It's now becoming very clear that the year of Linux will never come. It is obviated by a cross-platform free web browser that gives you access to an open web "OS." For now, Chrome (or Chromium) is that browser. Google snuck around the free operating system problem by changing what "operating system" means.
None of those are remotely Unixy from a user perspective.
Neither the iPhone nor WebOS ever expose a filesystem to the user. Neither Android nor WebOS expose Unix to App developers. Android doesn't even present a real Unix system to the OS developers!
I wonder what will happen in the super long term. I think things sortof oscillate. We have desktop apps, people stopped buying them because of web apps, but then mobile web apps aren't as good as local mobile software so that started up again. Perhaps some day in the future we'll be talking about installing local apps onto ChromeOS like it's a brand new feature.
ChromeOS uses X Windows, just without a reparenting window manager.
ChromeOS is not Unix 'hidden' under a pretty UI -- neither the filesystem nor the userland are exposed to the user or to application developers. There's no Finder as the main UI with a Terminal in the Utilities folder: both metaphors are simply gone.
And that Android brought Linux to the mobile masses.
I suppose you can just think of Chrome OS and Android as just another couple of distros among the thousands already available. Same kernel, customized UI & apps.
Fundamentally, that's what Linux is about - a FOSS platform that anyone can modify and redistribute.
Android uses a heavily hacked kernel and essentially ignores the standard interfaces for everything it's userland stuff touches. It's also less of a distro than Slackware was -- every update has to be ported to the existing hardware platforms! Google finally wised up and will soon start shipping all the user-facing stuff through the Marketplace instead of tied to a kernel release.
My startup, AppStoreHQ (http://www.appstorehq.com), has been private alpha testing a mobile web app (HTML5) app store for the past couple months. It's been working out great and having Google come in for standard web apps should be a big wind in our sail, I think.
Our mobile web app store focuses on distribution and monetization of developers' work.
Here's an example app (it's an HTML5 frontend to foursquare) called fortysquires:
"having Google come in for standard web apps should be a big wind in our sail, I think."
I'm not sure I understand, Isn't google now a direct competitor. They are now competing with you, recurly, autorize.net and every other billing company. Everyone will now use google checkout to pay monthly fees and their Gmail user name to sign into web apps?
1) A rising tide lifts all boats. I suspect this isn't a zero-sum game.
2) Chrome Web Store appears to be more geared towards desktop web apps. We're solely focused on mobile web apps. This is a key differentiator as the use case, experience, and expectations are all different. But Google pushing web apps as a legit form of apps will help us as we push mobile web apps as a legit form of mobile apps.
Hello, my name is Aaron, and I will be your question answerer for today.
- Anything that works in Chrome can be an app. Flash apps are OK (one of the demo apps at the keynote, Plants vs Zombies, was a flash app)
- We haven't finalized exactly how payments will work. We know it has to be good, because it needs to be better than what you could do yourself.
- The web apps in the store will most likely work with Mozilla, Opera, Safari, etc, because those are pretty modern browsers that implement most of HTML5. But the tight integration with Chrome (the "install" experience) is Chrome-only, at least for now.
I don't understand the last point: if I can only purchase and install them from Chrome, how do I run them in another browser? Just point the non-Chrome browser at the (un-DRMed?) local install after purchasing/installing with Chrome?
Apps are essentially a metadata wrapper around a URL^. The payment authorization is done server-to-server. Thus you could buy an app in Firefox and authorization would still work fine, the UX just won't be as nice.
^ We are also considering supporting fully packaged apps that are completely offline, like Chrome extensions are today -- these won't work in other browsers because they don't understand our package format.
Using bundled packages for offline support is entirely the wrong strategy. The coupling is too tight even before you start trying to use the same format for multiple browsers.
You should just copy Apple and just make it easy to save web pages as local apps, though you might want to explicitly offer with an infobar when there's a manifest available. What they already look for is all you need:
<html manifest="..."> and window.applicationCache
<meta name="application-name"> and <meta name="application-url">
<link rel="icon" sizes="..."> or <link rel="apple-touch-icon">
The capability model for requesting extra permissions is terrific, but it should have been done using <meta> or <link> tags -- reusing their bundle format for extensions was a terrible idea, especially since they're going to have to implement all the DOM-based stuff anyway.
I think they would just post their app as "free" in the store and handle payments themselves. Or support both payment systems if the distribution is worth it to them.
Will there be support for one login to multiple apps? Would be useful to essentially have a Chrome Apps login that then provides auto-login/off for the other apps installed on Chrome
Gears was the alpha implementation of the work that went on in WHATWG. All the Gears APIs (except for WorkerPool) were refactored significantly and integrated into the spec that became HTML5.
This is really great and I've been waiting for GOOG to do this for a while. The web is the dominant platform for the future. Having companies like Google making rich HTML5 apps become the standard is a huge boon. Some have already asked us how this plays into Cloudomatic long term. Like Iseff said, it's a big wind in our sail. Not just ours or appstorehq's but for web app devs as a whole.
Soo, it looks like the store is kind of advanced app review directory where instead of urls you click on shortcuts which smartly translate the url into an icon + metadata inside the browser, i.e. "install" the app, and if need be charge you for that.
Seems like a problem will be with having to pay for something which you can get for free by just going to the web site directly. If on the other hand developers don't allow direct access to the apps via normal urls, then obviously the competitors will do.
So, the store will be filled with loads of free apps, promoting paying versions. That's probably how its best to be used.
Our startup, Favetop (http://favetop.com) provides similar shortcut functionality for web apps along with direct site search capabilities. We also allow users to save all their favorite online media alongside their web apps with the ability to share these with friends and followers if they choose.
Awesome! This is the first real confirmation that Google is going to ship Native Client to end users as a supported technology. They'd intentionally kept it off the agenda when introducing ChromeOS before, so it wasn't clear to me that they were going through with it.
I need a VNC/RDP client and an terminal emulator with SSH as NaCL apps. I'd prefer if they were implemented in Go as wrappers around standard C libraries, exposing JS APIs to the DOM where at least the UI chrome would be implemented, and with the settings/data saved in HTML5 LocalStorage.
It's been on my to-do list for a while, but maybe now that Google is promoting it someone else will fill the hole so I don't have to :)
Haha, that one. I thought it was the title of a slide and didn't even give it a second thought because the video was not playing. Funny how attention works.
I wonder what long term effects this has on domain names. If people need only to remember how to use iTunes or Chrome to find and use an app, will you really need your app.com domain anymore?
There are lots of finicky details to get right about entrance/exit from apps. Hacker news is an example of one important class of apps that we hope to make work well.
I don't quite understand the concept of this AppStore for the web. What kind of apps does it sell? a stand-alone Flash based game/app? stand-alone web-app (with possibly no server behind it, pure JS)?
A fair amount of people have complained that they switched to Chrome from Firefox because FF is overly bloated, takes a long time to start up, and so on. Seems to me that a browser that also runs separate apps like this would take a considerable hit in load times and performance.
Chrome does really well to make sure that one tab doesn't interfere with another. Furthermore these paid apps would only load when you want to use them. Why would running a paid app in a tab have anything to do with startup time or performance?
Yes. But it's not just that. The key is that the user is buying an app that essentially get's "installed' onto his computer like a traditional application.
I'll be interested into seeing what happens if a user wants to install an app while browsing with IE. Will the App store download and install Chrome, and set Chrome up just to run the app?
