Letter to Adobe: Kill the Apps
I have followed Adobe since its beginning back in the early 1980s. The little company John Warnock and Chuck Geschke built from a tiny rented office near 101 in Mountain View almost 30 years ago has grown into a world leader valued at over $23 billion. Adobe’s culture embraces innovation, partnerships, support of standards, quality of life for employees, green initiatives, and corporate responsibility.
In this essay, I claim that Adobe’s product line is showing its age and isn’t working at web scale. I want to point Adobe in a new direction, based on cloud software and massive collaboration. I’ll start by describing today’s Creative Cloud, next I’ll define cloud computing, and then I’ll get to the main theme: what Adobe can do from here, and how that transformation has already started.
Creative Cloud Today
Adobe’s Creative Cloud has been a huge success. It’s a service that puts all Adobe’s products online, ready for immediate download and install to any computer you’re using; it includes web-based fonts, an online community, and a suite of new tools that help designers create and publish to different platforms. For $50 per month, you can download any app any time, and you get $20 gb storage to store your files, essentially replacing Dropbox. It’s a step up from sending out DVDs or charging for each program separately.
Is it cloud computing? No, it’s not cloud computing. It’s web storage and delivery, and a new way to charge for a confusing array of stand-alone desktop applications.
Cloud Computing, Defined
Many companies have put their traditional desktop apps on a set of servers connected to the Internet and called it cloud computing. But naming it cloud computing and adding cloudlike logos doesn’t make it cloud computing.
I believe the two defining characteristics of cloud computing are elasticity and interoperability. Most uses center around the first property, but it’s the second that really fulfills the promise of the worldwide web.
Elasticity means the replacement of dedicated hardware with shared resources, so you can add capacity or capability easily, even on the fly as you need it, and reduce when you don’t need it. For example, AutoDesk cloud rendering lets designers send their rendering jobs to the cloud for processing and get them back and share them quickly, without having to dedicate specific machines for rendering. The more people who use a service like this, the better and cheaper it gets for everyone. Done properly, the cloud gives you infinite storage and computing power, and you pay only for what you use.
Interoperability means that data and applications work seamlessly across systems, making mash-ups and creating new apps easy. If you do cloud computing right, you never duplicate any piece of information — you leave it in one place and then refer to it by name. An example is your LinkedIn profile. Even though it’s not really data, there’s only one, and you don’t “send” it to anyone, you link to it. Ideally, when you “log in with LinkedIn,” you would see your LinkedIn profile on the new site, and if you edit your profile on LinkedIn, you’d see the changes on all systems that refer to it. Sadly, that is not the case. Instead, other systems just copy the text and keep the copy, making it harder to keep track of everything. While several branded ecosystems (Google, Facebook, SalesForce, Linkedin, etc.) are interoperable within their system, very few interact this way with other systems. Two good examples are YouTube and Google Maps, which let everyone share and embed the same resources across the entire web, without duplication.
A New Adobe Platform
The Creative Cloud isn’t cloud computing, but I’m sure there are a handful of people at Adobe thinking about how to go from here to there. My goal in this section is to give people in all companies, and Adobe in particular, a vision that may help steer toward building a 21st century work model.
We all know that our familiar desktop apps will eventually go the way of the dinosaurs. But even though I’m typing these words into a browser-based app and haven’t used Microsoft Word in years, it would be asking a lot to do in a browser the kinds of things today’s high-performance desktop apps do. Besides, there are times when you want to work offline, times you want to work on a touch-enabled device, and you may even want to work on a netbook someday. Fortunately, there’s a way we can have both the performance we want and the cloud-computing advantages that true collaboration requires.
Imagine that you download a single new program from adobe, let’s call it Adobe CoLab, for collaborate. It’s not a browser, it’s not an app; it’s a platform. On the CoLab platform EVERYTHING is a service. You start by indicating whether you want to work in 2D or 3D (you can change later), and you establish a work space for your project. Now, rather than going out and looking for tools, the tools come to you.
In a 2D workspace, everything is in layers. When you’re sketching and prototyping, you have access to a wide range of services: sketch libraries, sketching tools, pre-made art, photography, environments, lighting, models, surfaces, textures, video, sound, music, fonts, animation, spell check, typographic tools, etc. On this platform, hundreds of people can collaborate on a single workspace, similar to the way we share Google Docs today. As the design matures, you bring in new people, new services, and new tools: rendering, shading, review, comments, A/B trials, analysis, output, color spaces, programming platforms, etc. As you dial in the details, you work with dozens of different tools, each one providing a service to do its particular job. People can easily work in parallel on their part of the project and everything plugs in to fit together. When the project is ready for production, the final files go off to a printer, moldmaker, web site, or whatever the final product may be.
You can have a 3D model in a 2D environment, of course. A poster showing airplanes or cars would be a good example. You can also turn the environment into a true 3D workspace, allowing people to collaborate on a model they way they do using Dassault Systemes today. In this scheme, hundreds of collaborators have access to the workspace, which manages an infinite number of 3D “layers,” allowing people to plug in parts, check for fit, and do “what if” scenarios, motion testing, stress analysis, and much more — with the accompanying documentation, video conferencing, and iteration management built into the space. In fact, this is the only way airplanes are designed now, because making thousands of flat drawings, printing them out, and mailing them — or emailing attachments — to partners wasn’t scaling very well.
In the CoLab system, the entire web serves as a storage device. All the data for a project stays in its online space without being sent, imported, or exported anywhere. For example, shoot a video or photo and the raw file goes online; now anyone who has permission can work with that file to create derivatives — no copying, transfer, or “lost data” issues, ever. Everything gets reused, and much of what you modify you can share with others, creating an “open source” creativity platform. Want to add some digital minions to a scene? Using the CoLab search feature, find the minions you want somewhere on the web, and purchase the rights to use them instantly. You can even describe your environment, so for example you could “turn off” gravity and the various elements would respond.
This past July, Adobe announced Adobe Anywhere, a new service aimed at managing enterprise video assets. The service lets you upload video once and share across many of Adobe’s apps, maintaining the integrity of the original footage and even doing the heavy rendering on the cloud. Think of it as DropBox for video, done right. Both these attributes — elasticity and single-source data — make Adobe Anywhere a true cloud platform. It’s only available for large companies, and it’s just a small part of the eventual work-flow platform I hope to see, but it’s a great start.
Accelerate the Revolution
The CoLab approach requires new namespaces, new kinds of searchability, new standards, new architectures, new programming languages, interaction models, and an active market of participants — all built around the always-on web. It will require a combination of proprietary technology and open standards like HTML 5. Anywhere is a small but positive step in the right direction, and there is much more to go. I’ve written all about this future in my book, Pull, and on my web site, ThePowerOfPull.com. There, you’ll find an 8-minute movie describing the personal data locker, which takes this concept several steps further.
Just as Microsoft Office has become far less relevant, so will the Adobe apps some day. In the future, to get working on your project, you won’t fire up an app and bring the data into the app. You’ll fire up the project and bring the apps/services to it as needed. Starting now, the future of the company rests on Anywhere and other projects that I hope will be based on the true principles of cloud computing. If Adobe doesn’t build this, two entrepreneurs in a tiny rented office space will.
Adobe and other large companies can learn how to take advantage of agile approaches at businessagilityworkshop.com.