App Stores are not the Future, says Google
Vic Gundotra, VP Google Engineering:
"What we clearly see happening is a move to incredibly powerful browsers."via Chris Nuttall, FT.com
"Many, many applications can be delivered through the browser and what that does for our costs is stunning."
"We believe the web has won and over the next several years, the browser, for economic reasons almost, will become the platform that matters and certainly that’s where Google is investing."
"I think Steve [Jobs] really did understand that, over the long term, it would be the web, and I think that’s how things will play out."
Why do I blog this:
The discussion of web apps vs. mobile apps is getting more and more interesting by the day. While discovery (65k apps vs. billions of websites) and local storage/data (e.g. access to phone sensors) are important issues to be considered, the most important is monetization. Most web apps have no established billing system, rather earn cents than dollars (which adds up only if you're Google and the like), while mobile apps do.
Interesting comments over at FT as well.
Posts
7 Comments:
There are differences for the distributionmodel of mobile apps and desktop apps. The model Google wants to see is the centralized one, or the "all your data belongs to us"-approach. Centralized distribution allows for the best control. This is also the reason why app stores flourish.
Web apps - apps which cannot work without beeing connected - are the dumb terminal approach. The Webbrowser is only used as the gui-display, all logic is behind the scenes on a central server.
I really understand that this *must* be Gogle's favourite approach to the future. Taking the perspective of a user, there do exist situations where you would really like to work without any connection to "the net". Especially if you think about privacy. Every package routed to your machine allows to more and more geolocate you and profile what you do.
I do not think, that centralization is the final answer to computing. Centralization carries huge risks usually incorporated into centralized solutions and monocultures.
I think there will soon be a market for decentralized approaches and products which respect the user better.
just my 2 cents
Interesting thoughts there, Helge. Thanks for weighing in.
I do agree that there are scenarios where a user would prefer offline capabilities: be it because of privacy concerns, roaming costs, speed or reliability.
Of course there are applications that are designed for offline use and make perfect sense not being connected (games, for example). Increasingly apps only unfold their full potential when they are connected - which does not mean that they have to be connected all the time.
A good example that comes to mind in which Google is already supporting offline functionality to broaden the use of their online apps is Google Gears. I'm using it to download unread items to my Google Reader before travelling. So Google's of course well aware of that :).
What about games?
The App Store is a very good system to distribute video games. 30% of applications on the App Store are video games and those apps could not be played in a browser.
@Ben: I also though of Google Gears before I posted my comment. That is clearly a solution to the problem of "I case of lost connection, please keep on collecting data."
When I think e.g. of using Xing to lookup someones contact data, why should that action leave a trace of someone having looked up something? Why not have a Desktop App, which allows me to access the data of MY contacts as often as I want without notifying the whole universe about this lookup-action? And there are similar functionalities built into many other services.
More and more webapps work this way to keep the steady arousal of attention in a flow-state. You get notifications about things which you do not really want to know and you expose parts of your own behaviour wether you want it or not. Using web-apps you commit to the model of exposure and the loss of control.
This takes away the choice to use computing functionality AND BEEING in control of the personal data. At the moment no one has a chance to break into a computersystem which is switched on only several hours a day - which is how most personal computers at home are used.
Compared to cloud-computing, a system like 24x7-cloud-computational data-centers is much more attractive for the hackers and all kind of people which want to get access to this data. This data is stored centrally and can be attacked 24x7. It's a data paradise!
Breaking into these centralized systems to find the "ultimate backstagepass" to other peoples data might be a vital target for some individuals. The huge security issue of twitter beeing hacked recently exposes some of these risks of centralization.
"Powerful Browsers" could also be interpreted as "Newspeak", because at the same time it means powerless users not beeing any longer in control of their data. The amazon-kindle "bookburning" just beeing another impressive demonstration of the danger of centralized systems disempowering the user step by step.
I agree, that having a connection enables great services. But I disagree with the way the users data is handled building centralized datastores simply can not be the solution to a stable infrastructure. If we continue building he web of the future this way, we will soon begin to suffer from the dependencies of centralized components failing with a huge whoohoo.
let us face the truth: the web browser is just the least common denominator in building dumb terminal apps which run everywhere.
If we want dumb terminal apps which loose many of the conveniences modern OS provide for the GUI (many are rebuilt with flash-extensions), than the browser is the most powerful OS you could get.
Yeah, HTML5 will and 3D-GFX-functionalities in CSS will blur this line even further, but it never will be the same. Thats, why the iPhone experience of native apps is so dramatically better than using a browser.
If the web-browser will win it will win at the expense of the user... beeing not any longer in control of his data, trading in fragile, centralized appfarms which carry the same risks any monoculture evokes, plus the risk of handing over one's own privacy of data to a huge company.
This is not exactly what I personally would prefer as a personal computing device for myself. ;-)
@Sane ... I mentioned games in my comment. And I do agree with you :)
@Helge ... thanks for this long comment. Let me try to find some replies to the points you make.
To start off, I believe we should clearly differentiate between apps that don't need any connection at all (pure offline apps like e.g. games) and all other apps - be it mobile or web apps that do connect to the web.
Whenever an app has to connect to receive or send information this information could potentially be intercepted, compromised or used to profile and track the user. This is no different for web (browser-based) or mobile apps.
Whether the continuation of a centralized handling of data will remain or whether each user might take care of his own data remains to bee seen. As long as it's an extra cost to the user (not only pecuniary but also in terms of convenience) I don't believe we'll see this move.
I agree with you that owning one's data is important (at Tagcrumbs, for example, you own your data and can export and download it) and should be respected by more companies.
@Ben: I also think games are somehow different here. Though many game console products and many successful games rely on the connection to the net (and even having own app stores).
I just was catched by the word "Powerful Browser" and I do think you have to ask: "Powerful, in regard to which aspect exactly?"
The data question just is deducted from the fact a browser is a dumb terminal and needs to store its data somewhere. For me a more powerful browser would be a browser capable of copying ad hoc data from large data silos onto my HD for offline usage. E.g. climate data, financial data, bibliographical data, and so on...
I love web apps which allow me to take my data with me. Delicious is a perfect example for this. I use it because I can do my own backup of all my bookmarks 24x7. I am free to go at any time! Delicious profits from every entry I tag and share. So this is a fair deal, while feeding data into something like qype.com does not really catch me and I am not able to export my own/favourited qype spots as KML-file e.g. to reuse it in Google Earth. It's the all-your-data-belongs-to-us-approach, while all-your-community-processes-belong-to-us would be totally sufficient.
just another cent.
Post a Comment