There. I said it.
For a while now I've been getting quite frustrated and annoyed by the writing I've read in various places about the idea of a WebOS (or this view), Cloud OS (if you're Microsoft), Google OS, Yahoo OS (yuck--too many "o"s), and so on.
There is no Web Operating System. There will be no Web Operating System.
Computers need operating systems but networks don't (not at the OSI layers I'm interested in, at least). A Web Operating System is a myth propagated by people who either don't understand the Web, don't understand operating systems, or both. A good example is this TechCrunch story where mimicking a desktop interface is confused with an operating system.
Anyone who thinks so is on crack and I just didn't know how to explain it until recently. But, as usual, the moment of clarity came when I was talking to someone else and not paying attention to what I was saying until the words were already out of my mouth--a problem that has landed me in hot water more than a few times.
Luckily, two of my coworkers caught on to what I was saying and managed to help put it into context a bit. First off was Matt McAlister (who runs YDN, the group I work in). In The Business of Network Effects he does a good job of explaining how businesses and services in a network are fundamentally different from those which are isolated islands.
Recalling a brief conversation we had a couple weeks ago, he says:
Jeremy Zawodny shed light on this concept for me using building construction analogies.
He noted that my building contractor doesn’t exclusively buy Makita or DeWalt or Ryobi tools, though some tools make more sense in bundles. He buys the tool that is best for the job and what he needs.
My contractor doesn’t employ plumbers, roofers and electricians himself. Rather he maintains a network of favorite providers who will serve different needs on different jobs.
He provides value to me as an experienced distribution and aggregation point, but I am not exclusively tied to using him for everything I want to do with my house, either.
Similarly, the Internet market is a network of services. The trick to understanding what the business model looks like is figuring out how to open and connect services in ways that add value to the business.
Bingo.
The web is a marketplace of services, just like the "real world" is. Everyone is free to choose from all the available services when building or doing whatever it is they do. The web just happens to be a far more efficient marketplace than the real world for many things. And it happens to run on computers that each need an operating system.
But nobody ever talks about a "Wall Street Operating System" or a "Small Business Operating System" do they? Why not?
Ian Kennedy followed up to Matt's post with The Web as a Loose Federation of Contractors in which he says:
I like Jeremy’s illustration - an OS gives you the impression of an integrated stack which leads to strategies which favor things like user lock-in to guarantee performance and consistency of experience. If you think of the web as a loose collections of services that work together on discreet projects, then you start to think of value in other ways such as making your meta-data as portable and accessible as possible so it can be accessed over and over again in many different contexts.
Bingo again.
Lock-in is what vendors like Microsoft and IBM had relied on for years and years. Today's on-line lock-in is focused mainly on things that are less concrete: your social graph or the list of television shows you like to watch, for example.
But when you stop and look at how technology has changed and opened in response to networks, the Internet, simple/open protocols, view source, and self-serve business systems, there's a clear pattern emerging. Strong forces are at work here--strong enough that you're better taking advantage of them rather than fighting them. You will lose. I don't care if you're Microsoft, Google, or even Facebook (the latest golden child).
Open beats closed. Simple beats complex. Freedom of choice beats being told what technology to use.
The idea of starting now (or recently) to build up a full blown "stack" of services for the next generation of web-based applications (the kind that run on the Internet) and deciding to market it that way seems rather insane to me. You're using strategies and tactics from the previous war. It's last year's thinking (where "last year" is about 2001).
That may work in corporate IT shops (which are known for their brain-dead approach to software projects), but on the big bad Internet, it's a very different story.
The web is open and decentralized. Everything is one click away. Remember that.
Mashups are not toys. They're a good illustration of this point... a hint of the future.
Posted by jzawodn at August 20, 2007 02:09 PM
There is no X. There will be no X.
So many times we have heard this in the past. However the reality is a bit different.
Facebook platform is a kind of operating system for web. People build apps for Facebook exactly as they build for windows.
Jeremy - your comment confirmation system is rather annoying. Is it possible to make it that after some threshold of approved comments from the same person people don't need to bother ?
I tend to agree.
How much of the clustering around terms like "Web OS", "Web 2.0", "viral marketing", etc. can we attribute to SEO?
It's not that the people writing those terms don't know what they are, but they use those terms because everyone else uses those terms and people search on those terms and you want the traffic?
BTW, our reading lists seem to overlap a lot. 3 out of the 10 stories in your "recent links" section showed as visited when I got to this post.
i don't agree that the internet is open. technologies might be but services are just as locked in, if not more, than the platforms of old. it's a lot easier for me to transfer pictures or music from a windows pc to a mac and vice versa than it is for me to transfer my flickr pictures to somewhere else. i can't think of a single thing on my computer that i can't transfer somewhere else. Blogs, social networks, etc are all more locked in than anything i have on my desktop.
I disagree.
I mean - yeah, a "Web OS" is kind of a lie in that it's not nearly the same thing as a desktop OS, especially if we look at an operating system by the narrowest definition possible - the thing that makes the hardware components work together.
But, an operating system is generally assumed to be more than just something that talks to the hardware - it's a platform. And that's generally what people are talking about in the context of the web os.
Think about what made Windows succeed. In the early days, they made it very easy to develop powerful applications for Windows. They provided the libraries, the environment, the GUI - they made it very easy for a developer to write an application without knowing how to detect mouse clicks or talk to the printer. Windows did it for them.
In the web operating system, API's take the place of DLL's and libraries. Just like a programmer would let Windows figure out when and where a mouse click occurred, today he let's Google do the searching, Yahoo offer the maps, Amazon s3 handle the storage.
Mashups are the case in point - they couldn't exist without the big guys offering the platform to build these applications on top of. And once they're done, these applications are every bit as dependent on (locked into) Google's, or Yahoo's, or Amazon's platform, and subject to their whims and fancies.
well this will devolve into "what is an OS"
an OS isn't just something that "talks to hardware"...because anyone writing assembly is talking to hardware. the kernel just has the distinction of being the hardware-aware process started earlier that people generally consent to mediating hardware issues due to convenience and other benefits (stability, etc).
an OS isn't just a "manager of executing processes", on a logical level we subsume activities into virtual platforms (jvm, flash, the browser, the database) on a regular basis.
is the OS an api? arguably it is "the" api.
a single network-aware api could be seen as serving the same intrinsic purpose to a developer - the single operating architecture that pervades all codepoints.
now as to the value of the web "OS"s...well how do you play WoW in one? until the web gobbles up everything else people do with their systems and networks, there seems to be little value in this. you can go back to the 90s and people like the former rocketmail founder were doing this with desktop.com
jay, transferring your photos out of flickr is much easier than transferring them out of most other places. Flickr has given you all the tools to do so, and if you look around a bit, I am sure you will find apps that use the public flickr api to do this.
It is as simple as:
Get a list of all your photo sets:
http://www.flickr.com/services/api/flickr.photosets.getList.html
For each set you want to transfer, get a list of all the photos:
http://www.flickr.com/services/api/flickr.photosets.getPhotos.html
That's about a 20-line PHP script to migrate your photos on a per-set basis, or to write something that grabs all of them. For completeness sake, in case some of your photos are not in sets, you can get a list of all your photos not part of a set:
http://www.flickr.com/services/api/flickr.photos.getNotInSet.html
Thanks for the link Jeremy. The fundamental problem I have with the "web as OS" argument is that it skips over a tricky problem that everyone forgets. A stable OS requires updated hardware drivers to make everything run smoothly.
Until there is a layer of services that normalize how data stores can talk with each other, the standards complaint, open network is always going to be more ubiquitous (valuable) and closed systems are only going to be useful to other components locked into and ever shrinking "stack"
The Web OS is like a normal OS (operating system) but its entire infrastructure is located on many remote servers that you can access at any time you want via a browser as long as you have an Internet connection. Many people take the concepts behind the Web OS too literally so they come up with ridicules solutions that serve no purpose at all. In that respect neither Google Custom Home Page nor Microsoft Live Custom Home Page is the beginning of the Web OS. The Web OS is not a desktop environment so don't even think about starting the next AJAX site that has useless features like a calculator, a RSS reader, a movie player, an audio player, etc.
Your boot loader is your browser. The browser is also responsible for launching any application you want to interact with. On the top of the browser there are many abstraction layers that help the Web OS to function. The more versatile your browser is the more things you can do.
Different vendors provide different applications. However, the Web OS allows for these applications to communicate with each other the same way processes do in conventional operating systems. Web developers call these inter-web-application communication mechanism “service”. Applications that use services are called Mashups.
etc, etc, etc
You have to think about the (potential) consumer benefit of a supposed 'Web OS'. If it makes things that are currently hard, easier, then the concept could be a big winner. Look at people flocking to services such as flickr, gmail, etc. What are these? At the highest level they're just web applications maintained by some centralized third party. Are there alternatives to these? Certainly, not only in terms of other centralized services, but in open source applications that you can install yourself. The problem? Most people dont want go through the entire process of buying a domain, then buying hosting which may or may not support the right backend technologies, coupled with the nightmare that can be actually installing and configuring these apps.
The million dollar idea is that a 'Web OS' could consolidate much of the hassles. Preconfigure a bunch of backend technologies (PHP4, PHP5, Ruby, Mono, MySql, etc), and make a shiny frontend that consolidates file management (Flex maybe?), and develop the equivalent of sudo apt-get or install shield (make it shiny too) for web apps, and you have something that the average pc-savvy consumer might actually shell out money for! We're slowly getting closer to this with the various C-panel like applications, but we're not quite there (yet). Anybody hiring? :p
To continue on the construction metaphor, IMHO the Web is like the Tower of Babel, except that the builders have adopted a super simple pidgin language to communicate amongst them.
Spoken HTTP is just 2 verbs, GET and POST. Webapp APIs add their own nouns and adjectives, like Rasmus' Flickr example. Technology that sticks close to HTTP (LAMP, REST) gets used. Technology that strays to far from it (.NET, SOAP), less so.
I think the 2 main factors that take an operating system to the top, while compared to its competitors, is
*speed and
*simplicity.
I found that very few online OS's had these features. Also another factor being multiple functions.
I have tried many Online OS's but i have found that Edeskonline, a realtively new contender is faster and simpler to use than Goowy. It has some brilliant features like an accounts and online tax package as well. Comparing Goowy to edeskonline will be an interesting task.
Hi Jeremy,
I agree with your opinion that there would be no Web Operating System implemented. Moreover, I consider of changing the term WebOS to WebROM (web-resource operating mechanism), which raises the agreements of web operations from the method level to the methodology level.
The Web is more than a platform. It is also a society. When a society has to agree on a set of global methods, this society is a dictating society. When a society only needs to agree on a set of global methodologies, this society is a democratic society. If a society even does not want to agree on any global methodology, this society becomes an anarchistic society. Among the three options, a dictating web is too hard to maintain; an anarchistic web is too difficult to manage. Therefore, a democratic web is the one we should and could achieve.
I have more thoughts about WebROM at http://yihongs-research.blogspot.com/2007/09/initiative-of-stage-transition-view-of.html
How do you think of it?
-- Yihong
Jeremy and Yihong,
I deeply appreciate your posting and that of the others on this thread. I'd ask you to look at the issue a little differently and tell me what it would take for an OS to be a real no kidding Web OS. I'm currently studying this issue and I'd deeply appreciate your insights and that of others that posted above. Thanks!
Kip
ROTFLMAO ~ What a load of LoL!
I love parody sites.. anyways;
Eric hit the nail on the head:
"an OS is the thing that makes
the components work together."
With WebOSs the components are
Network API, SaaS and Servers.
To disarm the hardware purist:
WindowsXP on a virtual machine
is no less an operating system
while resources are emulated.
And though it may very well be
true that the first WebOS able
to loading up WoW will win.
(zOMG that's teh funniest)
That doesn't disqualify DOS or
UNIX as true operating systems.
What we have here is a failure
to appreciate a new generation
of technology and some jokers.
IF the author's opinions were
true there wouldn't be enough
crack to go around, MUHAHAHA!
I don't agree. As Wikipedia says, a webOS is not a real OS. It's all just for taking it everywhere. For example you save your documents in a webOS. Then you go to your Wii from another point of the city and edit it again. You forgot your presentation on a USB Stick. No prob. you have it online.
First off, decent article. Anyways, Web OS's. Interesting concepts in my opinion. Why bother? Simple. With the road that the computer industry is pushing everyone, web os's "make sense". Decentralize everything so that there is no "real os" on the computer. The PC will ultimately end up being the storage facility for our music, movies, etc. Even games are becoming browser based in some cases. What does a decentralized OS have to offer? A bit more freedom, in some cases. But if you loose internet, what then? This is the conundrum we are running into. Theoretically, Web OS's are great, but for the average user, not sound.
