Which web browser is best?

[Albert Wiersch]

And yet another update...

Web Browser Grand Prix: Chrome 25, Firefox 19, And IE10

Looks like Chrome is best - best standards compliance and best performance, and Opera is lagging pretty far behind. Firefox was a close 2nd.

I will use Chrome more on my new Dell notebook because Firefox (my usual browser) displays the browsers tabs ridiculously large (double height I think). I don't know why and couldn't find a quick answer so it's a good excuse to use Chrome more.

[MikeGale]

I've found a few issues with my latest checks.

Chrome "phones home" more than I'd like. (Trace the conversation over HTTP, from it, and you can see what I mean. Some people don't mind, so it may not be a factor for you.) Though I haven't tested it much, I've noticed something called "Comodo Dragon" which is Chrome repackaged without the phone home tendency. If anybody else has tested this more thoroughly I'd appreciate feedback.

I've found that Firefox 19 fails to operate on a fair number of sites. This is things like placing orders (I haven't seen the problem before). Normal traffic fine. This is disruptive, and seemingly quite common. (I have to stop working in FF and go to another browser.) Haven't researched that one. Last time I reported a bug to FF I found the experience unsatisfactory, though ultimately successful, so not inclined to do the work needed. FF also "phones home" more than I'd like, sometimes to Google services!. Again Comodo has a solution to that, "Ice Dragon" this time. Again not evaluated thoroughly, by me, yet.

If you try any of the Comodo offerings. They suggest you use their secure DNS system. (I'm using OpenDNS so I didn't go down that road.) I believe that the versions of the underlying browser engines are a little behind the official Chrome / FF releases.)

I continue to use a variety of browsers. That includes various versions of IE. I find that web sites are getting worse. Some only work in IE, some don't. Some insist of FF / Chrome, some fail in one of those. Some give very different behaviour in different browsers. My presumption is that a fair number of developers are not working cross browser from the start. Maybe they've become used to "apps" so expect things not to be usable on many platforms. Pity they don't integrate decent web checking into their work flow!

I see Opera recently decided to drop development of their own browser engine and have adopted WebKit. The game is narrowing further!

[Albert Wiersch]

[list]Chrome "phones home" more than I'd like. [i](Trace the conversation over HTTP, from it, and you can see what I mean.

I haven't read much about this, but I suppose I shouldn't be surprised. I've kind of just accepted the fact that I'm being tracked in various ways.

I continue to use a variety of browsers.

Me too. You never know when you're going to run into a site that has issues with one browser... and of course I need them for testing.

I find that web sites are getting worse. Some only work in IE, some don't. Some insist of FF / Chrome, some fail in one of those.

I think that is unfortunate but not unexpected. The more complex a site is, the more likely it's not going to work on one or more browsers, or work poorly (like have bugs).

Pity they don't integrate decent web checking into their work flow!

That would definitely help solve some issues. My guess is that sometimes when a developer tries a validator (usually the W3C validator) then

1) it throws up a bunch of issues but they just don't want to spend the time to fix things because it "works" in the browser(s) they quickly (but not thoroughly) tested it in, or

2) the validator throws up issues that they know they want to ignore or disregard and so it makes the validation more of a pain and hassle than a help. CSE HTML Validator tries to resolve this issue by being more practical as well as configurable so developers can eliminate many validator messages that "get in the way" and concentrate only on the more serious issues - but developers actually have to spend (and have) the time to figure this out - or be required to do it by their boss or the law (like for government websites).

I see Opera recently decided to drop development of their own browser engine and have adopted WebKit. The game is narrowing further!

Yes, another interesting development. It has to be an immense amount of work to make and maintain a browser engine these days - considering how large and complicated HTML, CSS, JavaScript, and other web technologies then need to support are these days. On top of that, it has to be rather hard to support that development with funding, considering how good the free browsers are. Google and Microsoft and Apple clearly have a lot of resources, but I don't see how smaller companies with fewer resources can justify supporting substantial software development endeavors for another free web browser.

[MikeGale]

My guess is that sometimes when a developer tries a validator (usually the W3C validator) then...

From time to time I try other validators. Like the W3C.

When I look at their results I realise that helping developers doesn't seem to the core aim. They are instead taking a pedantic view of adherence to standards without thought of whether each standard makes sense or is useful to each developer.

Long story short, the information typically isn't useful enough to the developer.

As you say the developer (my view) needs to figure this out for himself. Developers who don't do this are, in some sense, incomplete.

I suspect that a lot of people are only aware of the less functional validators. One thing is for sure, I've seen them, and only them, integrated into a lot of software packages (presumably by people who haven't themselves learned to get it right). This wide, and misleading, visibility has, I'm sure, convinced many competent, thinking developers to give up on validation. (To the detriment of the web.)

[Albert Wiersch]

When I look at their results I realise that helping developers doesn't seem to the core aim. They are instead taking a pedantic view of adherence to standards without thought of whether each standard makes sense or is useful to each developer.

Exactly! I suspect most developers are interested in "real-world" or "practical" results. I know there are some who just have to stick with the standards (many in academia?), but that has its own drawbacks as well, probably more so than taking a "real-world" approach (as long as one is developing for the "real-world"). Don't get me wrong, standards definitely have value, but so does being able to cope & deal with real-world issues that might involve a little bending or contorting of the standards.

[MikeGale]

I'm surprised when I hear talk of standards being just about the only thing. Browsers to some degree do their own thing, within limits this is good. Great design comes from coherent thinking, trying things out etc. Committees and approval organisations stifle and actively prevent doing the right thing, all too often. So I believe that limited exploration is a good thing.

At root the standards may not reflect what happens in reality and are insufficient to be a single indication of what to do. That means developers need to do some thinking, which is great. It also means that any devloper who thinks that standards are just about the only thing, doesn't understand as much as he could, or maybe should.

On the browser as spyware issue, I saw a survey result. About two thirds of users had a high dislike. I imagine most haven't taken a look, for themselves. If they did I wonder what would change.

[Lou]

I agree standard are great, but why should I give away - block - any segment of my potential audience because some popular browser doesn't implement the standards in a way similar to others?

The objective is effective communication. To that end it makes no differences whether I don't follow the standards or the browser implementer didn't. If my message is not conveyed, I loose not the arbiters of standards.