Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A faster BBC News front page

wildlyinaccurate.com

Post reply on HN
Print view View original

71–80 of 133 posts

Re: A faster BBC News front page

#71

Post by butler14 »

ralfn wrote: post #65

Until the internet marketing team loads it up with scripts that inject 12 different versions of jQuery to do some a/b testing. I wish engineers would actually be able to take responsibility for the whole, and dismiss marketing tags based on technical grounds.

they can and often do -- code your own bespoke lightweight A/B testing software

that'll be the easy bit, of course, the hard bit will be influencing people to part with the £silly amount of money it'd likely cost vs. off the shelf alternatives

Re: A faster BBC News front page

#73

Post by talldan »

zb wrote: post #46

This does feel a lot faster, but hasn't fixed my pet peeve (on bbc.com): clicking on a link to a news article and getting taken to the wrong one because the ads loaded and everything moved just as I clicked. This often happens to me 3 or 4 times in a row. Space needs to be blocked out for ads as soon as the page renders.

Yeah, I get the same issue with another part of the page - the little 'bell' icon they use for notifications at the very top. It appears just as I'm about to click on the link for 'sport', and I end up in 'news' instead.

Maybe I should just read more news.

Re: A faster BBC News front page

#74

Post by afandian »

wldlyinaccurate wrote: post #66
afandian wrote: post #36

Earlier quoted context omitted.

Thanks for your reply! - Chrome Version 57.0.2987.133 (64-bit) - No-Script Live 0.2.4 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…

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). On…

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.

Re: A faster BBC News front page

#75

Post by DanBC »

untog wrote: post #58
afandian wrote: post #40

Earlier quoted context omitted.

I thought as much. My tweet that contained the screenshot: https://twitter.com/joewass/status/788071468131295234 Maybe you could resolve the conflict by including a "don't show any more breaking news" button and then taking the data back to Editorial. Or analyzing the clickstream to see who is being shown breaking news that they already saw, or they're actually reading at the time.

> Maybe you could resolve the conflict by including a "don't show any more breaking news" button and then taking the data back to Editorial. I suspect an optional button would not be very popular, so the data wouldn't achieve the aim you want.

Maybe include it in either [settings] or [Customise your Homepage]

Re: A faster BBC News front page

#76

Post by DanBC »

nailer wrote: post #30
wldlyinaccurate wrote: post #21

Earlier quoted context omitted.

The breaking news banner is a source of much internal conflict, at least within the engineering teams. I think most people agree that it's too big, and most people also agree that it's over-used. Unfortunately it's one of those things that editorial will not back down on — they must be able to push their breaking news to people, and it must be prominent (and therefore quite intrusive). There is a small improvement to…

Who cares if editorial thinks it matters and you think it makes things worse? Act like engineers, work out what you want, test it and get some damn data on it.

You're downvoted because you display ignorance of what it's like to work within a huge rigid bureaucracy where "just test it" mostly can't happen.

Re: A faster BBC News front page

#77

Post by inimino »

malza wrote: post #34
wldlyinaccurate wrote: post #21

Earlier quoted context omitted.

The breaking news banner is a source of much internal conflict, at least within the engineering teams. I think most people agree that it's too big, and most people also agree that it's over-used. Unfortunately it's one of those things that editorial will not back down on — they must be able to push their breaking news to people, and it must be prominent (and therefore quite intrusive). There is a small improvement to…

It reminds me of the early 2000s when websites would initiate a pop-up ad and the first thing I would do is automatically close it without looking. Thankfully browser makers stepped in and blocked them automatically. That's my hope. That news outlets will spam the feature so much that web browsers will make it easy to switch them off.

Won't happen. Pop-ups were a specific browser API and could be easily turned off without affecting anything else, banners are just content and can't be singled out by the browsers (unless by an adblock-type arms race).

Re: A faster BBC News front page

#78

Post by DanBC »

afandian wrote: post #11

Can you ditch the 'breaking news' whilst you're at it too please? It's never really 'breaking' and is usually the headline of the landing page anyway. Usability wise, it's as bad as a pop-up ad. Like this, for example: https://pbs.twimg.com/media/Cu_KsFAWgAADCJR.png

Normally I'd just upvote this and not leave a comment, but BBC can't see those upvotes.

Please, BBC, I really dislike the breaking news banner and I'd love the option to turn it off.

Re: A faster BBC News front page

#79

Post by acdha »

DamonHD wrote: post #57
acdha wrote: post #33

Earlier quoted context omitted.

Compression shows every sign of being enabled so there's little downside to any whitespace in the HTML: https://redbot.org/?uri=http%3A%2F%2Fwww.bbc.com%2F https://www.webpagetest.org/result/170424_79_P0Q/1/details/#... If you're not seeing transfer encoding, is it possible that you're behind some sort of proxy – network or local antivirus – which is stripping it?

Always direct. Today is the first time I've seen gzip actually happen though! If only all my other problems were so easily solved by whining! But still: sending, compressing, uncompressing stuff that definitely adds no value and would be trivial to statically omit almost all of is still a downside...

The question you should be asking is why it's there: is saving an extra nanosecond or two of easily-cached compression time really worth having to rearchitect a backend templating or content assembly system?

It's not a bad thing to do but I'd be surprised if they didn't have higher priorities for their engineering staff. If you look at the trace I linked before, note that even if you could eliminate all of the time to generate the page the most you could save is less than 100ms on page which uses multiple seconds of CPU time to render. Look at the video – the page doesn't even render for almost 5 seconds in a low-end browser:

https://www.webpagetest.org/video/compare.php?tests=170424_7...

Chrome is faster but still over 3 seconds:

https://www.webpagetest.org/video/compare.php?tests=170424_N...

If you were in charge of that project, would you commit your limited developer time to something cannot possibly produce a human perceptible benefit or instead have them work on removing a borderline-uncomfortable delay?

Re: A faster BBC News front page

#80

Post by ezequiel-garzon »

React is only used on the server. We do not load it in the browser.

My apologies in advance for my ignorance, as I have so far read about React, but not actually used it. Isn't the main advantage of React the virtual DOM it sets up and its reconciliation algorithm with the DOM? How can this be applied without using on the client side? Wouldn't all the responsive aspects be lost, requiring a round trip to the server for each event?

Thanks in advance.

Post reply on HN