Why people minify assets and not the HTML?
Why do people suggest minifying web assets, such as CSS and JavaScript, but they never suggest the markup be minified? CSS and JavaScript can be used on many various pages while the markup gets loaded each and every time, making minification of markup far more important.
Solution 1:
The answers written here are extremely outdated or even sometimes does not make sense. A lot of things changed from old 2009, so I will try to answer this properly.
Short answer - you should definitely minify HTML. It is trivial today and gives approximately 5% speedup. For longer answer read the whole answer
Back in old days people were manually minifying css/js (by running it through some specific tool to minify it). It was kind of hard to automate the process and definitely required some skills. Knowing that a lot of high level sites even right now are not using gzip (which is trivial), it is understandable that people were reluctant in minifying html.
So why were people minifying js, but not html? When you minify JS, you do the following things:
- remove comments
- remove blanks (tabs, spaces, newlines)
- change long names to short (
var isUserLoggedIntovar a)
Which gave a lot of improvement even at old days. But in html you were not able to change long names for short, also there was almost nothing to comment during that time. So the only thing that was left is to remove spaces and newlines. Which gives only small amount of improvement.
One wrong argument written here is that because content is served with gzip, minification does not make sense. This is totally wrong. Yes, it makes sense that gzip decrease the improvement of minification, but why should you gzip comments, whitespaces if you can properly trim them and gzip only important part. It is the same as if you have a folder to archive which has some crap that you will never use and you decide to just zip it instead of cleaning up and zip it.
Another argument why it pointless to do minification is that it is tedious. May be this was true in 2009, but new tools appeared after this time. Right now you do not need to manually minify your markup. With things like Grunt it is trivial to install grunt-contrib-htmlmin and to configure it to minify your html. All you need is like 2 hours to learn grunt and to configure everything and then everything is done automatically in less than a second. Sounds that 1 second (which you can even automate to do nothing with grunt-contrib-watch) is not really so bad for approximately 5% of improvement (even with gzip).
One more argument is that CSS and JS are static, and HTML is generated by the server so you can not pre-minify it. This was also true in 2009, but currently more and more sites are looking like a single page app, where the server is thin and the client is doing all the routing, templating and other logic. So the server is only giving you JSON and client renders it. Here you have a lot of html for the page and different templates.
So to finish my thoughts:
- google is minifying html.
- pageSpeed is asking your to minify html
- it is trivial to do
- it gives ~5% of improvement
- it is not the same as gzip
Solution 2:
One likely reason is that markup typically changes MUCH more often, and would have to be minified for every page load. For instance on a given Stack Overflow page, there are timestamps, usernames, and rep counts that could change with every page load, meaning you would have to minify for each page load as well. With "static" files like css and javascript, you can minify far less often, so in the minds of some, it is worth the work up front.
Consider also that every major web server and browser support gzip, which compresses all of your markup (quickly) on the fly anyway. Because minifying is slower and much less effective than gzipping anyway, webmasters may decide that minifying for every page load isn't worth the processing cost.