Goodbye Firefox

[Tannin]

The purpose of the web, Buck, is not just to communicate, it is to communicate in an open and accessible way. This is why we have strong (and widely supported) rules requiring that content (HTML) and presentation (CSS) are kept conceptually separate and (in common practice) kept apart in different source files. This is why any properly written web document can be easily understood by any person with even very basic training. The web shouldn't just be open, it should be seen to be open, and readily verifiable as such.

Present an untrained person of moderate intelligence with a properly written HTML document, even a complex one, and he can see for himself exactly what it contains, how it is structured, and what it does. He is in a position to decide what to do with that document, decide if it is safe, decide if it presents a threat to his privacy. Present that same person with one of the sadly commonplace convoluted Javascript chunderscapes one sees everywhere these days and he is powerless. Even a skilled computer expert, unless he happens to have specialised skills in that particular language (and few do), cannot decode that document or see what it does without a great deal of effort.

The only, repeat only valid excuse to use Javascript on a web page is when there is no sensible alternative method.

 

[Tea]

Javascript is pox. It will rot your teeth, ruin your marriage, corrupt your kids, crash your browser, hasten the apocalypse, make your babies cry, spoil milk, create large dead yellow spots on your lawn, slow your computer to a crawl, wreck your credit rating, ruin the beauty of your website, harden your arteries, and upset Tannin.

On balance, and taking into consideration that last item, it overall worth the risk.

 

[Handruin]

The only, repeat only valid excuse to use Javascript on a web page is when there is no sensible alternative method.

Which there isn't an alternative when it comes to client-side (or even server-side ajax) validation and improved UI interaction with the consumer of a website. You can't do this kind of work with HTML5 of CSS3.

 

[Mercutio]

Which there isn't an alternative when it comes to client-side (or even server-side ajax) validation and improved UI interaction with the consumer of a website. You can't do this kind of work with HTML5 of CSS3.

Sure there is. You could build your interactive application with Java, Flash or Silverlight. Or you could just stop using the HTTP as a replacement for the top three layers of the OSI model and deliver a proper binary for whatever it is you need to do.

 

[Handruin]

Sure there is. You could build your interactive application with Java, Flash or Silverlight. Or you could just stop using the HTTP as a replacement for the top three layers of the OSI model and deliver a proper binary for whatever it is you need to do.

Seems like a heavy-hitting and less usable way to do basic programatic work on a web page. I really don't believe in plugins like the ones you've suggested.

 

[Mercutio]

Seems like a heavy-hitting and less usable way to do basic programatic work on a web page. I really don't believe in plugins like the ones you've suggested.

I'm not arguing that, just that there are options available. Options that for the most part make javascript seem like the least of all evils, even.

 

[Tannin]

The only, repeat only valid excuse to use Javascript on a web page is when there is no sensible alternative method.

Which there isn't an alternative when it comes to client-side (or even server-side ajax) validation and improved UI interaction with the consumer of a website. You can't do this kind of work with HTML5 of CSS3.

That's my point. I might have added that, in 90% of cases, this "improved" UI interaction used as a bad excuse to pox up a web site in fact makes it worse, not better. Mostly, Javascript is used to degrade up the load times, compatibility, and responsiveness of the site, as well as to defeat the purpose of the web by obfuscating the code. Whenever I see Javascript navigation menus, for example, I barf. There is no excuse for Javascript navigation.

(This is not to exclude the possibility of using Javascript for minor tweaks to an otherwise pure CSS menu, of course. But it is a last resort for use only when none of the good methods work.)

 

[timwhit]

Trying to do what is possible with Angular.js, Backbone.js, or React with straight html is just impossible. Have you read about or done a proof of concept project with any of these modern web frameworks?

 

[Tannin]

Most web frameworks are bloated, ugly crap, Tim. I am yet to find even one framework that I would want to use. Well, not unless you count the extensive library of fast, concise, efficient hand-written PHP and CSS code that I have written myself for various projects over the years and can draw on whenever I need to do something new. (And yes, there is Javascript in there for those few - very few - things that simply can't be done in CSS. Out of many tens of thousands of lines, there might be 50 or 100-odd lines of Javascript. When I get time one day, I'll revisit them and, with any luck, reduce them a little further.) The thing is, when I write a page I use the code that page requires and only that code. So it's fast and generally bug-free. And it works with any browser. People regularly comment on how fast and nice to use my sites are. (They also comment, from time to time, that Site X hasn't been updated for 100 years and looks like it. This has nothing to do with my coding practices, however, and everything to do with having one life and enough things I want to do at any given moment to soak up every spare minute out of seven lifetimes.)

Learning a new framework is a non-trivial task, and the time I'd devote to it is time wasted that I could devote to actually doing something useful with the proper tools. (Never mind the fact that the result is always slow and ugly when compared to doing things the proper way in the first place.) If people who call themselves web designers spent half the time they waste mucking about with bloated frameworks actually learning how to code properly in the first place, the world would be a better place.

Doubtless those ones you mention have a function somewhere. If I ever want to do something I can't do with the first-class tools (HTML, CSS, and whichever is the most convenient server-side scripting language at the time), I'll look at second-class tools like Javascript libraries. As a matter of detail, a year or so ago I needed a way to scroll some navigation horizontally which simply wasn't possible with CSS. This is unusual - you can do almost anything with modern CSS and HTML5, including 95% of what you see bloody Javascript used for. It turned out that the only documented way to do it was via Javascript and the huge, clumsy Jquery library. Thousands of lines of bloat! Yuk! So I figured out how to write a raw Javascript function to do exactly what I wanted, no more, no less. Well, I tried to but it was too hard, and buggerising around with Javascript is like swimming naked in a sewer full of industrial effluent. I hate it! In the end I asked Tea to write it for me, which she did 'cause she is such a helpful little ape whenever she wants to be. (This is usually between 11:05 and 11:10 on 5th or 6th of most Octobers.) Tea's version is about five lines of code. Works perfectly. Site still works perfectly (bar this minor enhancement not being there) with Javascript disabled. This is how it is supposed to be done.

Now, if you are writing programs rather than pages, and you happen to want that program to run synchronously and use client side processing, and you want to use HTML as a convienient front end for your software, that's a different matter. Javascript is apparently the language de jour for that. I regard that as a sad thing, (a) because Javascript is such a horrible, horrible language - I'd prefer writing stuff in bloody Fortran or even APL, never mind something clear, elegant and structured like Pascal or Modula-2 - and (b) because the proliferation of active programs disguised as passive web pages and accessed by the unsuspecting user through a web browser is downright frightening. This is where nearly all of modern security breeches (viruses, spyware, and so on) come from, and it was a shocking bad idea in the first place.

Excuse me, I feel the need to rush out and find a school or place of worship now. Watch for me on the nightly news.

Sorry. Cancel that^. I forgot that you are supposed to go armed with semi-automatic deadly weapons. I don't have any. Not unless you count a text editor and a Javascript library.

 

[Chewy509]

I personally hate the misuse of javascript for sites... (blog sites that fetch content via AJAX, and the page itself is nothing more than a JS blob and that blob is several MB).

But I think both sides are talking about two slightly different things, one about content (web-site with mostly static content), and the other about redefining the web browser as an application hosting service (web-application with very dynamic content).

There does appear to be a lot of developers who seem to mistake one for the other, but there are a few fundamental flaws with web-applications, namely, you are trusting your user to execute a lot of code that talks to sensitive back-end systems, especially where that code is available to the user in plain sight...

Now image handing over all your source code that includes references to data structures (JSON/XML), references to input validation, etc... and the user then carefully looking for holes in the communications/message protocol you are using, and exploiting that! And you're also really on the users environment to do things right as well... (There are lots of bugs in JS engines that require work-arounds).

But do we ditch the whole web-application paradigm and move back to heavy applications (be they .NET, Java or native) and deal with those problems? I don't know...

For me, I prefer full applications that are native, vs web-applications, that may be just my bias, as I remember back when HTML 2.0 was the latest and greatest.

 

[Chewy509]

@Tannin, well said.

 

[time]

I thought that this was just another Tannin rant (nothing wrong with that), but that last post gets a 9 out of 10 from me for real world insight. Chewy's post was, as usual, boringly accurate and relevant.

That question about native apps vs web apps has been bothering me of late. In the mobile space, native apps have clearly won. Is Google's vision of Chrome a complete dead end?

 

[Mercutio]

I suspect the relative ease of blocking ads in browsers vs. being able to make mobile apps devote a third of your screen real estate to them will settle that. I can see app-ification of formerly web based content as a possible evolution in response to ad blocking there as well.

 

[mubs]

The local Amazon copycat (founded by two ex-Amz employees) bought out a fashion dress portal and later made it pure mobile app only. Now they are pontificating making their entire website app-only. The IT crowd (tech-savvy, with high purchasing power) is up in arms because they make a lot of purchases at work on their work PCs. Of course Luddites like me will never do banking or shopping on a mobile. I can envision a future where the internet is app-ized and people like me will go back to brick-and-mortar stores and paper news, assuming, of course, that they exist. I believe that in my lifetime they will, and the new generation (like my daughter) doesn't care enough about privacy and security anyway.

 

[Buck]

Most web frameworks are bloated, ugly crap, Tim. I am yet to find even one framework that I would want to use. Well, not unless you count the extensive library of fast, concise, efficient hand-written PHP and CSS code that I have written myself for various projects over the years and can draw on whenever I need to do something new. (And yes, there is Javascript in there for those few - very few - things that simply can't be done in CSS. Out of many tens of thousands of lines, there might be 50 or 100-odd lines of Javascript. When I get time one day, I'll revisit them and, with any luck, reduce them a little further.) The thing is, when I write a page I use the code that page requires and only that code. So it's fast and generally bug-free. And it works with any browser. People regularly comment on how fast and nice to use my sites are. (They also comment, from time to time, that Site X hasn't been updated for 100 years and looks like it. This has nothing to do with my coding practices, however, and everything to do with having one life and enough things I want to do at any given moment to soak up every spare minute out of seven lifetimes.)

Learning a new framework is a non-trivial task, and the time I'd devote to it is time wasted that I could devote to actually doing something useful with the proper tools. (Never mind the fact that the result is always slow and ugly when compared to doing things the proper way in the first place.) If people who call themselves web designers spent half the time they waste mucking about with bloated frameworks actually learning how to code properly in the first place, the world would be a better place.

Doubtless those ones you mention have a function somewhere. If I ever want to do something I can't do with the first-class tools (HTML, CSS, and whichever is the most convenient server-side scripting language at the time), I'll look at second-class tools like Javascript libraries. As a matter of detail, a year or so ago I needed a way to scroll some navigation horizontally which simply wasn't possible with CSS. This is unusual - you can do almost anything with modern CSS and HTML5, including 95% of what you see bloody Javascript used for. It turned out that the only documented way to do it was via Javascript and the huge, clumsy Jquery library. Thousands of lines of bloat! Yuk! So I figured out how to write a raw Javascript function to do exactly what I wanted, no more, no less. Well, I tried to but it was too hard, and buggerising around with Javascript is like swimming naked in a sewer full of industrial effluent. I hate it! In the end I asked Tea to write it for me, which she did 'cause she is such a helpful little ape whenever she wants to be. (This is usually between 11:05 and 11:10 on 5th or 6th of most Octobers.) Tea's version is about five lines of code. Works perfectly. Site still works perfectly (bar this minor enhancement not being there) with Javascript disabled. This is how it is supposed to be done.

Now, if you are writing programs rather than pages, and you happen to want that program to run synchronously and use client side processing, and you want to use HTML as a convienient front end for your software, that's a different matter. Javascript is apparently the language de jour for that. I regard that as a sad thing, (a) because Javascript is such a horrible, horrible language - I'd prefer writing stuff in bloody Fortran or even APL, never mind something clear, elegant and structured like Pascal or Modula-2 - and (b) because the proliferation of active programs disguised as passive web pages and accessed by the unsuspecting user through a web browser is downright frightening. This is where nearly all of modern security breeches (viruses, spyware, and so on) come from, and it was a shocking bad idea in the first place.

Excuse me, I feel the need to rush out and find a school or place of worship now. Watch for me on the nightly news.

Sorry. Cancel that^. I forgot that you are supposed to go armed with semi-automatic deadly weapons. I don't have any. Not unless you count a text editor and a Javascript library.

Tony, you should learn to write in a more concise manner, and with less drama, like Tea.

 

[Chewy509]

Chewy's post was, as usual, boringly accurate and relevant.

So unsure how to take that?

 

[Tea]

Tony, you should learn to write in a more concise manner, and with less drama, like Tea.

Buck, you should learn to read faster, like me.

 

[Handruin]

So unsure how to take that?

Take it as a compliment. You provide detailed and well thought-out explanations.

 

[Mercutio]

Palemoon no longer uses Gecko as its underlying rendering engine, but rather a fork called Goanna. Absolutely the only change I've noticed because of this is that Youtube now gives me a message that I need to upgrade to Firefox or Chrome when I browse there, and defaults to a much less functional view than what it had before this weekend. Interestingly, Youtube worked FINE all weekend, even after I upgraded to the newest release of Palemoon, which suggests that this is web code that it added today, about five days after the updated version of Palemoon was released.

 

[timwhit]

Can you get an extension that sets your User-agent to something Youtube recognizes?

 

[Mercutio]

I'm sure I can, but it's a case that I'd rather be seen in metrics as using a current, modern browser so that it can be counted and considered for support.

 

[sedrosken]

Is there much that Pale Moon does that classic theme restorer on pure Firefox can't? Right now I have it set up to look pretty much like Pale Moon, without the big toolbar buttons, mine are all tiny courtesy of the MicroFox theme.

 

[Mercutio]

Is there much that Pale Moon does that classic theme restorer on pure Firefox can't? Right now I have it set up to look pretty much like Pale Moon, without the big toolbar buttons, mine are all tiny courtesy of the MicroFox theme.

Palemoon isn't constantly packaging new third party software.
Palemoon has no plans to discontinue fully functional NSAPI plugin support.
Palemoon isn't fucking around with a touch friendly default theme that absolutely no one asked for.
Palemoon doesn't change user defaults when its sponsorship changes.
Palemoon doesn't look for excuses to find new places to put its own internal advertising.

At this point, I don't trust the stewardship of the Mozilla foundation to do anything correctly. Its current CEO has a marketing background and the result of that leadership has been a product that's sinking into the toilet.

... but it's still better than Chrome.

 

[sedrosken]

Noted. Thanks for clearing that up, in all seriousness.

Honestly I'm wondering why they don't have an official *buntu build? Everything I'm seeing is telling me to either use the official Linux tarball or use a PPA that hasn't even had the index updated for Wily.

EDIT: Well, this is somewhat surprising. What do you think the odds are of using this working out?

EDIT 2: The installer, while rather unorthodox and somewhat advanced for those new to Linux, certainly works and is easy enough to use. I just have to remember to use that instead of apt-get when I need to update it. I still say they need to make an official deb package. Pale Moon took my old Firefox settings just fine after I removed AdBlock Plus (to replace with AdBlock Latitude later on) and Classic Theme Restorer. No side effects that I can tell. YouTube works fine, even the bloated HTML5 player. No sites I visit try to get me to 'upgrade'. I'm running 26.0.2, is that the latest version? Firefox's addon plaza or whatever says it's equivalent to Firefox 24.9.

EDIT 3: YouTube doesn't like going any lower or higher than 360p in HTML5 for Pale Moon. I think that's what Merc was referring to? No matter, forcing it to use the flash player instead eliminates this annoyance. It's lighter on my resources that way too.

 

[Mercutio]

For about two days, I was getting a "Upgrade to Chrome!" message when I hit Youtube on Palemoon, the same message that IE6 or weirdo third party browsers (e.g. Dolphin) users get. And then I stopped getting the message. I don't know if that's something Palemoon did or something Youtube did. Either way, it's fine now.

 

[mubs]

NoScript and other popular Firefox add-ons open millions to new attack

 

[Mercutio]

Firefox runs in a monolithic process that is not sandboxed (or running in a least privileged session, which is something IE does) from anything else. That's not news. That's its design. It's a trade-off for memory efficiency, or at least it was.

 

[CougTek]

I use Adblock Latitude with Pale Moon 64-bit. It isn't listed on the table of vulnerabilities, but it certainly doesn't mean it isn't vulnerable. This isn't comforting.

 

[sedrosken]

Since Vivaldi hit its third beta I've been using it for a daily driver. The other day it got it's first official (1.0) release.

Above the hood, It's basically a flat version of Opera. Stacking tabs, panels, etc etc.

It's basically Chrome under the hood, which is good news for speed and somewhat bad news for RAM usage. Hence why I installed "The Great Suspender." It's an extension that suspends inactive tabs after a user-defined period of time or can do it manually as well. Tabs that are suspended take up far less RAM and Chrome actually lets go of RAM that it isn't using (unlike Firefox and co.).

 

[Handruin]

Here is some news on some changes coming to Firefox related to performance improvements over the next several weeks. If they can pull this off it might really help. I can only see this useful once they allow this to work with add-ons.

• The company is reporting a 400 percent improvement in responsiveness and a 700 percent improvement in responsiveness for loading large web pages.
• Mozilla has completed its initial tests on 1 percent of its user population and the initial numbers are good, according to Asa Dotzler, director of Firefox at Mozilla.
• Over the next week, multi-process will be coming to 10 percent of total Firefox users.
• For now, users with add-ons will not be getting the new architecture.
• In the coming weeks, Mozilla will push multi-processing to 100 percent of their initial cohort of users.
• For now, multi-process is limited to a single content process and a single browser process. Later versions will include multiple content processes and sandboxing.

 

[ddrueding]

Interesting. My patience expired yesterday; new OS install came with a switch to Chrome. We'll see if they can get me back.

 

[Mercutio]

Palemoon is a thing that exists.