Hacker News Reader

Comment by nindalf | Hacker News Reader

I thought the thread would have more substance to it. His claim is that Google products would have performance bugs or would explicitly block any non-Chrome browser. As a long time Firefox user, I’m with him so far. But he loses me on the next bit where he rules out incompetence and then jumps to org-level malice. I don’t think it’s either. A simpler explanation is that Google simply stopped caring about non-Chrome browsers. People building say, Inbox were told it was acceptable to launch a product that is only accessible to one browser at launch. I hated that decision because it affected me directly but that’s not malice, simply prioritisation.

Of course standard disclaimers apply. I don’t work for Google, never held Google stock etc.

View on HN

Have you tried using the new Gmail interface on Firefox?

It's ridiculously slow, to the point of being completely unusable.

A Google engineer claimed here on HN the reason was that the UI framework uses some deprecated API that is polyfilled in Firefox but available on Chrome.

Something like that should never have launched, but may have been a somewhat acceptable reason months ago. Now, after being in production for months, not fixing this is either saying "we don't care about those <10% Firefox users" or straight up intentional to force FF users to switch to Chrome.

Either of those amount to the same thing and classify as malicious to me.

I have completely shifted away from web UI to native application for the same reason.

For me it works fine. Whole page reloads in maybe 3 seconds, while browsing mails is pretty much instant.

You're both saying the same thing, but drawing different conclusions:

> Google simply stopped caring about non-Chrome browsers

> we don't care about those <10% Firefox users

Except that GP attributed that to incompetence and you to malice. I guess it's still hard to guess intent :)

Even if it can be rationalised as them not caring, or even just not prioritising support for Firefox, the point of TFA is you have to treat it as malice if you are on the other end. Maybe it's better characterised as neglect, but you have to actively call it out and fight it if you don't want to lose out.

It may be like being crushed by an elephant that's rolling over in it's sleep, but you need to wake the elephant up and tell it to sleep somewhere else.

This is "Shadow DOM v0" API and have been removed in latest Chrome:

https://news.ycombinator.com/item?id=18701765

https://caniuse.com/#feat=shadowdom

Gmail never used Shadow DOM v0.

I don't buy the "oops we dropped the ball on UAT". I was working in a bank once and we were making sure that our web-banking would work properly with EVERY POSSIBLE browser. And I don't mean IE, FF, Opera, Chrome.. I mean some end-of-corridor browsers that you could get from tucows. I don't buy that Google made mistakes like that with one of the top5 browsers. Did they ever pull a trick like that with Microsoft's IE?

According to an M$ intern they did which eventually caused Edge to fail and for M$ to say hey we'll just be a wrapper around Chrome and then they can't screw us.

"Never attribute to malice that which is adequately explained by stupidity."

But since Google's engineers are far from stupid, you may have a point.

This whole article makes me question that statement. I hear it all the time, but passive aggression is a very real tactic in a bloodless war.

> But he loses me on the next bit where he rules out incompetence and then jumps to org-level malice.

I think he has a reasonable justification for his position (two even):

- Google would have to be very incompetent for it to be incompetence (stated in the article)

- If Mozilla comes around every few weeks "hey guys, you broke it again" it's hard to argue they "don't care" or "don't prioritize" in the sense of "oh, we have so many things to do ..."

You even state it yourself: "People building say, Inbox were told it was acceptable to launch a product that is only accessible to one browser at launch."

Yeah. That's active malice if your company produces the one browser they choose. Whether you couch it as "simple prioritization" or not.

At the end of the day I still think the competition agencies should have thrown the law book at Google so hard that it bleeds. Start with a few billion and go up from there. Google had (has) a virtual search monopoly and actively used it to promote their browser. At least in the EU that's illegal. And that the agencies didn't kick Googles ass for it is for me one of their biggest regulation failures.

It's not malice against another browser to build for where the users are. When the users were on IE and Firefox, Google built for those platforms, ignoring e.g. Konqueror or Opera. Now that the users are on Chrome, Google can ignore Firefox. Each product team can independently decide that the best thing to help the most people is to work on more features for 1 browser rather than fewer features for several browsers.

Google does save on the frontend, e.g. they use pixel-oriented design, which is incompetent. Maybe they have competent engineers somewhere deep in the backend.

I was thinking about this comment and I thought of the perfect analogy: accessibility.

Everyone in the abstract would like to make accessible websites. Some people at Google are really passionate about accessibility and advocate for it, while others don't know as much about it or overlook it when building their product. There's surely some organization-level guidelines like "make sure your product is accessible before launch" that have likely changed over time as different people in power have prioritized different things, and with uneven enforcement when it comes time to launch a given product a given team will make a random judgement call as to whether to delay their launch or cut some other feature in order to make the product accessible. (Inevitably someone will protest here with "it's not that hard to make an accessible product, just do X Y Z and test configuration Q, and it's really important". That is true but doesn't change the reality that it has a cost that takes away from other work.)

Now substitute "Firefox compat" for "accessible" in the above and it all still applies pretty much perfectly.

I don't believe there's an anti-Firefox (or anti-accessibility) conspiracy at Google, but rather just that Firefox/a11y compat is not a hard blocker for launches. The net effect is the same, of course, in that it's as if every product somehow discovers a way to break a11y/Firefox, but it's actually just how entropy works -- unless you're continuously watching out for and verifying your a11y/Firefox it's inevitable you'll break it.

(disclaimer: I worked on Chrome but I wouldn't know either way if there is any higher-level conspiracy; as an engineer who has built multiple products in different orgs I have seen a general lack of specific guidelines/requirements for a11y/Firefox)

Even for smaller websites, I don’t think any web developper doesn’t test how the website renders on Firefox and Safari. Not doing so during the development process seems like a deliberate decision rather than an omission from engineers that do not know better.

Plenty and plenty of developers never tested on anything more than IE. And today on anything more than IE and Chrome.

And plenty of developers test on IE and Chrome and Safari For iOS and call it a day.

Its very much not obvious that this is a deliberate decision and not an omission.

This is an excellent analogy.

Personally I'm surprised by Google's attitude to a11y/Firefox because all their products have millions of users. 10% doesn't sound like much, but 10% (Firefox market share) of 50 million is more than the population of many countries.

I guess they rationalise it by saying "our product is new, with 0 users right now. Let's focus on the largest segments of users (Chrome, non-a11y) and ship an MVP. If it sticks, let's expand to Firefox, do the a11y work etc." If I was in that PM's shoes, this makes sense to me. As a user ... it feels wrong.

> 10% (Firefox market share)

Market share, across all platforms, is closer to:

* Chrome: 60%

* Safari 20%

* Firefox: 5%

* IE: 5%

* Edge: 2%

* Other: 8%

If you just mean desktop, though, then yes Firefox is about 10%. But mobile web is serious traffic these days.

I think the incentives to not support accessibility/minority platforms grows as the number of users grow, since the potential benefit of the features you would not implement because you have to work on a11y or Firefox support is multiplied by your number of users.

That's market forces for you.

If you just chase money all the time, you always end up with shoddy crap — there's no incentive to be any better than barely adequate, so you're constantly riding that line.

We only end up with nice things when someone dedicates time and effort _despite_ market forces. Mediocre businesses and their besuited human avatars spend so much time insisting they're “passionate” because even _they_ know that giving a shit is the only way you get anything other than crap.

>But he loses me on the next bit where he rules out incompetence and then jumps to org-level malice.

Google deliberately blocked Windows Phone users from MAPs. They gave bullshit reason that the browser wouldn't work. But it worked for those who knew how to change their user agent to something else.

https://www.businessinsider.com/google-admits-it-was-blockin...

It's stunning to me that in 2019 there are still people on Hacker News that don't believe in Google doing things like that intentionally.

I think Google's early culture and ethos (Don't be evil.) as well as their initial positioning as the anti-Microsoft helped them get a lot of "street cred" among tech workers. It's taking a while for that idealized image to change. I was still giving Google the benefit of the doubt right up until AMP started getting pushed hard.

This lack of care is awfully convenient. That each "oops" gave them more Chrome users might be a happy coincidence at first, but I'm sure some exec noticed the effect and suggested they continue to "prioritise" that way.

The distinction between accidentally building software that performs badly in any browser other than your own and intentionally doing so to deliver a degraded performance to people not using your browser is very small and hard if not impossible to determine from outside.

There is no good case that Google intentionally built the software in such a way that it would be slow in Firefox. But e.g. by not testing the app in Firefox or by explicitly using Chrome-specific optimisations or having the Chrome team create specific optimisations for those web apps (i.e. "good cross-team communication") you can effectively accomplish the same without ever having to spell it out as an intentional strategy.

If you build a browser with a significant marketshare and you also build widely used web apps, not spending any resources on fixing cross-browser compatibility or performance issues is effectively anti-competitive. Deciding not to correct "happy accidents" like "forgetting to test your apps in other browsers" leading to a competitive advantage is as bad as intentionally doing the same from the get-go.

We're talking about an international megacorp that is worth more than half a trillion dollars at this point. There are no accidents, especially not many identical accidents that remain consistent over years.

In my opinion, that is not really separate.

Pretty much all companies constantly whip their employees to get things done as fast as possible. And if that definition of "done" does not explicitly contain interoperability, then even the most well-intending employees will drop that sooner rather than later.

Which is the case for any side-goal. Security, code quality, documentation, tooling, test coverage etc. If you only prioritize one thing, other things will be neglected, unless you put explicit checks in place.

Which is why we have people whose job it is to manage these things. And if those managers failed to put checks in place, then we are back at either incompetence or malice of those managers.

I remember this "oops": http://fortune.com/2018/07/25/youtube-slow-mozilla-firefox-c... use of a deprecated API was slowing YouTube. Others may remember other oopses.

A simpler explanation is Google took a page from the Microsoft monopoly ‘Embrace, Extend, Extinguish’ playbook. It is illegal behavior designed to promote one part of your business by using the monopoly power of another part.

> People building say, Inbox were told it was acceptable to launch a product that is only accessible to one browser at launch. I hated that decision because it affected me directly but that’s not malice, simply prioritisation.

Would it have been acceptable if the one browser was Edge, Firefox or Opera? I bet dollars to donuts that they would not launch until they had it working on Chrome.

Reply on news.ycombinator.com