Stupid obituaries, not again


Updated: March 2, 2012

Why is that people so much revel in death? Every two months, something else happens to be dying. First, it's Firefox, then it's Firefox again, then it's Linux, then it's the computer, and Firefox once more. Kind of gets boring after the seventh iteration, don't you think?

But no, apparently not. The media is bleeding for this kind of drama. For some reason, whether it's fear, boredom or clickbaiting, people keep posting about things dying, aiming for that trollful angle that will anger their readers and make them respond, to point out the statistical fallacies and vent out their fanboyness. Now, this here article is mostly about Mozilla's browser, but we will be touching other aspects of the modern computing, too. Follow me.

Teaser

Will Firefox die?

Depends on who you ask. Some people will tell you that it's already died, only it does not know that yet. Well, for that matter, so are you. You are dead, but it may take another fifty or sixty years for your body to register and realize that. Now, you will certainly perish, but the browser may yet live.

Anyhow, what prompts bloggers to speak about evil things to come is the variation in browser statistics collected by different sites, as an indication of growth or lack thereof, of certain browsers. For example w3schools.com has its own metrics and shares them with the wider audience, who is free to interpret the percentages any which way they want.

To get a moderately accurate estimate of what's really going on, you need several sources, first and foremost. But that's not all. Here's the one big problem that everyone ignores. People ignore the variations between different sources for a particular point in time.

Here's a warmup example, with completely fictional numbers, so you can fully understand what I'm trying to explain here:

Say you have three sources, A, B and C. All three sources show a slow, steady, linear decline from January to June, which averages some 1.5% loss. Then, you look at June figures, separately, which report that Firefox share was 19, 22 and 26 percent. All right, interesting. Now, you look at May figures: 20, 23, 29 percent. You go back, month by month. You create some nice linear trendlines. The average loss turns out to be 1.5%.

Now, you may assume that this means Firefox has lost 1.5% share to other browsers in the said period. But no. What everyone forgets, since they are too busy quoting and reposting the drama, is that the standard deviation for June alone was 3.5%. Even more than that in May. Looking back at the rest of data points, you get an average variation of 2.8% for your different sources. And since the standard deviation happens to be larger than the actual change, the market share drop can be considered a statistical noise. It might be accurate, but from the purely mathematical perspective, it is not.

Noise

If you look at what really happens in the real world, you will find out that numbers are a tricky thing. For instance, based on different sources Firefox had 24.8, 24.5, 25.5, 21.7, 19.7, 23.5 percent share in December 2011, with 2.2% deviation. The month before that, the numbers show a 2.1% deviation. For Chrome, the variation is 3.9%. In that regard, anything Chrome gained since August 2011 falls within the statistical noise.

So what do we learn here? Is there a trend? Yes, it seems like there's one. Chrome is gaining, Internet Explorer is losing, Firefox is maybe losing some users. Other than that, there's little else you can claim with certainty, unless reputation is secondary to glory.

Now comes the big question ...

Why shouldn't Firefox be losing some market share? When you only had two big browsers competing, the cake had but two pieces. Now there's another player. So it's only logical that an additional need, which neither Internet Explorer nor Firefox could deliver is being offered and satisfied by Chrome. Stands to logic. It's the natural order of things.

Most of Chrome gains comes from Internet Explorer. In this particular case, the change is obvious. Tens of percents of market share gained for an equal loss by the rival, exceeding the statistical variation by a factor of five or six, even hardcore skeptics will have to accept that for a fact. However, with Firefox, the rivalry is not immediately apparent.

Browser rivalry

Thus, premature obituaries are as useful as their content. My expectation is that the market will balance itself at a sweet one third share between the giants, more or less, and some extras for the rest. The situation where a single browser dominates the market is unhealthy, and look what it did to Microsoft. Google, if they are smart, will never want to get into a position where Chrome holds more than half the market, because it will become the victim of its own growth. Sustaining a competitive product is easier when you have ten million users rather than a billion. The speed of change is proportional to the size of it.

Furthermore, we learn that Firefox is losing market share DESPITE all of the new stuff they've done in the browser world, which includes simplifying, dumbefying and Chromefying their interface to the point of hangover nausea. What few people seem to realize is that market needs are far more subtle than just raw speed or the positioning of tabs.

What Firefox people did was take Chrome features and apply them to their model, but apparently, none of these features are considered important. On the other hand, Google people ported the one truly important thing from Firefox: the extensions.

There's going to be a balancing point. It has to be. If, for all practical purposes, Firefox and Chrome are identical, then for the average population, the choice is pretty much random. The competition leads toward standardization, which is a great thing.

Other things that will die

The drama does not end with the cute little fox. Linux is dying too. Again, we have numbers being molested. One of the most often quoted sources is Google Trends, which shows you the normalized volume of all Google queries pertaining to a specific search string. So if we look at Linux:

Trends

You would assume that Linux is less popular. However, this is incorrect. What people forget is that this claim would be true if the size of the market did not change. But it did.

Back in 2004, Linux search was approx. three times more than what it is in 2012. All right, but back in 2004, Ubuntu, Linux Mint, Android, and many other flavors of Linux did not exist at all. The SIZE of the cake is bigger than it was. Again, let's use numbers.

In 2004, the Internet had 500 million users. Today, it has three billion. The market has grown by a factor of six. Now, everyone seems to write how Linux hovers around the golden 1% market share. But this means that there are six times more Linux users than there were back in 2004! The cake has grown bigger.

Proportionally, Linux as a bare search word is less popular in 2012 than it was in 2004, but this tell us nothing about the actual size of the user base, especially since Linux used to be synonymous with anything Linux-like, but not so much anymore. Many new users searching for Ubuntu and Android do not really care that the underlying kernel is Linux. For that matter, people using Mac OSX do not know they are running UNIX. If you Google UNIX, you will see a super-steady decline, but in that period, Mac has gained sizable chunks of the desktop market.

It is important to look at the weights before making any decisions. And then, reference other sources to get an even more accurate impression. Linux weight in the last eight-year period is 1.00, Ubuntu is 0.32, and Mint virtually zero. If we normalize by Linux Mint, the overall volume of search for Linux itself is about 200 times less than what people search for Linux alone in this time frame. But then, Linux Mint did not exist so far back then.

Trends, normalized to Mint

Moreover, DistroWatch reports Linux Mint to be its top daily hitter for a while, with almost twice the popularity compared to Ubuntu. So things are definitely changing and they might not be reflected in averages. Or the other way around. And let us never forget that for each data point shown above, the market size was different.

Indeed, the market has many more search strings today, including many variations and mutations. For instance, if you search for Linux Mint, you also cover Linux and Ubuntu categories, without explicitly naming them, so to speak. Is this significant, yes, no? Maybe. The big question is, what if you take the top ten Linux-related searches and then sum them all together. What do you get then? What if you dump Android there, too?

Computer is dying too, not

Yes, yes, we had this one before. Everyone has a proverbial hardon for the chimp emulation computing model, also known as craplets and moronphones. I can appreciate the enthusiasm for new market segments. But everyone is making the same mistake by assuming the market remains fixed and that new technologies come at the expanse of old ones.

No. The market is scaling out horizontally and vertically. Having new gadgets opens up new possibilities, it does not replace old ones. There's no conservation of CPU cycles, you mathless dolts.

There's massive hysteria around this new stuff to the point of dumping all of the collective user experience into a single basket, then setting that basket on fire with napalm. Touch screens truly need their unique big shiny buttons and whatnot, I agree. Consumerism, live streaming, all that nonsense. But the new possibilities are a logical continuation of the existing, not a replacement. To help you cope, here's the analogy.

TV channels, cables. When you have only one channel, you must watch it. If someone offers you another channel, you may spread your time equally between the two, or not. With a whole range of options, your daily repertoire will eventually reflect your time, needs and taste. So far so good. That's your conventional, traditional entertainment model. Now, someone introduces live streaming. Someone else creates the radio. You get 3D cinema, you get virtual reality. All of these offer a similar experience.

Now, if you were to follow the smartphone nonsense, we should abolish old technologies or make them identical to the new ones, when in fact they are complementary in nature if not in the functionality. However, the simple reality is that TV lives happily alongside all other entertainment options. New needs are created and met.

The same with computers. Fifty years from now, it is quite possible that everyone will be using ultrabooks, superbooks and smartphones. But people will still continue using their traditional tools. In the relative sense, the PC dominance may drop from 80% to just 15%. But in absolute numbers, the absolute share will still be bigger than it was before. Because, 1% out of a million is more than 10% out of a thousand. Get it?

Cake

PC, laptop, netbook, and other gadgets were mostly created to fulfill real needs. If not, then they might truly disappear. Indeed, some of the gadgets today have been created as fashion commodities, which is why they are being hyped so much. When there's no real need, you must fight hard to survive. That's why your brains are being pumped with crap about Web 3.0, instant access and all that bullshit.

As long as the needs exist, the tools will exist. Once people no longer require brute computing power, hardcore gaming, typing, or large screens, the classic computer might disappear. Will this ever happen? Well, we've been typing for the past 400 years, so I'm not sure. Life dictates size. And thumbs don't get any smaller. Nor can our eyes see details just as well on a 4" screen as they can on a 22" screen.

Can you comfortably work for 9 hours with only one hand, while the other is holding a tablet? Can you comfortably run CAD tools on a smartphone? Would you write a book using one of these new-age gadgets?

Hey, if we're into making things cool and ignoring reality, why not replace the piano with a smartphone? I mean you can have the keys emulated and all that. So why not? One size fits all is like digital communism. Not gonna work.

So is anything dying yet?

The answer is simple. Which of the would-be dying things are fashion and which ones answer real human needs? That's the only question you need to ask. If they are used to better humanity in some way, they will stay. If not, they won't. That's why MySpace died. And this is why Failbook will most likely die. Because it does not have a real purpose in human life, in the long run. Look around you. What do you really need? Is there a technology that answers that need? Yes, good, life. No, bad, death.

Death

Note: Image taken from Wikimedia, licensed under CC BY-SA 3.0.

Conclusion

In the digital age, browsers have become an important element of our life, since we use them to obtain and share knowledge and information. Since this is inherently a human survival trait, we will need browsers or their equivalents for decades to come. There must be market fluctuations. The laws of natural selection apply here. The survival of the fittest. The good thing is, unlike nature, which has a finite pool of resources, the Web keeps growing. So the total extinction of your rival is not necessary.

Firefox's share might be dropping, but the absolute number of users is quite likely steady or even increasing slightly. This means that Firefox usage model has reached the top of its capacity and serves people whom this model meets their needs. Chrome is gaining, because it is answering a new need. Firefox took about seven years to peak. Let's see when Chrome will exhaust its pool. There must be some cross-browser migration, but things will balance out. Nothing will die. Neither Linux, nor the computer. So cut the crap already. And learn how to do proper math.

The teaser images here and on the index page and the TV image are in public domain.

Cheers.

RSS Feed icon

del.icio.us del.icio.us stumbleupon stumble digg digg reddit reddit slashdot slashdot



Advertise!

Would you like to advertise your product/site on Dedoimedo?

Read more

Donate to Dedoimedo!

Do you want to
help me take early retirement? How about donating
some dinero to
Dedoimedo?

Read more

Donate