The cloud is not a computer
My hat goes off to Preston Monroe, the developer of iCopy, an online service that adds cut and paste functionality to the iPhone's browser and e-mail apps. As you probably know, Apple's handheld computer bizarrely omits this feature.
iCopy is a clever hack that lets you select text or a link from a Web page and paste it into another page, or an e-mail. It gets around the lack of iPhone-native copy and paste by sending selected text to a temporary online repository when you "copy," and retrieving it when you "paste." In operation, it's a horrible kludge--it requires a lot of Web page switching and too many visits to the iCopy site to do a simple copy/paste operation. But the fact that Monroe figured out a way to make the Web a giant clipboard in the sky is pretty cool.
iCopy illustrates that while the Web can be employed to do a lot of things that we've formerly thought of as belonging solely in the domain of local computing, it doesn't mean we should do so.
I edit a blog about Web 2.0 apps. It's my job to push the vision of Web-based products and cloud-based resources. But even I realize that local processing has a place. I find it curious that many people I talk to think Microsoft's rumored Midori project, for instance, is a "cloud OS." While there's no question that an operating system written from the ground up today should use Internet resources in a more native fashion than most OSes do today, the change should be seen as one of degree, not replacement.
The Internet can be used to deliver apps and updates, for storage and backup, for social networking and person-to-person communications, and other functions. But for the moment and the near future, you need local processing to maintain speed and robustness of applications, and native graphics capability to present the interface. One of the reasons Web 2.0 apps can work well today is because today's browsers have deep user interface and graphics capabilities, and because they run on powerful local PCs. Many popular Web apps--like Google Docs and Microsoft Live Search Maps--rely on capabilities that were simply not present in PCs only a few years ago.
That's why I continue to refer to Web operating systems like G.ho.st as science fair projects. They're really cool, and they provide glimpses of the evolution of personal computing. Much of what we do on a PC today can be done over the Web. But a lot cannot, at least not well. To deliver the best experience--the best user interface, reliability, collaboration, and so on--smart developers don't force all their apps either onto the Web or the local PC. Today's architectures make distributing applications among platforms easier than ever. They even make it possible for apps to adapt to their environment and redistribute themselves depending on circumstance (see Google Gears). The really interesting upcoming apps and operating systems will not just be hybrid (online/offline), but adaptive.
Meanwhile, if you're interested in how copy and paste might work on the iPhone, check out Proximi's Magicpad, a text editing app that offers cut and paste controls. Proximi has also published video proposing a user interface for general cut and paste on the iPhone. This is the work Apple should have done. Although for all we know, the company has done it already, but in secret.
Rafe Needleman writes about start-ups, new technologies, and Web 2.0 products, as editor of CNET's Webware. E-mail Rafe. 



I have a comment about Web OS in general and G.ho.st in particular. IMHO a new concept such as web OS takes time to grow roots. Take a look at Windows 1.0 ,back then it wasn't clear it's going to take off either.
Since the 70s the OS concept hasn't changed a single bit - we are still walled into a single physical computer that we have to back up, upgrade and secure in order to keep our precious software, settings and data safe.
Recently things began to change and today we have many Web 2.0 apps that become more and more useful. Still you have to remember many different URLs, logins, passwords and, most importantly, these apps do not collaborate at all.
Take the example of Google - when you compose an email on Gmail you cannot browse, choose and attach a document from Google Docs, and these apps some from the same provider!
According to Go2Web20.net there are thousands of Web2.0 services. Do they have an open API? (only a few). Can they collaborate? (most of them cannot). Do they support open community standards? (each one its own)...
That's where G.ho.st comes in handy - it's designed to provide a common layer of open APIs based on open community standards such as WebDAV, OpenSAM, OAuth, OpenID etc. that will allow many distributed an heterogeneous apps to collaborate simply and efficiently.
I would give it a chance :-)
For the forseable future (i.e. 10 - 15 years out), local resources and OS will be critical to delivery of anything we would consider progress in software. Blending this with cloud resources will deliver a very rich and transportable experience.
As to Midori being purely a "cloud OS" maybe the exercise is or is not, but Ray Ozzie is at the center and controls the direction of MS (including Mesh, Midori, etc.). If you look at all his comments since taking over as Chief architect for MS in 2004 through today, it is clear that he strongly believes in a blended (local and cloud) approach to delivering software as service. Any commercialization of Midori will surely be practically grounded to allow for the exploitation of powerful local resources (mobile, or otherwise) deeply integrated with cloud resources.
If you listen, MS (Ray specifically) has been preaching this for the past 4 years.
I still hate to see it when apps -- like the Office suite -- don't take advantage of the Web for cheap (or free) collaboration, though. Google Apps' real-time group editing simply blows MS out of the water both in capability and ease of use (compared to Sharepoint). And it's totally free.
Boy, I wish every developer would take just a second to think about that statement before deciding their application is a should be web-based.
Porous solutions (e.g. AJAX) to performance-based issues with complex user interfaces are better solved on the desktop (although deployment of the app could be via the web). Even the cross-platform quandary is easily solved using one of the many downloadable runtime interpreters (flash, silverlight, even java and now JavaFX). I can hear the screams now. Let us not forget where the precious RIA's run. They are nothing more than a glitzy basic program with superior UI and network resource support.
- by Lerianis August 7, 2008 1:14 AM PDT
- The reason that they didn't put a copy and paste on the iPhone is simple: they figured that most people would be writing e-mails and text messages and therefore would not NEED a copy and paste function on their iPhone.
- Like this Reply to this comment
-
(6 Comments)This isn't a case of 'simple oversight'. It's a case where Apple reasonably said "Who in the world would need a copy and paste function on a cellular phone or even webphone?" The answer: no one that I know of, bubu.