Maybe it's time that MacOS X get its own conference. Since this one was sold out with 100% of folks being iOS developers, it's quite probably that this survey does not represent the spread of folks who would have liked to attend such a conference.
Maybe it's time that MacOS X get its own conference. Since this one was sold out with 100% of folks being iOS developers, it's quite probably that this survey does not represent the spread of folks who would have liked to attend such a conference.
If we can believe the AI story, attendance would be down 93% compared to the current setup. Only 7% of attendees develop for OSX. And according to last year's numbers, OSX devs are dropping like flies. By this time next year, who knows how many OSX devs will even be left?
Apple is now a mobile device company, and the devs know it too.
A new survey of attendees at Apple's annual Worldwide Developers Conference found that nearly half of iOS developers attending the conference support Android, while just 7 percent write applications for the Mac.
This article title is a perfect example of a desperate attempt by AI of "trolling for hits". By comparing developers who write for Mac .... a desktop environment .... to Android ... a mobile environment is meaningless in its headline. The same charts that supply info for the headline .... state that 100% develop for iPod/iPhone and 93% develop for the iPad .... while only 47% develop for Android.
Of course a headline stating that would not attract the hits that your sleazy headline attracts.
AI, I swear, you're getting more like The National Enquirer everyday. I think it's time to look for a more credible outlet for my Mac news. ...\
The Android number is misleading and irrelevant. It means absolutely nothing. Might as well ask the attendees if they like pepperoni or anchovies on their pizzas. This is a Mac/iOS developer conference, people are not there to learn about Android. Everyone who paid to go to the conference was interested in programming Apple devices (though there might have been a few Google employees show up just to check out the competition).
A possible better metric would be to sample the individual sessions to see how full they are. Were the iOS ones packed with the Mac ones half or almost empty? Even that number would be inaccurate as the conference sold out very quickly. The might have easily sold double the number of tickets. And the Mac OS developers are a more seasoned group, since Mac OS has been around for 10 years. While some iOS development is done by these experienced Mac developers, iOS is newer and has attracted a following of developers that are not experienced with Cocoa programming.
Lower barrier of entry. Crap-ass developers have an easier time getting their apps on Android stores. Bug-laden app can go prime-time on Android in a heartbeat. The App Store actually screens your apps/code and easily figures out you have a grand total of 3.5 hours of experience in writing code.
The same reason a lot of people go to work for the government. Great pay, less work.
This article title is a perfect example of a desperate attempt by AI of "trolling for hits". By comparing developers who write for Mac .... a desktop environment .... to Android ... a mobile environment is meaningless in its headline. The same charts that supply info for the headline .... state that 100% develop for iPod/iPhone and 93% develop for the iPad .... while only 47% develop for Android.
Of course a headline stating that would not attract the hits that your sleazy headline attracts.
AI, I swear, you're getting more like The National Enquirer everyday. I think it's time to look for a more credible outlet for my Mac news. ...\
Let's not kill the messenger here - I didn't see anywhere that indicated that AI commissioned this survey or is making any sort of predictions or conclusions based on the results.
Yes the sample size and random sample etc are problems - as are the venue as well as the questions themselves can affect the outcome - but you have to start somewhere.
I see this as an interesting tidbit but not terribly useful - lets just watch that some media pundit doesn't start touting this as an indication that Android is replacing Mac OS any second now.... wait for it...
OK... So, according to http://www.mobiledevhq.com/developers, (the only reference I can find at short notice) there are at least 61,000 iPhone developers.
This means that Piper Jaffray's "survey" of 45 developers is a sample of around 0.074%.
As any qualified researcher will tell you, this sample is simply too small to be able to draw any meaningful conclusions about the entire body of IOS developers. To then go further and translate these small numbers into tables of percentages demonstrates heroic cretinism skills. And to then try to infer conclusions about Android or IOS market changes from this microscopic speck of data... Well that's beyond ridiculous.
This is like me stopping 45 people wearing shoes, asking them some basic questions and then claiming to have some insight into the global shoe market.
The fact that this sort of bogus, primary school spin is being hyped as 'research' by a so-called investment company suggests two things: (1) they are morons; and (2) they may have a hidden agenda...
Idiots!
Actually, the idiots are the ones who insist on commenting on topics they don't understand.
There is no such thing as 'too small' a sample size. Statistically, the larger the sample size, the more precise the results, but even for a sample size of 45 people, SOME conclusions can be reached.
At a 95% confidence level, the standard error for a sample size of 45 people is around +/- 15%. So, if the sample was random, one would have a 95% confidence level that between 32 and 62% of iOS developers were developing for Android.
That's not a great level of precision, but it IS valid.
There are two REAL problems with this survey:
1. There's no way of knowing if the sample is random. At the very least, the developers attending WWDC are a self-selected group and may not be representative of iOS developers as a whole. To compound that problem, they may have selected an unrepresentative group from the WWDC audience.
2. It's really useless. What conclusions can one draw from the survey that mean anything? Even if the number is correct and 47% of iOS developers also develop for Android, what about the converse? What percentage of Android developers also develop for iOS? And why should anyone care?
In terms of overall developers, a far more extensive survey, the 2011 edition of the VisionMobile Developer Economics survey, found that 67 % of developers polled are working with the Google platform, and 59 % with apple.
Way too small of a sample size to mean anything. This number represents less then 0.9% of the attendees this year and less then 0.4% of attendees in the past year. It is fairly rare to have a developer that writes for both, so somehow this survey is skewed. Maybe the developers had a co-worker that wrote for android? My impression was that the number is much smaller. I rarely found someone working on Android at the conference. I heard a lot of things like: Sometimes we write Android apps but generally iOS only. Most developers say Android owners don't buy anything, so I don't know where the growth potential comes from. Maybe the most growth for free apps.
Another question: why did only 45 developers reply? A number that low usually indicates that the survey was done on a subsample that isn't truly random, for example, at a meetup of Android developers at one of the after parties. Of course, it might not be that obviously skewed, but if it was a survey that appeared in every attendee's goodies bag, there should have been hundreds of responses.
Actually, the idiots are the ones who insist on commenting on topics they don't understand.
There is no such thing as 'too small' a sample size. Statistically, the larger the sample size, the more precise the results, but even for a sample size of 45 people, SOME conclusions can be reached.
At a 95% confidence level, the standard error for a sample size of 45 people is around +/- 15%. So, if the sample was random, one would have a 95% confidence level that between 32 and 62% of iOS developers were developing for Android.
That's not a great level of precision, but it IS valid.
There are two REAL problems with this survey:
1. There's no way of knowing if the sample is random. At the very least, the developers attending WWDC are a self-selected group and may not be representative of iOS developers as a whole. To compound that problem, they may have selected an unrepresentative group from the WWDC audience.
2. It's really useless. What conclusions can one draw from the survey that mean anything? Even if the number is correct and 47% of iOS developers also develop for Android, what about the converse? What percentage of Android developers also develop for iOS? And why should anyone care?
+++
You made the same points I was going for earlier, but far more eloquently!
At the keynote, Steve Jobs said the number of WWDC attendees this year was about 4500. The number of respondents in this survey (45) is relatively small (1% of total attendees). Thus, a simple formula can be used to calculate the single-sided margin of error: Sqrt( p(1-p)/45 ), where p is the fraction (between 0 and 1) of affirmative responses to each question.
The maximum margin of error at 95% confidence is +/- 7.5%, associated with a 50% affirmative response rate within the sample population. (The 95% confidence interval for the total population of attendees is from 42.5% to 57.5%).
For 3 out of 45 (~6.7%) affirmative responses, the margin of error is +/- 3.7%. (The 95% confidence interval for the total population of attendees is from 3.0% to 10.4%).
For 45 out of 45 (100%) or 0 out of 45 (0%) affirmative responses, the margin of error is effectively 0.
That was just an example for the person who didn't understand why the percentages added up to more than 100% for one of the survey questions. I was contrasting a multi-answer question with a multiple-choice (single-answer) question. I could have used this example for the latter instead:
Maybe it's time that MacOS X get its own conference. Since this one was sold out with 100% of folks being iOS developers, it's quite probably that this survey does not represent the spread of folks who would have liked to attend such a conference.
Incorrect, this survey was limited to the iOS developers present - 100% of the attendees were not iOS developers.
How does the market share compare between iOS and Mac OS?
I would imagine iOS has a much larger install base and much larger market share compared to Mac OS so it is not illogical to assume there would be more developer for the iOS Platonism.
Edit.
In Q1 2011 Apple sold
7.33 million iPads
16.24 million iPhones
19.45 million iPods
and
4.13 million Macs.
So 7.33+16.24+15.0=~40 million iOS devices
(Lets say the majority are iPod touches)
So out of ~44 million devices sold, 4.13/44=.09 or 9%
So, the developer figures for OS X seem on par with its market share compared to OS X in terms of devices sold by Apple in Q1 2011.
I'm interpreting the percentages as percent of the total population (n=20, n=45) and that there appears to be duplicated reporting of the data (e.g., an individual developes for multiple platforms).
So besides the poor research design, we now have confusing depicting of the data in tabular format.
Something else to consider is: of what relevance to the survey is the reporting of the average number of apps in the app store? First, there are only apple apps in the app store, no andriod or symbian. Second, they're reporting mean data? What's the range, median, mode and SD? Average means very little, and is often misrepresentative especially with regards to non-parametric data. Which is one reason why the median housing price is reported, not the mean.
In essence, collecting this data is irrelevant to the study, reporting the results is meaningless...
Actually I heard our dog, Rocky, saying it was 243%. Guess I'll have to cut out all that coffee he drinks.
I think this only reinfoces the view that OSX and iOS will merge into one sometime over the next few years: iOS X
erm - how on earth do you get to there from the article? The tiny overlap? If anything the article shows that the development bases are diverging and that there is no benefit in merging the OSes.
I believe you need a random sampling size of roughly 1,050 developers to get a realistic representation of the entire dev populous, assuming there are 61,000 of them.
No you don't. That sample size (1000+) is often used when extrapolating to millions in elections. It's not too hard to work out roughly what you need to get the "19 times out of 20" statement in the error. Like any sample the size required will depend on how alike the developers are.
Comments
Maybe it's time that MacOS X get its own conference. Since this one was sold out with 100% of folks being iOS developers, it's quite probably that this survey does not represent the spread of folks who would have liked to attend such a conference.
If we can believe the AI story, attendance would be down 93% compared to the current setup. Only 7% of attendees develop for OSX. And according to last year's numbers, OSX devs are dropping like flies. By this time next year, who knows how many OSX devs will even be left?
Apple is now a mobile device company, and the devs know it too.
A new survey of attendees at Apple's annual Worldwide Developers Conference found that nearly half of iOS developers attending the conference support Android, while just 7 percent write applications for the Mac.
This article title is a perfect example of a desperate attempt by AI of "trolling for hits". By comparing developers who write for Mac .... a desktop environment .... to Android ... a mobile environment is meaningless in its headline. The same charts that supply info for the headline .... state that 100% develop for iPod/iPhone and 93% develop for the iPad .... while only 47% develop for Android.
Of course a headline stating that would not attract the hits that your sleazy headline attracts.
AI, I swear, you're getting more like The National Enquirer everyday. I think it's time to look for a more credible outlet for my Mac news. ...
A possible better metric would be to sample the individual sessions to see how full they are. Were the iOS ones packed with the Mac ones half or almost empty? Even that number would be inaccurate as the conference sold out very quickly. The might have easily sold double the number of tickets. And the Mac OS developers are a more seasoned group, since Mac OS has been around for 10 years. While some iOS development is done by these experienced Mac developers, iOS is newer and has attracted a following of developers that are not experienced with Cocoa programming.
The same reason a lot of people go to work for the government. Great pay, less work.
This article title is a perfect example of a desperate attempt by AI of "trolling for hits". By comparing developers who write for Mac .... a desktop environment .... to Android ... a mobile environment is meaningless in its headline. The same charts that supply info for the headline .... state that 100% develop for iPod/iPhone and 93% develop for the iPad .... while only 47% develop for Android.
Of course a headline stating that would not attract the hits that your sleazy headline attracts.
AI, I swear, you're getting more like The National Enquirer everyday. I think it's time to look for a more credible outlet for my Mac news. ...
Let's not kill the messenger here - I didn't see anywhere that indicated that AI commissioned this survey or is making any sort of predictions or conclusions based on the results.
Yes the sample size and random sample etc are problems - as are the venue as well as the questions themselves can affect the outcome - but you have to start somewhere.
I see this as an interesting tidbit but not terribly useful - lets just watch that some media pundit doesn't start touting this as an indication that Android is replacing Mac OS any second now.... wait for it...
OK... So, according to http://www.mobiledevhq.com/developers, (the only reference I can find at short notice) there are at least 61,000 iPhone developers.
This means that Piper Jaffray's "survey" of 45 developers is a sample of around 0.074%.
As any qualified researcher will tell you, this sample is simply too small to be able to draw any meaningful conclusions about the entire body of IOS developers. To then go further and translate these small numbers into tables of percentages demonstrates heroic cretinism skills. And to then try to infer conclusions about Android or IOS market changes from this microscopic speck of data... Well that's beyond ridiculous.
This is like me stopping 45 people wearing shoes, asking them some basic questions and then claiming to have some insight into the global shoe market.
The fact that this sort of bogus, primary school spin is being hyped as 'research' by a so-called investment company suggests two things: (1) they are morons; and (2) they may have a hidden agenda...
Idiots!
Actually, the idiots are the ones who insist on commenting on topics they don't understand.
There is no such thing as 'too small' a sample size. Statistically, the larger the sample size, the more precise the results, but even for a sample size of 45 people, SOME conclusions can be reached.
At a 95% confidence level, the standard error for a sample size of 45 people is around +/- 15%. So, if the sample was random, one would have a 95% confidence level that between 32 and 62% of iOS developers were developing for Android.
That's not a great level of precision, but it IS valid.
There are two REAL problems with this survey:
1. There's no way of knowing if the sample is random. At the very least, the developers attending WWDC are a self-selected group and may not be representative of iOS developers as a whole. To compound that problem, they may have selected an unrepresentative group from the WWDC audience.
2. It's really useless. What conclusions can one draw from the survey that mean anything? Even if the number is correct and 47% of iOS developers also develop for Android, what about the converse? What percentage of Android developers also develop for iOS? And why should anyone care?
Read more:
http://www.visionmobile.com/rsc/rese...omics_2011.pdf
Way too small of a sample size to mean anything. This number represents less then 0.9% of the attendees this year and less then 0.4% of attendees in the past year. It is fairly rare to have a developer that writes for both, so somehow this survey is skewed. Maybe the developers had a co-worker that wrote for android? My impression was that the number is much smaller. I rarely found someone working on Android at the conference. I heard a lot of things like: Sometimes we write Android apps but generally iOS only. Most developers say Android owners don't buy anything, so I don't know where the growth potential comes from. Maybe the most growth for free apps.
Another question: why did only 45 developers reply? A number that low usually indicates that the survey was done on a subsample that isn't truly random, for example, at a meetup of Android developers at one of the after parties. Of course, it might not be that obviously skewed, but if it was a survey that appeared in every attendee's goodies bag, there should have been hundreds of responses.
Actually, the idiots are the ones who insist on commenting on topics they don't understand.
There is no such thing as 'too small' a sample size. Statistically, the larger the sample size, the more precise the results, but even for a sample size of 45 people, SOME conclusions can be reached.
At a 95% confidence level, the standard error for a sample size of 45 people is around +/- 15%. So, if the sample was random, one would have a 95% confidence level that between 32 and 62% of iOS developers were developing for Android.
That's not a great level of precision, but it IS valid.
There are two REAL problems with this survey:
1. There's no way of knowing if the sample is random. At the very least, the developers attending WWDC are a self-selected group and may not be representative of iOS developers as a whole. To compound that problem, they may have selected an unrepresentative group from the WWDC audience.
2. It's really useless. What conclusions can one draw from the survey that mean anything? Even if the number is correct and 47% of iOS developers also develop for Android, what about the converse? What percentage of Android developers also develop for iOS? And why should anyone care?
+++
You made the same points I was going for earlier, but far more eloquently!
The maximum margin of error at 95% confidence is +/- 7.5%, associated with a 50% affirmative response rate within the sample population. (The 95% confidence interval for the total population of attendees is from 42.5% to 57.5%).
For 3 out of 45 (~6.7%) affirmative responses, the margin of error is +/- 3.7%. (The 95% confidence interval for the total population of attendees is from 3.0% to 10.4%).
For 45 out of 45 (100%) or 0 out of 45 (0%) affirmative responses, the margin of error is effectively 0.
But who picks only one?
That was just an example for the person who didn't understand why the percentages added up to more than 100% for one of the survey questions. I was contrasting a multi-answer question with a multiple-choice (single-answer) question. I could have used this example for the latter instead:
What's your favorite barbeque food?
1. Hamburgers
2. Hot Dogs
3. Ribs
Maybe it's time that MacOS X get its own conference. Since this one was sold out with 100% of folks being iOS developers, it's quite probably that this survey does not represent the spread of folks who would have liked to attend such a conference.
Incorrect, this survey was limited to the iOS developers present - 100% of the attendees were not iOS developers.
I would imagine iOS has a much larger install base and much larger market share compared to Mac OS so it is not illogical to assume there would be more developer for the iOS Platonism.
Edit.
In Q1 2011 Apple sold
7.33 million iPads
16.24 million iPhones
19.45 million iPods
and
4.13 million Macs.
So 7.33+16.24+15.0=~40 million iOS devices
(Lets say the majority are iPod touches)
So out of ~44 million devices sold, 4.13/44=.09 or 9%
So, the developer figures for OS X seem on par with its market share compared to OS X in terms of devices sold by Apple in Q1 2011.
I'm interpreting the percentages as percent of the total population (n=20, n=45) and that there appears to be duplicated reporting of the data (e.g., an individual developes for multiple platforms).
So besides the poor research design, we now have confusing depicting of the data in tabular format.
Something else to consider is: of what relevance to the survey is the reporting of the average number of apps in the app store? First, there are only apple apps in the app store, no andriod or symbian. Second, they're reporting mean data? What's the range, median, mode and SD? Average means very little, and is often misrepresentative especially with regards to non-parametric data. Which is one reason why the median housing price is reported, not the mean.
In essence, collecting this data is irrelevant to the study, reporting the results is meaningless...
Actually I heard our dog, Rocky, saying it was 243%. Guess I'll have to cut out all that coffee he drinks.
I think this only reinfoces the view that OSX and iOS will merge into one sometime over the next few years: iOS X
erm - how on earth do you get to there from the article? The tiny overlap? If anything the article shows that the development bases are diverging and that there is no benefit in merging the OSes.
I believe you need a random sampling size of roughly 1,050 developers to get a realistic representation of the entire dev populous, assuming there are 61,000 of them.
No you don't. That sample size (1000+) is often used when extrapolating to millions in elections. It's not too hard to work out roughly what you need to get the "19 times out of 20" statement in the error. Like any sample the size required will depend on how alike the developers are.
philip