5 ms·

Wow, this is awful. Would you mind telling me what combination of settings/add-ons you have so that we can reproduce this? Our "no JS" fallback styles are not

Wow, this is awful. Would you mind telling me what combination of settings/add-ons you have so that we can reproduce this? Our "no JS" fallback styles are not done in a very nice way at the moment: they rely on the browser reading noscript tags. We never did any testing with NoScript or similar add-ons, and I remember hearing somewhere that those sort of add-ons don't always evaluate noscript tags.

afandian 9y ago
Thanks for your reply! - Chrome Version 57.0.2987.133 (64-bit) - No-Script Live 0.2.4 http://mybrowseraddon.com/noscript-lite.html http://mybrowseraddon.com/noscript-lite.html - Privacy Badger 2017.4.19.1 FWIW Privacy Badger isn't blocking anything, I think this is all from script blocking. I notice that low-res images get sent by default, which suggests that the design conflates 'no script' with 'mobile'? That's not true at least in my case. Here's my browser log. https://gist.github.com/afandian/b97fc7bb9f5e15f924acb7703220df42 https://gist.github.com/afandian/b97fc7bb9f5e15f924acb770322...
- corobo 9y ago
  > I notice that low-res images get sent by default, which suggests that the design conflates 'no script' with 'mobile' Could also be javascript-powered lazy loading. Load the low res using HTML then swap out high-res when the user scrolls it into view
  - afandian 9y ago
    I'm talking about the thinking that led to this, not the mechanism by which it works. Why would you send a low resolution copy of the image first? Probably because you assume someone is on a low-bandwidth connection.
    - corobo 9y ago
      To make the page load faster (the topic of this post)
- wldlyinaccurate 9y ago
  Thanks! We'll take a look at this ASAP. In the mean time, I guess all I can do is apologise for the broken experience... We use a technique called cutting the mustard to determine whether a browser is modern (read: >=IE9). This allows us to write one set of CSS that should work in pretty much any web browser (we call this the "core" experience), and another set that uses modern-ish CSS (the "enhanced" experience). One of the major drawbacks to this technique is that the technique relies on JS — it is literally just modernBrowser = ('addEventListener' in window && 'querySelector' in document && 'localStorage' in window). So when you browse the BBC News website with JS disabled, you are getting a page that has been styled to work on a Nokia E65, BlackBerry OS4, IE8, and everything in between. Unfortunately right now we don't have a way to give you the enhanced version of the page without JS. In the future it might be possible for us cut the mustard at the server (or the CDN edge) based on User-Agent but that isn't something I can see happening in the foreseeable.
  - afandian 9y ago
    It's great that you're trying so hard to accomodate such a broad landscape; of course there will be trade-offs. Being a weird intentional corner-case you get what you ask for, I suppose! Thanks for engaging with questions.
kevingrahl 9y ago
Looks pretty much the same here though a slightly different setup: - Chrome Version 58.0.3029.81 beta (64-bit) - uMatrix 1.0.0 - uBlock Origin 1.12.1 - Privacy Badger 2017.4.19.1
UnquietTinkerer 9y ago
I found that the "new" site looks reasonable with javascript completely disabled (i.e. turn it off in the browser settings). The problem appears to be when using a plugin to selectively restrict javascript (I use uMatrix) - in this case, the browser apparently doesn't interpret <noscript> tags. This is unfortunate, but I think its more of a bug/limitation in the script blocking plugins than a problem with your design.

afandian 9y ago

Thanks for your reply! - Chrome Version 57.0.2987.133 (64-bit) - No-Script Live 0.2.4 http://mybrowseraddon.com/noscript-lite.html http://mybrowseraddon.com/noscript-lite.html - Privacy Badger 2017.4.19.1 FWIW Privacy Badger isn't blocking anything, I think this is all from script blocking. I notice that low-res images get sent by default, which suggests that the design conflates 'no script' with 'mobile'? That's not true at least in my case. Here's my browser log. https://gist.github.com/afandian/b97fc7bb9f5e15f924acb7703220df42 https://gist.github.com/afandian/b97fc7bb9f5e15f924acb770322...

corobo 9y ago
> I notice that low-res images get sent by default, which suggests that the design conflates 'no script' with 'mobile' Could also be javascript-powered lazy loading. Load the low res using HTML then swap out high-res when the user scrolls it into view
- afandian 9y ago
  I'm talking about the thinking that led to this, not the mechanism by which it works. Why would you send a low resolution copy of the image first? Probably because you assume someone is on a low-bandwidth connection.
  - corobo 9y ago
    To make the page load faster (the topic of this post)
wldlyinaccurate 9y ago
Thanks! We'll take a look at this ASAP. In the mean time, I guess all I can do is apologise for the broken experience... We use a technique called cutting the mustard to determine whether a browser is modern (read: >=IE9). This allows us to write one set of CSS that should work in pretty much any web browser (we call this the "core" experience), and another set that uses modern-ish CSS (the "enhanced" experience). One of the major drawbacks to this technique is that the technique relies on JS — it is literally just modernBrowser = ('addEventListener' in window && 'querySelector' in document && 'localStorage' in window). So when you browse the BBC News website with JS disabled, you are getting a page that has been styled to work on a Nokia E65, BlackBerry OS4, IE8, and everything in between. Unfortunately right now we don't have a way to give you the enhanced version of the page without JS. In the future it might be possible for us cut the mustard at the server (or the CDN edge) based on User-Agent but that isn't something I can see happening in the foreseeable.
- afandian 9y ago
  It's great that you're trying so hard to accomodate such a broad landscape; of course there will be trade-offs. Being a weird intentional corner-case you get what you ask for, I suppose! Thanks for engaging with questions.

kevingrahl 9y ago

Looks pretty much the same here though a slightly different setup: - Chrome Version 58.0.3029.81 beta (64-bit) - uMatrix 1.0.0 - uBlock Origin 1.12.1 - Privacy Badger 2017.4.19.1

UnquietTinkerer 9y ago

I found that the "new" site looks reasonable with javascript completely disabled (i.e. turn it off in the browser settings). The problem appears to be when using a plugin to selectively restrict javascript (I use uMatrix) - in this case, the browser apparently doesn't interpret <noscript> tags. This is unfortunate, but I think its more of a bug/limitation in the script blocking plugins than a problem with your design.