How Browsers Work | Hacker News Reader

How Browsers Work

html5rocks.com

633 pointsmartius36 comments

> According to the StatCounter browser statistics, currently (August 2011), the usage share of Firefox, Safari and Chrome together is nearly 60%. So nowadays open source browsers are a substantial part of the browser business.

Somehow this never occurred to me to aggregate it this way. I recall having read somewhere that the server-side ratio is about 70-30. It's quite a milestone as now both ends of the WWW platform have an absolute majority being open source.

It's hard to lump Safari into that, even if webkit is open source.

Why so?

Because Safari isn't?

Neither is Chrome.

If you use that criterium, you also can’t lump Chrome into that. I would count the engine, not the browser.

But Chromium is, and there's no difference between the two beside branding and Chrome shipping with some non-free stuff: https://code.google.com/p/chromium/wiki/ChromiumBrowserVsGoo...

Chrome isn’t open source. Full stop. If you don’t count Safari you also can’t count Chrome.

I think it’s equally stupid to not count Chrome and Safari.

Why do you insist on it being black or white? You wouldn't agree that more of Chrome is open source than Safari?

More … how? What matters is the engine, who cares about the chassis? It would be truly idiotic to include Chrome but to exclude Safari. That just does not make sense.

Sure, more of Chrome’s chassis is open source (though not really), but who the hell cares about that? All things considered, Safari and Chrome are both pretty open source. What matters, anyway.

I do? I think the parts of the browser that aren't the rendering engine are important too. That being said, I honestly don't really care about the original argument. I agree that Safari should be included as open source.

My point was just that I think Chrome is measurably more open source. I know that's vague and you can quantify by number of lines, whatever. Since Chrome is "more" open source, wouldn't someone be able to fairly draw a line that includes Chrome but not Safari?

In my mind where one draws that line is completely subjective. I think where you draw the line is equally subjective.

The part where you were treating Chrome and Safari as equal in terms of open-souce-ness is what I disagree with.

If all that mattered was the rendering engine, then Chrome would never have gotten more than 1% browser marketshare, and Firefox would reign supreme. It's pretty obvious to anyone that's paid attention over the last few years that the surrounding UI, Javascript engine, and experience encapsulating the rendering engine is just as important, if not more so.

With Chrome basically the entire UI, including preferences, extensions engine, syncing, and automatic and partial updates are completely open source as part of the Chromium project. Chrome merely puts the Google logo on the cover and packages some pieces of software that can't otherwise be distributed due to licensing constraints. Developers don't consider Ubuntu closed-source just because it has the ability to package closed-source software and drivers with it.

In short: there's no way you can consider Safari an open-source project. There are several ways to classify Chrome as open-source, to which many developers agree.

How so? Percentage of code? # of features? If you want to make claims like that then actually make them, don't just wave your hands.

You're arguing over the degree of something which is just silly. Either both count or neither counts. Who cares if 10% of Chrome's code is proprietary and 15% of Safari's code is proprietary, or vice versa? If one counts as an open source browser then so does the other.

The important pieces that actually deal with content are all open so I would count them both, but that's just me. If you want to draw lines and make it sound scientific then you better show us some numbers. You don't get to draw imaginary lines because you just think that's where a line should be, if Chrome is more open tell us how.

I guess you could quantify it with % of code. I've never actually snooped around inside of there, and I'm basing what I was saying off of things I've read here on HN. Maybe I'm wrong, but I've gotten the impression that a larger percentage of the Safari code base is not open source than Chrome.

Saying that any percent >0 causes the software to not be considered open source is also drawing a line. Either way in my opinion it's subjective where you choose to draw it.

I agree that Safari should be included. I was simply disagreeing with the parent, because I got the impression that he or she was saying they are equal. Also, if someone decided they wanted to draw the line somewhere where it would exclude Safari, I would understand where they're coming from, and agree, as Chrome is more open sourcy.

Edit:

> "If you want to draw lines and make it sound scientific then you better show us some numbers. "

I'm not sure where you're getting this. I wasn't trying to draw lines or making anything sound scientific. All I said was that more of Chrome is open source, that's it. Like I said, I feel it's subjective as to where you draw the line, which is why I was disagreeing with parent. I felt that he or she was making an objective claim with how it should be. In my opinion if you insisted on being objective about it, you would probably have to measure the % code or whatever. I don't believe that the objective view on it is treating any software with greater than 0% proprietary software the same.

It depends on what you care about when you say "open". If you care about security/privacy and the fact that you're running potentially Trojan-infected code on your machine, then any non-open code makes a difference.

The fact you use a connection to the internet opens yiu to many nefarious means to "screw you over". I can think of subtle things like changing adservers to backbone providers, router protocol redirections, ssl master cert spoofs (care of us govt), and many others.

Unless you were the types that freaked out on mm256.dat and nsakey, use standard security and you're fine.

If you did freak out, im sure we can work out a GPS-login script that watches location, bloodtype, retinal scan, passphrase, and voice scan to get a guest account.

Google do not promote builds of Chromium since they want users using Chrome. If you want to run Chromium, the open source browser, but do not want to build it yourself, you can download the latest builds (and older builds) for all the different platforms from the buildbot server:

http://build.chromium.org/f/chromium/snapshots/

This looks like an excellent place to link someone who knows a bit of HTML, maybe some Javascript, but is curious about "real CS stuff".

Very interesting. I am always amazed how quickly browsers can update the layout when you work with jQuery to update CSS or the DOM.

The bit about how browsers parse malformed html is very interesting - a useful lesson for all developers having to deal with the crap that users tend to produce! ;-)

It's also a useful lesson for people thinking about security. When something that is not actually HTML can be interpreted in a browser as HTML, there is a big potential security hole.

For example: At one point, I sanitized incoming user content by parsing it as HTML and nuking almost all markup (I allowed links and italics to pass through). THEN I realized this wasn't good enough, because Firefox was parsing unclosed image tags as though they were closed. The (strict) parser I was using would pass through a user's unclosed image tag as plain text, but Firefox would load it as an image. An attacker could have used that to not only collect visitor IPs (by loading an image on his server) but also to do various things with Javascript -- an image tag can include an onload attribute. (I have no idea if Firefox still works this way, it was about five years ago.)

It would make more sense to HTML escape the input than to strip out tags.

The goal was to allow link and italics tag, as I mentioned.

They refer to Tali Garsiel as "she". Sure it isn't a "he" ?

The authors picture looks like a woman to me.

Tali is a woman, correct!

From her bio on Blogger: "My name is Tali Garsiel. I worked for many years as a web application developer in various projects and technologies. Both on the client and server sides. I live in Israel with my husband, 3 daughters and my cat. Yes, I'm actually a woman with a technical blog!" -- http://www.blogger.com/profile/00737791624070565735

Lol. It seems I need to do more web research... Thanks for all the down votes. What a stupid and childish site. Bye Bye Hacker News.

HN has elitism problems: it's excessively worried about quality (something it can't really control as the site grows) to the extent that its group action violently downvotes certain posts--such as your rather innocent question.

I guess everyone thinks you should have googled it, but we didn't have the decency to politely request that you google such information in the future?

I feel you.

It's an irrelevant question and if a discussion were to spawn from it, it would detract from the actual content. This is a perfectly valid comment downvote.

It's a question about the author. It's a stupid question about the author that could easily be answered with five seconds on Google, but that doesn't make it irrelevant, except in the eyes of an indifferent, exclusionary mindset that prizes signal-to-noise-ratio above simple politeness. I ain't asking for people to _upvote_ it, just to ease off on the downvotes and maybe provide a courteous explanation.

The question was a sexist double whammy (given both the assumption that "the author of a technical document can't possibly be a woman" and that "the author's gender is at all germane to this piece") that has no place anywhere, let alone HN. If anything should be downvoted, that is the perfect example of one such thing.

Or maybe the speaker is from a culture that doesn't know that Tali is a Hebrew name for both boys and girls, or from a culture where Tali is only a dude's name, or... the hypothetical exceptions to your interpretation go on and on. [Ninja-edit] Your ninja-editted assumptions as to what the person meant by his question are an over-reaction: you don't _know_ that's what he meant, you're only seeing hints of it.[end ninja-edit]

It could _well_ be an innocent question, and there's no sense in generating righteous indignation over two sentences.

Given that female pronouns were used several times in the opening paragraph, even if one didn't scroll down to the bio at the bottom, to assume that she must be a he, even on the basis of something such as a name (cf. George Eliot) is entirely unwarranted without further reason to back it up. But, you're right, it could have been a case of bad assumptions and only single-whammy sexism (the author's gender is still not germane to the article in any way). And it's still a downvote-worthy comment.

You made a benign claim about the article that was incorrect, and in the interest of not confusing other HNers your comment was downvoted, why are you taking this personally?

thanks for the link!

Reply on news.ycombinator.com