Stability comes in many forms, not simply up time.

For example, Linux has a long, long history of badly managed architectural transitions:
a.out to ELF
libc to glibc
virtual memory manager musical chairs
filesystem flavor of the month
32bit to 64bit
package manager du jour
sound
MAKEDEV/devfs/udev.

Stack on top of that the variety of distributions, with their own often wildly different ideas about where things should live and how they should be managed, frequently causes stability issues by introducing human error points. Many of those ideas are also inherently bad and affect stability, such as RedHat and friends throwing everything and the kitchen sink into /usr. -Yes, some packages can be retargeted...but not many, and doing so breaks convention (albeit a bad one) causing the same sort of management stability issues that multiple distros cause just on the local level.

All of that ends up being a make-work program for Linux System Administrators...honestly at leat 50% of your daily job only exists because of the instability of the Linux ecosystem.

Linux (all distros, all of it) is a Configuration Manager's worst nightmare.

While I feel for your plight, the idea of running without hardware acceleration in 2013 is pretty uncommon. It's certainly nothing any sane person would bother optimizing for. We're optimistically talking about 0.1% of users. You'd effectively be optimizing for actually broken systems. That's like optimizing a car's high speed handling for cases where one tire is flat; The answer isn't to optimize for conditions of only 3 healthy tires...the answer is to change out the flat tire.

Any chance you can offer a link to any of these particularly intensive PDFs?

I'm not discounting your experience, I just have yet to encounter anything like you're describing.

That's completely opposite of my experience.

On my not-so-hot computer I regularly open very complex, 400+ page PDFs (music scores mostly). We're talking 30MB w/o any imbedded images, just pure intensive processing instructions.

Chrome, from a total standstill (the process not even running yet), takes just slightly longer then it takes me to blink to start, load the PDF, and render. It's an order of magnitude faster in every way then every other PDF viewer I've tried, and I've tried quite a few.

It lacks features (PDF bookmarks, etc), but render speed is fantastic.

Programs are thought.

Wait, I thought they were math?

Math is not abstract. Math is not thought.

Math is a symbolic representation of the computational laws of the universe. When we eventually find intelligent life on another planet, they too will understand that 1 + 1 = 2. Guaranteed. The written symbols they use will differ, but the math won't. Not the teeniest, tiniest bit.

Programs are creative thought. There isn't anything even slightly creative about math any more then there's anything creative about discovering the Higgs boson.

Just because you can think about math in no way suggests that math is thought. I can think about cute puppies too, but that doesn't mean my dog is just an abstract thought.

By that logic programs are also all particles, electrons to be precise.

You can't say something isn't patentable simply because it can be described by math...because absolutely everything can be described with math, very much including the thought in your head at this very moment.

You can't use math as a distinction because math makes no distinctions.

I'll forgive Firefox when it can

A) Pull its own weight without lagging on common sites despite being run on incredibly powerful systems.

B) Stops forcing incredibly stupid "features" and UI design failures.

----

With our test labs and such we're running every browser across an array of configurations and performance levels. On Javascript intensive sites Firefox simply lags far behind everything else. That didn't matter so much pre "Web 2.0", but today when JQuery is considered lightweight compared to many common frameworks in use, it matters a whole hell of a lot.

Of course, hardware is all physics, and physics is all math, therefore hardware is all math. The universe is nothing but math.

If everything is math (and reductively, it is), then "math" has no useful meaning. It may be good for boosting the egos of mathematicians living in their own little bubbles, but as a practical matter it has pissall to do with anything back in the real world.

I won't argue for software patents (I think they're asinine), but trying to make that point using the "software is math" argument just isn't honest or helpful to the cause.

Because everyone and their mom must be running Firefox either without Javascript or with no more than a single tab open. And heaven help you if you open any dev tools with Firefox...yee gads!

Although I do find it amusing that if you trust Firebug's load time metrics it's just as fast as everyone else....never mind it hasn't actually rendered everything it says is "completed" yet, or gotten half way through the billion lines of javascript that most serious sites now run.

I run Firefox constantly, albeit grudgingly, because it's sadly a browser we have to support. Across dozens of systems of a variety of hardware configurations and platforms. This isn't a case of "just my PC is funky". If it ever actually improves I'll be one of the first to notice.

I haven't seen anything but significant degradation from Firefox for years. Every update is a downgrade. In performance, in functionality, in usability. You are right that Firefox it isn't at its low point today...only because it keeps digging its grave a bit deeper with each passing update.

I harped on IE just as bad when they were the crappiest browser on Earth, but that hasn't been the case for years. I still hate the UI, but at least it's fast. More over Firefox's wounds are all self-inflicted.

The only reason I pay Firefox any attention whatsoever is that it still has significant market share and thus, sadly must continue to be supported. I long for the future where it's as relevant as Opera and I can just not even bother having the turd installed at all.

The fact is they have to focus on performance, especially their Javascript engine...precisely because their performance sucks big harry balls, most especially their Javascript engine!

Call me back when/if they ever actually make progress. All I've seen as a user is a browser that has gotten nothing but slower and slower with each release, a sharp contrast to Chrome and IE.

It was very close second to chrome in Benchmarks.

Which seems more evidence of how badly Tom's Hardware botched the testing than anything else.

In real use Chrome is so wildly faster than Firefox it's incredible, thanks largely to its Javascript engine and the fact most modern web sites are incredibly JS heavy. I'm not even talking 20% faster, but closer to 200%, literally running circles around Firefox.

It's to the point that just taking raw performance into consideration, I can't stand using Firefox lately even on very powerful hardware. I frequently have to fire it up for work, but man...what a drag.

So you'll forgive me if I can't take Tom's "benchmarking" seriously.

A LOT of software has little to do with math.

Sorry, but ALL software is an expression of math..

It's much more correct to say software is implemented ultimately with math, but it's rarely ever an expression of math.

The vast majority of software is little more than an automated (via math...) flow chart... Flow charts that express business processes or artistic ideas. There's a reason why trained mathematicians on the whole make for some of the worst software developers you'll ever find, and that reason is because aside from a few niche domains the practice of programming has fuck all to do with math.

It really depends on what the patents are all about.

Chances are near certain that they are all a bunch of the typical BS software patents we see all the time and never should have been granted. It's just cheaper to pay the troll than to fight him.

You're spending how much video hardware and you're still running air cooling?

Put in a good water loop already, sheesh... :-/