Pretty soon, it will be faster and more efficient to just send us prerendered bitmaps.
Show this thread
"Why do I need a 4Ghz quadcore to run facebook?" This is why. A single word split up into 11 HTML DOM elements to avoid adblockers.pic.twitter.com/Zv4RfInrL0
-
-
-
For those of you who are just tuning in, the code in the image above is the equivalent of parents spelling out words so their kids can't understand it. Except kids is the adblocker and Facebook is your parents. It's just terribly inefficient and slow.
Show this thread
End of conversation
New conversation -
-
-
How do screen readers parse this (if they do/can at all)?
-
I can't imagine any screen reader smart enough to parse this. The crazy part is that this method also prevents the use of ".innertext" which normally foils this type of splitting.
-
How does it prevent innerText?
-
Misspoke, it doens't prevent it, but makes the data it returns useless .innertext returns "SpSpSononSsosoSredredSSS".
-
Oh man, I didn't even notice the duplicates. Wild.
-
Perhaps one day we'll have an AI-driven browser extension that detects ads based on what they look like. Then that mandatory "Sponsored" label would become the achilles heel of all web ads.
-
Until they start shifting pixels to fool AI networkshttps://www.theverge.com/2017/11/2/16597276/google-ai-image-attacks-adversarial-turtle-rifle-3d-printed …
- 5 more replies
New conversation -
-
-
I recently saved out the rendered HTML of the initial load of the Twitter timeline, just strictly from pressing Enter on the URL bar and letting the page load. The _rendered HTML_ is 2,025,102 bytes ...
-
Does that include media, css and js?
-
no that’s strictly just the markup being rendered on the page! it’s because they add insanely huge amount of data attributes to everything, even with embedded JSON…
-
Twitter’s notorious for that. Dumbfounded at the amount of metadata exchange for the simplest processes while building for iOS.
-
I can stop wondering why my laptop is having such a hard time when I just have a single tab with Twitter running.
-
Even worse with mobile safari. I knew something was up as other media heavy sites render smoothly.
-
I just checked, and a screenshot of the timeline in png (lossless compression!) is only about 260K (720p screen...) but the data transferred during F5 (using caching already!) is ~2.4M. So I guess the next step is to code everything in webassembly, and just ship pixels indeed. :(
-
I guess that's where adblocking would need to go anyhow. Personally, I'd like to believe people will just stop caring about twitter, fb, etc. Shoutout to the scuttlebutt folks :)
End of conversation
New conversation -
-
-
You need a quad-core 4GHz CPU to run facebook, because... Object Oriented Programming. The facebook code has 18,000 classes. 18K *object instances* is bad enough, but 18K *classes*? OOP put the state of technology back two decades.http://quellish.tumblr.com/post/126712999812/how-on-earth-the-facebook-ios-application-is-so …
-
36K functions would be better?
-
A competent programmer wouldn't need a gazillion lines of code to implement FB functionality. A client app for sharing pics and messages is NOT ROCKET SURGERY.
-
A competent programmer wouldn't blame a programming paradigm.
-
That's an odd way of looking at it. Why should the paradigm be above criticism?
-
I don't see any criticism, do you?
-
Bram opened the thread by criticizing OO. Since then, precisely zero details or facts about paradigms have been discussed. So I agree, I see no criticism, just a bunch of empty snark.
-
You are quite right, of course. But the "detail" I highlighted was that OO *can* get you in a situation where a client-side social media app ends up with 18K classes. So I see this as an *example* where OO is the enabler for overly complex engineering.
- 3 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.