Hey, fedi! I’m gonna hold a 10-minute talk at work about website bloat. How do *you* measure bloat? Is it the number of kB loaded? The speed of rendering? The number of calls made? The amount of JS loaded/used? The image:text ratio? Something else?

I’ve looked at as a starting point, and the oldie but goldie Motherfucking Website. Any other fun or interesting resources you’d like to point me to?

@tinyrabbit Speed of rendering / usage is the only thing that matters. Just make sure you're testing at different Internet speeds.

The thing I see most often that infuriates me is when a site requires Javascript to display content. Not to interact with it, but to display it.

@tinyrabbit number of third party SaaS / adtech / spyware things loaded is probably the thing i rely on most. watching the little counters tick up for all my adblock / privacy plugins is a pretty good index to how terrible an experience a site is going to be overall.

@tinyrabbit For me the term "bloat" suggests that it is not about absolute metrics (speed, size, response time), but about how much better it could reasonably easy be.

@tinyrabbit all of that isn't really a good way to measure bloat, because that depends on the functionality of the site. If a blog pulls in the whole tex ecosystem to display a post, that bloat, but if an html5 game loads 500mb of image assets, that can still be reasonable.

@tinyrabbit the number of different versions of jQuery that get loaded

@tinyrabbit A great example of bloty design is how much Javascript twitter has to load to display like a single line of text (a tweet)

@downey @tinyrabbit The way I measure #javascript bloat is when the whole machine is brought to a dysfunctional crawl to where mouse movement is 1 pixel per hour, and cntl-alt-F4 can't even get a terminal, the HDD LED is solid, and the only escape is to cut the power. Possible solution:

@tinyrabbit @downey Of course, the problem with that form of measurement is that when the measurement is taken, it's no longer useful. I never know which scumbag's javascript brought my machine down.

@downey @tinyrabbit BTW, shift-ESC in #UngoogledChromium gives you a task manager which shows the RAM & CPU consumption. When a tab is using over 100mb (which is quite often) I ask myself if I can live w/out that tab. I wish tabs that hit over 100mb would light up red, so I don't have to manually check the task manager. When things go bad, it's when I didn't think to have a task manager running.

@tinyrabbit I measure bloat in needlessness, if that makes sense. what I mean is, how much extra, pointless stuff has been added that has no bearing on the desired goal of the website? measured against the weight or cost of its implementation. html and css are perfectly suitable for the majority of web content. chiefly, most webpages should just be that: pages. text or information being provided upon access. at the end of the day the goal should be delivering information. widely, however, websites are essentially becoming applications unto themselves and browsers basically serving the purposes of operating systems. I think technology such as javascript and others are fine, but they should only be used if necessary—as tools. not abused and vomited everywhere almost as a means unto itself.

it's a breath of fresh air coming to a simple, old-school, clean html website devoid of scripts or cookies or trackers…just a web page delivering information and perhaps stylized or formatted with css. they're lightweight, they load quickly, they don't consume resources, and they serve their intended purpose efficiently.

web bloat also brings up greater ramification such as the need for faster and faster data connections with ever-growing storage and hosting costs. this renders older technologies obsolete in many respects while requiring greater resource consumption. it prohibits access to users with slower connections. the way the internet has gone has widely been down a path of pointless inefficiency, wastefulness…bloat. and, in most cases, for no good reason.

@kino @tinyrabbit

I saw a good description recently, of "applications which happen to look like documents". If it renders as a plain page of text, it shouldn't load several MB of Javastuff to in order to generate that : it should just load THE TEXT.

@tinyrabbit I generally go on bytes over the wire for the initial load - enough to fail most sites these days

@tinyrabbit I'm fairly flexible but generally I'd expect things to load in <500k (compressed) for a fairly boring page of content

Of course I'd expect basic landing pages to be much less and it's reasonable for "web apps" to be bigger

The problem of course is even filtering out images your average “news” site fails

Sign in to participate in the conversation

For people who care about, support, or build Free, Libre, and Open Source Software (FLOSS).