitors bandwidth (this really matters if you're targeting people from less privileged backgrounds or where internet costs are extortionate).

Personally I view caching as important part of the web development and infrastructure design process as writing the code and building the web servers.

[notforgot]

> you'd want to do caching

That's not what I meant. I understand caching helps.

What I meant is will there ever be a way for caching to happen automatically behind the scenes without me ever having to do anything at all to make all this happen.

For example, the OS caches files, but I don't have to configure the OS file cache by hand. The OS takes care of it.

[laumars]

Your example is pretty bad for a number of reasons:

1) If you're building a high availability file server then you would expect the sysadmin to configure the way the OS caches files rather than run with the defaults. Likewise, if you're building a busy site then you'd need to configure caching to fit your specific application.

2) The OS file cache can run with pretty basic defaults because file system files are static (yes they can change, but you have to go through the kernel ABIs anyway so easy to track changes). Website content cannot be guessed upon because even static content can and often is dynamically generated. No assumptions can be accurately made. This is also why there's so many different levels of caching that happen on busy web sites.

3) File caching on the OS only needs to happen at one place (as touched on in previous point). However websites are built from a plethora of different frameworks which are literally far too numerous to name.

That all said, there will be some specific web frameworks which will ship with caching defaults (more typically in the case of browser caching) and some web applications will ship with recommended plugins for enhanced caching (eg Redis / memcached). However pragmatically I think if a web developer is smart enough to write code then they should be smart enough to implement caching. In the days of frequent website attacks, the unpredictability of which sites go viral nor when, and the ease of which anyone can build and host a website; I really do think some web developers need to up their game rather than blaming the complexity of the tools nor the lack of sane defaults. Yeah the current web model is a mess of edge cases and hidden traps, but if you're a developer then there's no excuse not to properly learn the tools you've been given regardless (or maybe especially because) of how poor those tools are at protecting you from sawing your own hand off.