Slashdot Mirror
Average Web Page Approaches 1MB
MrSeb writes "According to new research from HTTP Archive, which regularly scans the internet's most popular destinations, the average size of a single web page is now 965 kilobytes, up more than 30% from last year's average of 702KB. This rapid growth is fairly normal for the internet — the average web page was 14KB in 1995, 93KB by 2003, and 300KB in 2008 — but by burrowing a little deeper into HTTP Archive's recent data, we can discern some interesting trends. Between 2010 and 2011, the average amount of Flash content downloaded stayed exactly the same — 90KB — but JavaScript experienced massive growth from 113KB to 172KB. The amount of HTML, CSS, and images on websites also showed a significant increase year over year. There is absolutely no doubt that these trends are attributable to the death throes of Flash and emergence of HTML5 and its open web cohorts." If you have a personal home page, how big is it?
Average information content - does a page view give me more insight as a user now than it did 10 years ago?
And.... when running AdBlock Plus, this figure goes down to 100kB. I run AdBlock mostly for the massive speed increase that comes with it.
My fully featured CMS that used jQuery, jQuery UI, and a lot of heavy library takes 140kb. Learn to optimize people!!
This only matters if people go to the first page, and never go to any additional ones.
For most websites these days, you'll take the initial hit from javascript and the 'branding' images when you first get to the site ... but the changing content per page is much lower.
If websites are using standard javascript libraries being served by Google's CDN, then it's possible that someone visiting your page already has jquery, mootools or similar cached and doesn't need to load yet another copy.
I also didn't see if they had any comparison between transferred size vs. used size. (eg, javascript that's sent compressed) ... and as this is from an new archive ... does anyone know if Archive.org could analyze their holdings to see what the longer term trends are?
Build it, and they will come^Hplain.
On most sites that I go to that have a paragraph per page model, I just click the "Print" button/link on the site and they combine the pages for printing. Then I read it without needing to print it. Sometimes they require printing it. If they do, I am less likely to read the article at all.
http://www.the5k.org/
It seemed so long ago. Didn't /. have an entry as well?
This sig contains repetition and redundancy.
Yes, compression helps (and is generally done automatically in any good Apache configuration). What helps even more from a user's perspective is combining files; basically, in the backend we combine all our Javascript and CSS (or as much as is reasonable) into one file instead of serving it as multiple, separate files linked to the current page. This cuts down on HTTP requests massively and speeds site loading from a user's perspective. Yahoo has a great list of best practices for speeding up sites if you're interested.
It's better to vote for what you want and not get it than to vote for what you don't want and get it.
- E. Debs
Forgive me piggybacking here, but I've a web question. I read Slashdot predominantly on my phone (doesn't everyone?), but once you get 5 or 6 levels of replying in, the posts become unreadable. Each reply has a shorter width then the one above, meaning you end up with a handful of characters per line, and the rest of the horizontal space as just that - space. Is that really how it's supposed to be - completely unreadable? Is there no way of overriding it and saying 'look, I know it's a reply due to the context`. I've tried poking around in the various options within Slashdot, but I don't understand what most of them do, and the so-called help is completely useless and doesn't describe what the options mean nor how to use them. I think the problem is that the designers of Slashdot believe everyone is using a monitor so you'd probably need to be about 30 or 40 levels in to get to the same problem.
I'm using Dolphin HD on Android but a friend with an Apple phone has the same problem. Is there an answer?
The phone issue is interesting. I was just on the "How Much Data Do You Need?" page for a local provider. You slide the bar for various things, like how many web pages you visit in a month (as if anyone really knows that). Their assumption was 0.17MB/page. I know there are mobile versions of pages and such, but this still seems like a gross underestimate given this story.
Sorry to be a dick but you're bragging about that page? Really? You know when they say "size doesn't matter"? Yeah - sometimes it also means being as small as possible is not necessarily a good thing. I would have thought that page was trash ten years ago when Geocities webpages were everywhere so, now, it's really not good... Seriously, stop bragging about it and spend some time designing a real page.
Sorry to be a dick - someone had to tell you...
Persistent HTTP connections were tacked on to HTTP 1.0 years ago and are widely supported, but you still have the "can I have that bit now please?" overhead with the associated latency between retrieving each file on each connection. 100ms of latency multiplied by a dozen assets soon adds up. HTTP 1.1's pipelining means you can ask for many things at once so only suffer that hit once (or twice - page then assets), but in practice browser support for pipelining is poor.
Chernobyl 'not a wildlife haven' - BBC News
So a web page today is about 10x bigger than it was in 2003. I can accept that.
But in 2003, I had a baseline 2-megabit-per-second Internet connection and could have had a 3- or 5-megabit connection for a bit more cash.
Today, 8 years later, the "normal" connection speed for my ISP is 6-megabit.
So according to my observations and their statistics, folks are expected to download 10 times the amount of stuff using just 3 times as much available bandwidth.
In other words, the web is currently more than three times slower than it was in 2003.
Hooray!
Kid-proof tablet..