Apple WWDC14 Streaming Live

Understanding the fragmentation of different platforms is useful information to a developer.

And that's why I bought it up because percentages of things of so dissimilar size doesn't have much practical value. Yes, a Windows Store app can at best right now only reach 14% of Windows users. When the number is used to talk about one platform this percentage has meaning. It would also have meaning when comparing different platforms of similar size. It's meaning fades when speaking of percentages on platforms of vastly different sizes.

Right now, virtually 100% of all Windows 8" tablets run 8.1. But what meaning would that have for developer if I then compared that to say 10% of Android 8" tablets running Kit Kat? It doesn't really.
 
And that's why I bought it up because percentages of things of so dissimilar size doesn't have much practical value.
Ummmm what? Of course market fragmentation data has practical value. All things being equal, most developers prefer non-fragmented markets.

Knowing that a particular platform has a lot or a little fragmentation informs all sorts of development and management decisions.
 
Are you seriously suggesting that adoption rate and fragmentation data is not useful to developers? More data is better.
 
Pretty clear now why Apple limited the live streaming to Apple only equipment because anyone else clearly recognize all the newly announced iOS 8 and OS X 10.10 features as borrowed from others.
 
Glancing over this, even I get the idea. OSX developers have to develop for a new OS much quicker than Windows developers do since so many adopt the new OSX. With Windows, most are still on XP or 7, so you can still develop for those for quite a while until you have to move on. This has nothing to do with market size, more to do with developers developing for a new OS.

So we're supposed to compare 50% of x to 14% of y and ignore what the values of x and y are? One thing about Windows software sales that's been noted over the years. The newer the device, the much more likely someone is going to buy software for that device. The number of just Windows 8 tablets sold is coming close the all Macs sold. And I bet the first thing that many Apple folks will say is that the iPad sells a lot more than Windows 8 tablets. Which for now is true.

You can't talk about 50% of x and 14% of y without the values of x and y coming into the discussion. That's just not how it works.
 
This argument doesn't matter. Company X will always present numbers in a way that sheds the best light on them and the worst light on competitors.
 
This argument doesn't matter. Company X will always present numbers in a way that sheds the best light on them and the worst light on competitors.

I completely agree. And I'm not criticizing Apple for doing it. Again, it's just not possible in the real world to compare 51% of x to 14% of y without the values of x and y coming into the discussion. Because otherwise the comparison mathematically has little meaning without values for x and y.
 
Because otherwise the comparison mathematically has little meaning without values for x and y.
Of course it has meaning - developers are often concerned with adoption rates and market fragmentation - these are very important pieces of information when making business decisions, planning projects, design decisions, etc.

If you're involved in cross-platform development work you ought to know this.
 
Of course it has meaning - developers are often concerned with adoption rates and market fragmentation - these are very important pieces of information when making business decisions, planning projects, design decisions, etc.

If you're involved in cross-platform development work you ought to know this.

So do you know of any project that has made any plans around the 51% adoption rate of Mavericks compared to the 14% adoption rate of Windows 8.x? How many applications for OS X require Mavericks? There are probably considerably more Windows applications that require 8.x as the 165k Windows Store apps require 8.x. If comparing adoption rates between completely different OSes with completely different market shares, why would anyone develop Windows Store apps which only run on at most 14% of Windows PCs? Why not take that effort and develop applications that only work on Mavericks with its 51% adoption rate?

Again, it's not the adoption rates that I'm talking about. It's about comparing two adoption rates where the market sizes are so dissimilar. When 51% of something is 1/4th of 14% of something else in absolute numerical terms, comparing the percentages alone isn't meaningful.
 
When 51% of something is 1/4th of 14% of something else in absolute numerical terms, comparing the percentages alone isn't meaningful.
You keep saying it isn't meaningful- you should think carefully about what it is for something to be "meaningful". This information is valuable to some people, even if you don't care about it.

I can list off reasons why someone might be interested in the comparative adoption rates if you want.
 
You keep saying it isn't meaningful- you should think carefully about what it is for something to be "meaningful". This information is valuable to some people, even if you don't care about it.

I can list off reasons why someone might be interested in the comparative adoption rates if you want.

Windows%208%20vs%20OS%20X%20Mavericks%20adoption.png


In the real world the first thing a person who understands anything about numbers or charts is going to point out "Where are the underlying numbers for these graphs? How would I know what the 51% compared to the 14% means otherwise?"
 
"Where are the underlying numbers for these graphs? How would I know what the 51% compared to the 14% means otherwise?"
I assume you know what a percentage is.

Sometimes people are interested in percentages; sometimes people are more interested in percentages than the absolute numbers behind them; sometimes percentages do a really good job of conveying some information.


Let's note that some things are meaningful in some contexts and not in others.

For example, to someone investigating traffic congestion, the comparative adoptions rates of operating systems is mostly meaningless; however to someone investigating methods of delivering software updates or how to price their software, it is meaningful.

I assume you have some context in your mind where maybe the comparison is not meaningful, but not everyone shares that.
 
1. Given (51/100)x and (14/100)y which one is larger?

Yes I understand percentages and basic algebra.

What you're trying to get at, I suppose, is that we can't know which is larger without knowing x or y. That's fine, except many people are concerned with more things than simply knowing which number is larger.

You'll be surprised to know that there is a whole world of numbers, mathematics, statistics, characterization techniques, etc., that provide really useful information and analysis for people and go well beyond simply just counting how much of something there is.
 
You'll be surprised to know that there is a whole world of numbers, mathematics, statistics, characterization techniques, etc., that provide really useful information and analysis for people and go well beyond simply just counting how much of something there is.

I understand that. But you couldn't calculate these percentages without making the counts in the first place. So simple counting in this instance is the whole basis of the percentages.
 
I understand that. But you couldn't calculate these percentages without making the counts in the first place. So simple counting in this instance is the whole basis of the percentages.
Numbers are too confusing. Pretty pie charts explain everything.
 
Given that it's a relative comparison, yes.

So if I only target for Mavericks then I get 51% of the OS X crowd which is 40 million potential customers for that version. If I only target for Windows 8.x for Windows folks I get only 14% of the Windows crowd but gee, that's 160 million potential customers. So hey, why would I target specially for Mavericks or Windows 8.x and lose 49% and 86% respectively of those customer bases?

If 53% vs 14% means something to developers then great. But 40 million vs 160 million probably means something developers as well.
 
So if I only target for Mavericks then I get 51% of the OS X crowd which is 40 million potential customers for that version. If I only target for Windows 8.x for Windows folks I get only 14% of the Windows crowd but gee, that's 160 million potential customers. So hey, why would I target specially for Mavericks or Windows 8.x and lose 49% and 86% respectively of those customer bases?

If 53% vs 14% means something to developers then great. But 40 million vs 160 million probably means something developers as well.

that's majestic and all, but the audience was composed of Apple developers.
 
that's majestic and all, but the audience was composed of Apple developers.

You do realize that this afternoon that you said this quoted here entirely and verbatim:

holy shit, dude, context really is key, here. ~70% of the developers in attendance were new to Apple development. that means there's some probability they've seen some Windows development. the point being made wasn't "hurr durr our numbers are bettar!" but rather "adoption rates and, consequently, development targets, move faster in Apple land."

http://hardforum.com/showpost.php?p=1040872215&postcount=34
 
So if I only target for Mavericks then I get 51% of the OS X crowd which is 40 million potential customers for that version. If I only target for Windows 8.x for Windows folks I get only 14% of the Windows crowd but gee, that's 160 million potential customers.
Clearly you're more interested in how many users are on Mavericks and on Windows 8; the charts are about adoption rates though.
 
"there's some probability." also, it's an easy comparison to make to make the simple point: we move fast.

but yet again you're intentionally being obtuse.

If someone is going to talk about 51% percent of some number vs. 14% of some other number the first thing I'd want to know is what those numbers are. And if ~70% of the people that were in attendance at that conference were new to Apple development like you said, those people would have naturally wanted to know the same thing.

Arguing common sense in a place like this I admit is just entertaining for me. Because so many people love to argue against it.
 
If someone is going to talk about 51% percent of some number vs. 14% of some other number the first thing I'd want to know is what those numbers are.

You're really going to make a blanket statement about the utility of percentages and how you would always ask for the raw data????

Person: "I got 87% on this test"
You: "Meaningless! How many questions were on it and how many did you get right?"

Percentages are often much better metrics than raw data.
 
You're really going to make a blanket statement about the utility of percentages and how you would always ask for the raw data????

Person: "I got 87% on this test"
You: "Meaningless! How many questions were on it and how many did you get right?"

Percentages are often much better metrics than raw data.

Wow, you talk about all that stuff earlier about a numerical count not necessarily being important but then use an example of a single test score having significance? Ever heard of a grade curve? Indeed my beef was never about a single numerical percentage but a comparison of two that have so little in common.

On a curve Mavericks wins because 51% of OS X users are using it only compared to 14% of Windows 8.x compared to Windows. That's fair in a sense. But I don't think I ever saw a grade curve where students didn't want to see the raw numbers.
 
On a curve Mavericks wins because 51% of OS X users are using it only compared to 14% of Windows 8.x compared to Windows. That's fair in a sense. But I don't think I ever saw a grade curve where students didn't want to see the raw numbers.
It's not even a good comparison, since Mavericks is an upgrade to an existing OS while Windows 8 is entirely new.

Now if it were comparing the adoption rate of Windows 8.1 for users of Windows 8 to the adoption rate of Mavericks for users of OS X, that might be relevant.
 
Percentages are pretty meaningless when you are comparing two sets of data of vastly different size.
Percentages, or any statistical measurement and analysis, can be valid or invalid for a variety of reasons. As a general statement, what you've said is false - there are many circumstances where it is actually desirable to compare data sets of significantly different sizes.

In this particular case of estimating adoption rates the amount of data for both MS and Apple is completely sufficient. (they don't mention any confidence intervals but it's reasonable to assume they are *extremely* small)
 
here's some interesting stats about a certain somebody:
- the subject uses 0 apple products
- apple products constitute 0% of their usage time
- the subject has written no apple software
- the subject, as of this post, has made 24.5% of the posts in this thread
- over 50% of the thread's content consists of discussion with the subject that has made exactly 0% progress
 
Wow, you talk about all that stuff earlier about a numerical count not necessarily being important but then use an example of a single test score having significance? Ever heard of a grade curve? Indeed my beef was never about a single numerical percentage but a comparison of two that have so little in common.
You said, generally speaking, you would ask for data when presented with a percentage comparison for any situation - apparently doubting the usefulness of a percentage measure in the absence of raw data. My point is only that this is obviously not true - there are many circumstances where it is totally sufficient (and even better) to simply see the percentage.


On a curve Mavericks wins because 51% of OS X users are using it only compared to 14% of Windows 8.x compared to Windows. That's fair in a sense. But I don't think I ever saw a grade curve where students didn't want to see the raw numbers.
I think I understand your position: you see this as winning vs losing. You think that presenting adoption rate statistics makes it look like OSX is winning and Windows is losing - but Windows has a hell of a lot more users so it must be winning. You think they should have presented a metric on how many Windows user's there are vs OSX users.

I don't care about winners and losers. I don't give a shit if Apple or MS go bankrupt or if the public is massively uninformed about which is more popular.

Let's remember: this was a developer's conference, and adoption rate data was presented to developers, who may quite reasonably be interested in this data. This is a very reasonable thing to do.
 
It's not even a good comparison, since Mavericks is an upgrade to an existing OS while Windows 8 is entirely new.

Now if it were comparing the adoption rate of Windows 8.1 for users of Windows 8 to the adoption rate of Mavericks for users of OS X, that might be relevant.

Maybe a curve isn't the best way to do it but comparing percentages of things where one is an order of magnitude large isn't useful either.
 
comparing percentages of things where one is an order of magnitude large isn't useful either.
This is totally untrue. Comparisons are done across orders of magnitude all the time.

Have you ever seen a logarithmic graph? It's entire purpose is to compare data across multiple orders of magnitude.
 
You said, generally speaking, you would ask for data when presented with a percentage comparison for any situation - apparently doubting the usefulness of a percentage measure in the absence of raw data. My point is only that this is obviously not true - there are many circumstances where it is totally sufficient (and even better) to simply see the percentage.

When comparing the percentages of two numbers that are based on those numbers that's never going to happen.

I think I understand your position: you see this as winning vs losing. You think that presenting adoption rate statistics makes it look like OSX is winning and Windows is losing - but Windows has a hell of a lot more users so it must be winning. You think they should have presented a metric on how many Windows user's there are vs OSX users.

This isn't it at all. You bought up a good point about cross-platform development. What does the fact that 51% of 80 million OS X users are on Mavericks mean compared to the fact that 14% of 1.3 billion Windows users are on Windows 8.x? Anyone trying to make an argument beyond this very simple point is just clouding the point.

I don't care about winners and losers. I don't give a shit if Apple or MS go bankrupt or if the public is massively uninformed about which is more popular.

I think it unlikely that either company will in our lifetime will

Let's remember: this was a developer's conference, and adoption rate data was presented to developers, who may quite reasonably be interested in this data. This is a very reasonable thing to do.

What's more interesting to a developer is the maximum total audience for any given version. No one here has even tried to touch the argument I've made with Mavericks vs. Windows 8.x because we all know that gaming on OS X is beyond a lost cause.
 
What does the fact that 51% of 80 million OS X users are on Mavericks mean compared to the fact that 14% of 1.3 billion Windows users are on Windows 8.x?
It means that OSX users are faster to switch to Mavericks than Windows users are to switch to Windows 8. This was a slide on OS adoption rates. Was that not clear to you?


What's more interesting to a developer is the maximum total audience for any given version.
There are lots of things that are interesting to a developer: adoption rates, total audience, developments tools, hardware architecture, etc. This particular slide was about adoption rates, not about those other things.


You said you didn't understand the point about comparing adoption rates. Many of us have pointed out that this information is interesting and/or useful to some people. Do you understand the point now?
 
here's some interesting stats about a certain somebody:
- the subject uses 0 apple products
- apple products constitute 0% of their usage time
- the subject has written no apple software
- the subject, as of this post, has made 24.5% of the posts in this thread
- over 50% of the thread's content consists of discussion with the subject that has made exactly 0% progress

But the subject does understand that when two pie charts are presented that people who think about numbers are going to want to know the numbers that are represented by those two pie charts.

Furthermore, say what you will about my use of Apple products, I've on many occasions defended Apple. I've flat out said that the iPad is a revolutionary device for some time now. Apple does some very great stuff. Google does some very great stuff. And Microsoft does as well. None of their core competencies overlap at this time. But I'm an old fart. I stick to what I know, in many ways I'm like the guy that couldn't figure out Windows 8. Except I spent 20 minutes to do so. Hell because I've used Windows tablets for so long I know more than the kiddies on Android tablets and iPads. It's all the same stuff with a different wrapper. Ignore the wrapper and it's pretty damned simple.
 
It means that OSX users are faster to switch to Mavericks than Windows users are to switch to Windows 8. This was a slide on OS adoption rates. Was that not clear to you?
This doesn't make any sense. Mavericks, like Mountain Lion, Lion, Snow Leopard, etc. before it, is OS X. Mavericks is simply OS X version 10.9. Windows 8 is not just a newer version of Windows 7.
 
Back
Top