written on 10-15-2013

1. A new opportunity for everyone, not just the big players.

I have always considered the big players of Internet invincible, even in the Mobile Era. But lately I have a 'Eureka moment' when browsing the the apps installed on my iPhone: of the ~50 apps installed, only less than 10 are products of the big players. I guess most smartphone users will have a suite of apps with similar make-up. This means, in terms of visibility on the mobile platform, the big players don't have an overwhelming advantage compared to small players. They surely get the upper hand in brand recognition, marketing budget, etc. But mobile platforms are by and large uncharted territories (not so much today), and the new entrants stand a good chance competing with any mature internet company.

There are several differences between the mobile era and desktop era:

A much larger market. More people are buying smartphones, more people are using smartphones to get things done, people are spending more time on their phones than on PCs, annual PC sales drop for the first time in computing history. In a word, computing is transitioning from desktop to mobile.

This market can accommodate more players. In the desktop era, you expect users to remember your domain name. But users just don't hold many domain names in their memory, or their bookmark folder. Ask anyone the websites they browse regularly (no less than once a week), I guess the list is very short. In desktop era, the browser is the only entrance to Internet, and most websites are just not visible from there. But in the mobile era, when the user looks at their phone screen, there are 20 apps in just a single page, and they are just one tap away. Even if you are Google or Tencent (in China), you occupy just one slot in these 20, the user is as likely to tap on any other app as Google Maps or QQ.

When users have 100 apps on their smartphone, each app gets only a very small percentage of the user's attention, on average. The app usage follows a long tail curve, some apps gets opened every day, others just got forgotten slowly. But no single app will dominate the internet for a long time. There are limitless choices out there, when people get bored, they get a replacement. Apps come and go. To survive in the mobile era, a company needs a strategy to deal with the short life cycle of products.

To sum up, in the mobile era, there will be more players. All of them could be moderately successful, but they will not become monopolies like Google or Microsoft did. The BIG-3 or BIG-5 will be replaced by Best-100 who produce value for different aspects of people's lives. As Paul Graham said, it will be a 'high-resolution' society.

2. Domain name is dead, a new service is called for.

In desktop era, when the browser is the only entrance to Internet, domain names are very important. Whether a domain name is easy to remember could mean life or death for a company (take the lesson of kaixin001.com). Technically, a domain name is used to map a name to IP addresses, so that when a website swap servers, users don't notice a thing.

But in the mobile era, the domain names are becoming useless. Apps are accessed by a tap on an icon, mobile web pages are mostly accessed through a QR code scan. People use a service without knowing its domain name, so domain name loses its value as a mnemonic device.

Technically, you can map any number or character string to IP addresses. This string doesn't need to be meaningful, like the domain name, since it's invisible from the end user. And, as the internet technologies evolve, companies need more fine-grained control than the IP addresses. Let's call them access points, they could be APIs defined for a service deployed on different servers. If we map strings to these access points, the services become more flexible.

I think a new kind of service will be born, It will slowly take the place of DNS system. And the process will not take long if it beats DNS in efficiency.

3. App or Web? It is a question.

Right now, when you talk about Mobile Internet, it's all about apps. Mobile Web still lags behind in user experience and performance. But are apps the final form of Mobile Internet? I guess not. Firefox released Firefox OS, featuring web apps on home screen. And some says Google bets more on Chrome OS than Android, in spite of Android's predominance in mobile OS market.

Mobile Web is still superior than apps in 3 aspects:

Development: Development teams can be really agile with the web, releasing new features/bug fixes on a daily/hourly basis. Apps have much longer release cycles, not to mention the annoying 'review process'. There is no A/B testing with pure native apps. So it is hard to achieve fast iteration. And with native apps, you have to develop at least 2 versions: iOS/Android.

Interconnection: Web sites are inter-connected. This is one key value of internet - making more information available to you by following links. Apps, on the other hand, are mostly isolated. Other than authentication with open platforms, few apps direct users to other apps. All try to keep the user using his app for as long as possible. But open inter-connection provides the most value to users, that is why the internet thrives in the first place.

Searchablity. Web content are searchable. Search engine are so important, that Google actually marks a new era in internet age. But with apps, information are sealed within the app itself, there is no easy way to crawl the app content, thus they are not searchable. A new kind of search engine that index the app contents will be needed.

So in my opinion, the native apps have to solve all these problems, or mobile web will gradually take over. The web use experience doesn't need to exceed that of native apps, it just need to be 'good enough', because these 3 factors will push both the service providers (developers) and users to the web side.

To facilitate agile development, the apps need to be more flexible in terms of software structure. Now for most apps, only data are loaded from server, the code (program logic/user interface) are still quite static. It is possible that a large part of logic/UI can be dynamically loaded from server as well. The remaining static native part of code are just basic scaffold or native UI components. The dynamic code can be the form of a declarative format or scripts. Either way, the native code need to cache and parse them, there will be some overhead, but it will not matter when Moore's law kicks in. Currently, the hybrid apps are just a form of what I described: PhoneGap uses HTML-JS in a web view and wraps it with native device features; Titanium uses a javascript engine to translate JS into native code.

The app SDK provides the API to redirect to another app. But it doesn't provide a way to 'go back', this is one of the reasons why apps are reluctant to link to other apps. If a standard is established to make sure everyone play by the rules (make sure the user has the option to go back when he finishes the task in the app redirected to), maybe links will form among the apps.

The app search engine will require the apps to open up their internal data for indexing. All services should implement a common interface for spiders to gather data periodically. Information security has to be guaranteed with authentication between the search engine and the app server. The search engine app has to get the list of supported apps installed on the user device. When user initiates a search, the search engine can rank the installed apps on the top of results, other results will provide either mobile web links or download links for the native app. It takes a leap of faith for apps to open up all their data, but even partial data like keywords can get the search service started.

When apps evolve like this, the app will resemble the web. Maybe it is not using HTML/CSS/JS, but the concept is the same. Whether this new web will coexist with HTML web, only time will tell.