Hacker News Reader

Comment by userbinator | Hacker News Reader

developers grouse about the other browser makers who are “holding the web back.”

Indeed, perhaps "holding the web back" is a good thing if it means websites will be more accessible overall, to even less common browsers like NetSurf, Dillo, and all the other text-based ones. IMHO the "feature war/race" between the major browser vendors has had an overall negative effect, as if all sites somehow need to turn into ridiculously bloated web apps instead of the simple and far more accessible hyperlinked collection of pages they once were. Keeping the browser choices diverse is a good thing, even if it means they will all display things slightly differently --- just find the lowest-common-denominator and emphasise the content, the stuff that people visiting sites really care about.

There's been some other related discussion on this topic recently:

https://news.ycombinator.com/item?id=15836027

https://news.ycombinator.com/item?id=15943689

View on HN

>developers grouse about the other browser makers who are “holding the web back.”

Oh get off my lawn! Anyone to complains about Chrome vs. FF Quantum vs MS Edge doesn't know or remember the dark of days of IE 5 and IE 6. The lowest common denominator these days is exponentially better than it has ever been.

Slight tangent, but I was happy to see the other day the Google hangouts works on Firefox again (I don't know if that is coincidence or related to the Quantum release or if Google finally decided to support WebRTC).

doesn't know or remember the dark of days of IE 5 and IE 6.

I remember that when IE6 was first released it was the fastest, most reliable, most standards compliant and feature rich browser. That's why it took over the corporate world. It was the Chrome of its day, if you will. How did that work out?

There was an incentive for MS to hold the web back, which isn't there now, although there's an incentive for Google to wall it up.

They released a good product, but weren't prepared to keep it up-to-date with the expectations that a de-facto industry standard should be able to handle.

The web kept on moving after IE6 was released - XHR, jQuery, and later on HTML5 and CSS3. Being a browser entirely developed and maintained by a private corporation subject to the shareholders priorities more than the Acid3 test results, Microsoft decided to ignore the calls to adhere to the new standards for as long as it could, and started developing its own internal awful standards to patch the lack of adherence to the W3C standards (ActiveX, VBScript, Silverlight...).

When Microsoft realized that a privately held company alone couldn't win a fight on your-own-propretary-standards Vs. what the rest of the world wants, it was too late. IE had become an unmaintainable patched monolith that had seen its market share drop from 90% to 20%, losing to the new rivals. They had to basically trash the old code and start from scratch with Edge to still be vaguely competitive.

Microsoft's story taught us that you can't push for your own standards while ignoring what the community wants and expect to survive for a long time. Google seems to have learned that lesson over these years (that's why Chrome still dominates the market), but if they start favoring their internal corporate priorities over the most pristine standard adherence, they might drift into the Microsoft case ad well.

For the record, the concept of XHR/AJAX originated at Microsoft, although it was not originally called that, and was first implemented in IE5.

https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequ...

Ditto for DOM Storage --- before the W3C, IE5 had userData.

The main issue with IE6 was that it was Windows only, making life miserable for those not using Windows. Chrome is on Linux and OSX as well. So there isn't really the "I can't access this site because I don't use Windows" problem that was a real problem then. Even if your favorite browser is Firefox, Edge, or Safari, you can still have Chrome on your system for Chrome-only sites.

… so the main issue is that Chrome is Chrome only so "I can't access this site because I don't use Chrome". Maybe a bit better than being Windows only, but not that much, in my opinion. I don't want Chrome on my system.

I was with you right up until the last sentence. However there are a few - though admittedly niche - platforms that Firefox has been ported to but which Chromium has not.

> I remember that when IE6 was first released it was the fastest, most reliable, most standards compliant and feature rich browser. That's why it took over the corporate world.

Most companies run Windows by default, that's why IE took over the corporate world.

Most companies run Windows by default, that's why IE took over the corporate world.

Netscape's plan was to so totally abstract the underlying OS that it didn't matter what it was.

I think there may even be a chicken and egg situation going on - for some reason, media websites such as CNN are pushing giant rich media content objects at us like videos, which obviously are meaningless to a text browser. Even webdev blogs are being posted with big GIFs every paragraph or two.

I'm not sure why content creators are feeling the need to go multi-media - perhaps it's getting harder and harder to tell who your audience is, or segment your audience, or maybe the "smearing" of society is causing an audience for a given topic (ES6 development, say) to have so many different desires it's impossible to create content on the subject without annoying some portion of them (either by having plain HTML and CSS for your blog post with no helpful GIFs, or a bunch of GIFs that piss off people in text browsers or slow connections).

Even on my iPhone I can’t browse CNN while simultaneously listening to music or the radio (iPlayer). Even though the videos are silent until clicked, they still take over the audio. Gets me every time!

What would a better experience be on iOS? I personally wouldn’t find mixing the audio and having to pause one or open a per-app volume mixer control panel to be a better implementation.

Not sure, but to have this happen even for silent videos seems particularly annoying

From what I know, even "silent" videos can have an audio stream, which just happens to have all its samples decode to 0 (or nearly 0.)

Videos embedded into a webpage have a "muted" boolean attribute, so the nice behavior would be to not pause the music player when muted video starts playing.

Play them with no sound by default if you are already listening to something.

"For some reason"

I think we should probably expect the Cable News Network to traffic heavily in video, yes?

If you've ever watched History or The Learning Channel, you might not expect much anymore.

Or if you're still a subscriber of American Telephone & Telegraph's services.

Back when going multimedia was very primitive and was clearly just a shiny feature with no direct benefit to anyone other than for the creators to partake in new world, I'd have agreed with your statement. In the current situation of muted Facebook videos being a major form of (dis)information distribution, everyone can point to all-encompassing multimedia communication as a potentially extremely lucrative medium. Therefore there shouldn't be any surprise as to the prevalence of that type of content.

This leads to some perverse incentives, though...I have started to see like-farming "tag your friends" pages will post a video of a static image, overlaid with a subtle movement effect like falling snow, undoubtedly because someone realized they'll get more eyeballs from the all-knowing News Feed algorithm if their content is video.

They go multi-media for preroll ads. These are contextually more resistant against adblockers, and bring in more revenue.

> I'm not sure why content creators are feeling the need to go multi-media

I was around for BBSes and the early public internet, and I do appreciate text interfaces... that said, I've always felt there was something quite magical about graphics and audio. The web allows the author quite a bit of freedom in this regard, so it ends up pretty chaotic and uneven.

I find that graphics help to get a point across quickly, even for folks who are very comfortable with tons of text (most people in the world really aren't, and the language you prefer is likely not their first anyway so it's quite a bit of work for them)

It can be misused just like anything else, and tastes can be subjective. It's nice, then, that web browsers being at their core agents for the user, you can also get a fair bit of control on what you want to block by default, or have animations be click-to-play, and so on, which override the author's design.

I find myself thinking of a line regarding early computer games development.

About how whenever game devs got a new toy to play with (faster CPU, better audio, etc etc) for years the games released would be heavy on bling and lite on content.

I suspect something similar happens in other parts of the computing world, and the web has long since hit its tech equivalent of "eternal september". Meaning that these days there are so many new "toys" coming into the web world that people can't help include them into whatever they are making, even if it makes little to no sense to do so in the long run.

i really wish the standards bodies would make a concerted push for a solid advance in CSS (flexbox and css grids are pushing that way but are relatively complex), so that we could get rid of DIV-itis. it's like the TABLE-itis of the past, but two letters shorter (i'm exaggerating, but still...). that would not only clean up the html but make both learning and development faster.

they should also really advance basic web functionality, like making contenteditable more flexible to build text editors on top of, and adding more (and more stylable) form controls (e.g., a good date-time picker, a true omni/combo box, etc). forms are all over most websites, so let's make those faster and less error prone to develop without a bunch of extraneous javascript handlers or resorting to formidable js frameworks.

in short, let's keep the learning curve shallow for web development--that's what made it great in the first place.

> flexbox and css grids are pushing that way but are relatively complex

Can you describe a concrete solution that is less complex than flexbox while still managing to achieve its goals?

CSS Working group person here.

We're trying. Grid (also Flexbox, but mainly grid) should be a large part of the solution, as it significantly reduces the need for divs that are just there for layout purposes. display:content should also help disconnecting your markup needs from your styling needs somewhat (go shout at browsers if it's not coming fast enough). We've also recently decided to add multiple borders, so that you don't need to add a bunch of nested divs just to have nested borders.

The last piece of the puzzle would probably be to be able create boxes (or trees of boxes) without markup, and inject content from the DOM into these boxes. That would be fantastically useful, and has been explored before, but it turns out it is a really hard problem. The first attempt at this is css-regions, which ended up being rejected, in part by Google (because they thought the complexity needed for the implementation was excessive), in part by Mozilla (because they thought the design didn't fit well with how everything else works and would break in too many cases. Also, complexity).

A more modest attempt has been outlined here https://drafts.csswg.org/css-overflow-4/#fragmentation but it is still only an early draft, and solving that isn't trivial either.

the problem is css is tied to and gets it's hierarchy from the html elements. In order to truly separate style and layout from content you need a separate hierarchy and that would be a huge change

https://games.greggman.com/game/css-grid-fail/

You could use XSLT which was supported by Firefox and IE over 10 years ago.

XML had it right with XSLT. It stands for Extensible Stylesheet Language Transformations. Because the language was clearly defined from the start for data not presentation the need for structure transformation was apparent.

There’s absolutely no reason why HTML5 shouldn’t reinvent something similar. Make it much simpler, just reuse the existing Web Template spec, which simply defines new elements with slots and selectors.

Example:

[ was super long so I moved it: https://ipfs.io/ipfs/QmQUszZRdHfdVUZwvQL2tK12XrW54TdJJqXEkm1... ]

More complicated transformations, like the toc here (arguably not that complicated but whatever), are better handled by user Javascript. Because let’s be honest XSLT like transforms are too hard.

Not immediately obvious is the fact that browsers could simply replace and rerun Javascript on page navigation, instead of rebuilding the entire page, if the templates match.

There are in fact many optimizations possible with this scheme. You could mark some regions that share a template as shadow DOM etc.

---

You can also provide polyfill for this TODAY using Service Workers. It’s not that crazy.

> means websites will be more accessible overall, to even less common browsers

Not to mention more accessible to people with disabilities. A forum like wc3 needs to make sure that new features works for everyone and all use cases. Call me judgmental, but I doubt that Chrome developers have the will or ability to foresee all such cases when whipping up cool new features.

In fairness: modern web browsers and platforms are wildly better for users with impairments than the historical baseline, and the accessibility story for many kinds of hot new shiny-shiny is to straightforwardly ignore the content.

I mean... at least in it's prototypical form new kinds of multimedia content, previously unavailable, can continue to be unavailable without degrading the existing content. To the degree that blind people, for example, want to enjoy VRML2020 it's not unreasonable to wait a few release cycles for specialized support and have that content ignored until such a time.

General accessibility of content, ie a preference for flash over HTML, is a content provider issue, not a browser issue.

Also: the Chrome team builds the browser that runs most of the web clients on the world (and soon in history)... it's been a leader in accessibility and standardization for years. Judge as we will, they're better positioned than most to analyze consequences and the market, and have a notable track record.

People investing money in the web are the ones who want to turn it into another sales vehicle.

I'm not sure but I suspect the DOM is the reason small browser implementations can't catch up. I think over the years specifying interactions with JavaScript and CSS have led to ambiguities in the standard and undefined behavior.

The difficulty of implementing the Web platform comes from many things, from the difficulty of implementing JS to the complexity of the DOM to how subtle CSS is to complex, far-reaching APIs like ServiceWorker and WebGL to unspecified behavior depended on by sites. There is no one reason.

I will say that the DOM is one of the better-specified areas of the Web platform. CSS 2.1, for example, is significantly worse. And the table layout specifications are in a miserable state (one of the reasons why it drives me crazy when people suggest going back to table layout for "simplicity").

Chrome is not the standard, BUT WebKit is!

About 90% of all users worldwide use a WebKit(/Blink/KHTML) based browser on all platforms (and mobile is a lot bigger than desktop).

Firefox/Servo, M$ IE/Edge, Dillo, Link, etc are minor browsers, I hope sites will continue to work almost okay with them too.

For traditional enterprise the new IE6 is IE11, it's still going strong in Win7-Win10. You can thank M$ that they forked IE11 trident engine and named it Edge instead of IE12. And IE11 is supported until at least 2020.

Reply on news.ycombinator.com