Cross platform is tricky. It seems like a small “technical” buzzword but actually, it is one of the biggest challenges for many technology companies and has different aspects for different people in the organization and outside of it.

Developer Point of View

It all starts with the fact that applications can potentially be targeted towards different computing devices. To get more people to use your applications you would like it to run on more and more device categories whether it is different smartphone operating systems or a desktop computer vs. a tablet.

I’ve met the term cross-platform in my first job (20 years ago) as a developer after I left the army and that was when we coded an antivirus scanning engine. We’ve built it purely in C to make it “compilable” and “runnable” on different desktop and server operating systems without being aware that we were building a cross-platform product. Today when you search for the term cross-platform on google you can find app developers challenged at running their apps both on iOS and Android. The aspiration to have a cross platform code base lies in the economic rationale of write once and run everywhere instead of developing again and again for each proprietary coding language and standards of each platform. Cheaper to develop and easier to maintain.

Sounds easy and good, no? Well, no. Even today after so many years of evolving development tools. The main reason it is not straightforward is the simple fact that each platform, when you go into details, is different than others, either by hardware specifications or by operating system capabilities, and at some point, you will need to have a piece of code that is platform specific.

For example, let’s take iOS and Android: on Android you have the ‘back’ button and on iOS you don’t. To make sure your code behaves “naturally” on Android you need to add some Android specific code to handle the ‘back’ action while it will be useless on iOS.

Cross-platform tools have evolved quite a bit, tools such as html5 based mobile app development environments.  Still, I’ve never seen a real application which was built in full using only cross-platform code. There is always the need to tweak something for a specific device or specific platform, there is no escape from it.

I always wondered why platform providers (Google, Microsoft, Apple…) have never bothered too much to support such cross-platform tools and even more, they seem always to make life much more difficult for such tools. I can understand the rationale of “not helping my competition” though I think that at some point in life the basic fact that not a single platform will win all the users sinks in. It may be more productive to apply cooperative strategies vs. only competitive ones. Indeed they may loose some developers to other platforms while at the same time they will win some switching to theirs and most important is that it will make developers’ life easier and with apps which will end with a good result for everyone.

QA Guy/Girl Point of View

For the QA team, cross-platform means usually a pain in the neck. First, you need to test it across different environments and life could have been so much easier if it was on one platform. Even supporting one platform is not easy nowadays due to versioning – iOS as a mild example for complexity and Android which is catastrophic due to its fragmentation.

The other aspect which is more problematic is the fact that developers which work with cross-platform tools are somehow shifting the “responsibility” of making sure their results are working properly and putting it on the tool itself to blame. As if they were doing the best they could and complying with whatever was requested of them and the fact that it does not work is not their responsibility. This state of mind automatically moves the blame to the person who found the bug, hence the QA person. Eventually developers are fixing whatever is needed but still, it is not the same situation as in the case of a platform specific developer and QA person. Maybe it is because the developer can not practically run all the tests on all the devices prior to handing the software which always leaves some quality gap “open”.

In general, QA have become highly challenged with the multitude of different devices out there which are very different one from the other. Previously ( a long time ago) you had Microsoft Windows for personal computers and Unix based servers. Now you have lots of operating systems, numerous hardware configurations and ever accelerating pace of releasing new OS versions so it does not make life easy, to say the least for the people who need to ship the software. Now add to that a cross-platform product:)

Product Manager Point of View

The product manager sees cross-platform from a whole different angle and that is more close to users’ perception. Cross platform is more about what people do with their devices, when and how they use them and how the product can adapt itself to the unique device-user context. For example on a smartphone, you might expect a “time wasting” behavior or a very efficient task-oriented behavior for getting something done vs. on tablets which can be used in more relaxed times driving different behavior. The challenge here is to really understand how your target audience can and may consume your products via each specific device and platform and how to adapt each platform specific version to serve that behavior. Of course, it contradicts the basic aspiration of the R&D division to write less platform specific code.

Marketing Team Point of View

The marketing team sees cross-platform as an opportunity. In a way, they are the only ones who don’t see the “burden” and try to enjoy the potential distribution hidden in the rich set of devices out there. More devices, regardless of their type, represent more users/consumers and that means more market. Sometimes each device reflects a specific market segment which carries on an overhead of reaching out to them such as in the case of specific gaming consoles and sometimes your target market just happens to be diverse in terms of consumption device and the users use a different kind of devices and platforms such as in the case of smartphone users.

The User Point of View

Users are kings of course and they want everything to run everywhere. Today it seems even “not ok” for an application to be only available on one platform as it can be even a sign of “laziness” of the provider or lack of attention the developer gives to the market. What really spoiled users is the web which is cross platform by nature and for users, it is too much to understand why Gmail is available everywhere and not my favorite iOS calendar app.

And that’s ok, they should not be bothered by that as they are kings.

One of the main problems with Android for app developers contemplating on Android vs. iOs is the fact it is highly fragmented. On  iOS you, unconsciously, know that you need only to build one version (Let’s keep the example simple) and it will work on all devices, you know that Apple is doing everything to make sure everyone has the latest version and that there is a decent level of backward compatibility.  For Android developers things have turned up differently, due to the way Android is “openly” distributed, you can not be rest assured that your app will run the same way or will even run at all on your users’ devices. Different incompatible Android versions, devices with different capabilities, OEM customizations and plain third party OS customization turn each Android device to be different then the other and that is usually a bad sign for developers. This infographics says it all.

Android fragmentation is a topic that has been discussed and acknowledged quite thoroughly in the industry and that is not what I want to uncover here. On aspect of the fragmentation, which has been neglected and left out of discussion but has no lesser impact on the apps industry is the variety in terms of screen dimensions, resolutions and input capabilities. This variety in devices input and output capabilities does not really impact you if you are developing apps with minimal user interaction. Thinking of it then it is quite hard to come up with such an example in today’s tablet/smartphones world. Most of the apps today are “intensively” focused on user experience needless to say judged totally by the experience itself. To design and develop a “good” user experience towards a single target device, with known screen and input capabilities is something feasible, taking into account even just one more device category then you are in a serious problem.

From a philosophical perspective, I think a “nature” law can be suggested on how good a user experience of an app would be based on how many target platform it addresses, the more platforms targeted, the worse the experience becomes.

Few thoughts in regards to this dark side of fragmentation are:

• web technologies ease the pain a bit since they allow “clean” separation of logic and presentation where the cost of customizations for additional platforms is just marginal. Of course for some app categories it is not an option since the experience has to be so “immersive” which make the browser as a container too restrictive.
• gaming and content driven apps suffer the most here
• google tries to minimize this by creating guidelines and removing the most problematic customize able “edges” from their user interface libraries but I think the problem is more rooted then that
• if someone feels a dejavu with Java ME then I got it too:)

One practical suggestion for Google to help developers in their decision of whether to target Android is two fold: first, admit the problem. Second provide a live decision making tool that will allow to filter the their existing user base by their devices (or whatever info they got from activations).

I am a big fan of apps! Both as an apps developer and as a smartphone user started way before the days it was even called a smartphone. I own several phones with all possible operating systems and never miss a chance to install any new app I encounter. I may be a major factor in the total 2011 downloads number in appstores:)

Following this self-proclaiming manifest and after I achieved credibility as someone who knows something about apps, I want to go back to the question in the headline. Sarah Perez story on the end of 2011  A Web of Apps  started with the following lines:“It is remarkable to think that we’re in the early days of the app era, when there are already close to 600,000 iOS applications and nearly 400,000 on Android (source: Distimo).”. For me, these lines assume a-priory that the number of apps will keep growing a lot, at least much more than 1,000,000 apps!, a common notion nowadays.

I think that every person that is somewhat related to the apps industry assumes whether publicly or silently in their “heart” that this number will grow; otherwise, there is no much of an industry in it, isn’t it so? Of course one can argue that each app can “fatten” up and make more money on its own which means to grow vertical and not horizontal. In my eyes growing vertical after a hyper horizontal growth period is usually a sign for the beginning of a market saturation phase.

To put it in the right perspective, I personally do feel it will grow much more (I got to, I am highly invested in this assumption:) so let’s not start pessimistically:) even considering the mere fact there are so many others not using smartphones and many audiences not being addressed etc.. etc.. then we are ok, aren’t we?

To answer the question above I want to see if we can somehow establish our beliefs on some rationale that is at least discussable. Or in other words, at least something to help the ones who are highly invested in it to be able to get a good night sleep calmly.

I think mobile apps and their acceptance present a major breakthrough in the computing world and this happened only thanks to the fact that people actually “met” these apps and discovered their existence (thanks Apple for creating the first effective apps distribution channel). The point of convergence of “capable” computing mobile devices, with good enough to go network connectivity and dynamically loadable small functional units called apps actually created something amazing – the ability to “upgrade” yourself instantly. It always reminds me of the scene from the Matrix where Neo loaded up the Kung Fu learning software and in a minute he was a Kung Fu master. I know we are not there yet but the metaphor has been established. The ability to load new functionality on demand on a computing device, which is actually your avatar since it goes everywhere with you, and then to be able to operate it is quite a leap from the usage perspective.

So if we follow this line of thought then we can predict two trends in the apps world for the near future:

1. Apps will become narrower in functionality within time. We are actually witnessing this trend already where you can see every day a new app achieving a very narrow purpose. This won’t mean there won’t be a place for big “Photoshop” or “Office” like packages which are more like a wholesaler in a box, we will see those but they will not be the majority, not even close. This trend will occur simply because people have shorter and shorter spans of attention and more and more specific needs which can be answered efficiently only by something narrow enough and simple enough to learn and operate it immediately. In general, I put the “blame” for this trend on the low bandwidth we have available today for communicating with our phones and unless someone will invent a direct injection of new functionality into our brains we are of stuck with the growing trend of simpler apps. Just to put it in context, narrow apps does not mean “dummy” apps. Actually, the narrowing of the apps happens in terms of explicit functionality while there is an expansion in the implicit dimension of making the app adaptive to the specific context of the user which enables easier operation of the suggested functionality. This adaptation may require more technological investment in the app then other explicit features.  For example, Siri has a very simple and narrow explicit functionality (she just listens and talks back with one button) while behind the scenes it holds a huge technological effort.

2. Thanks to the growing number of sensors and interfaces on mobile devices (and I know of few more developments in this area which are exciting) as well as the better connectivity options we will witness more and more human needs that will be addressable by apps. For example, the accelerometer which now assists people who run with their calculations and effort tracking, something that was not possible before this sensor was available.

These two trends point towards a growth horizon that is very clear that lies on the axis of the diverse set of unique and shared human needs. If we add to this equation locality, languages, gender, ages, cultures, religions and other social grouping criteria then we run into a very big number. A number that is big enough to take us to the point where we will stop counting how many apps are out there in the very near future.

One question that I still haven’t got an answer for is in the title of the original story that got me started “A web of apps”.  The question is whether we can do a comparison to the growth of the world wide web. I know it is not necessarily what Sarah Perez intended to discuss in her story, which was more about potential apps discovery by connecting apps. Still, there are many talks about whether apps will grow in the same growth pace as websites did. Websites grew and keep growing mainly on the axis of topical interest or knowledge areas and from a gut feeling it feels like it represents a much bigger growth axis. Maybe more on this in a later post.

What do you think? Will it grow forever?