Stax Brings more Standard App Models to the Cloud, Marginalizes Custom Platform Further

Stax, which recently launched in private beta, is a cloud hosting/deployment/ops platform based on Java appservers. The coolest thing about Stax is that it offers many flavors of JavaEE-deployable applications, including Python apps (via Jython) and Rails (via JRuby) with ready-to-roll built-in templates.

Stax has a very AppEngine-y feel, not just on the website, but in terms of the SDK interactions, local development, etc.

This is good news for all of the popular platforms ... and bad news for those rattling around the corners with non-standard APIs. As the app-hosting industry continues to mature, the emphasis will clearly be on established players like Rails, ASP.Net, JavaEE, Pylons, et al. at the expense of guys like AppJet.

It's not about the language (JavaScript) but the about learning a set of APIs, patterns/practices, and sustaining a community ... based on a roll-your-own platform.

It is true that some of these built-for-the-cloud platforms were designed from the start to default to hash-style or big-table style storage -- popular for content-oriented cloud apps because of its easy horizontal scaling -- where the "traditional" platforms focus on relational stores and have a variety of bolt-on APIs for cloud DBs.

But now that there are so many standard alternatives, it is unlikely developers will pay any custom-platform-tax no matter how elegant that platform might be.

Rails and .Net Had One Thing in Common Before ASP.Net-MVC: Magic

It's possible now to build a very Rails-like website via ASP.Net MVC. And at RailsConf, John Lam showed IronRuby starting to run Rails.

But long before all this, these two very different platforms had one thing in common: "magic."

By magic, I mean the sort of implicit, we-solve-problems-you-didn't-even-know-you-had kind of magic that makes it easy for people who don't know the details, don't want to know the details, or know but don't want to be bothered by the details to build an app.

From ASP.Net's first iteration, where you could drop an item on a "web form," wire a server-side event handler, and talk to items on the page through a .Net server-side object interface, ASP.Net has been heavy on the magic. Rails has been the same way, from scaffolding up through ActiveRecord, you point Rails in the direction you want to go and it drives.

Of course we all know about the leaky abstractions... in .Net, you used to get some funky HTML and Javascript, plus a server-based page lifecycle that didn't always match the model of what you wanted to do with the DOM. In Rails, we've seen perf issues, and discussions of the "worst practices" [pdf] that can result from letting the autopilot do too much and your brain too little when writing for Rails.

But in any case -- as opposed to a framework like Django (which included a big feature push called 'removing the magic' to make functionality more explicit) -- Rails and .Net give you as much magic as they can.

Alternate Reality: Java and .Net Both Went Open Source 8 Years Ago

The later-phase, heading-to-market end of application model innovation is screaming along.

Sun and Microsoft have announced their intent to get into cloud models that improve scalability and manageability. We don't know for sure how pleasant the programming will be, but it seems reasonable to assume it's at least as good as ASP.net 3.5 or J5EE. Microsoft seems to be aiming center-of-the-road with stuff like ASP.net, the MVC flavor,  and "dynamic data" pieces, while Sun is putting a big chunk of its chips on the Java side, and another big piece on JRuby/whatever-is-next-that-they-can-get-to-run-in-Glassfish.

Amazon is offering persistent disks for EC2, acknowledging that the RDBMS-backed app is not immediately going to disappear in favor of a S3 or SimpleDB layer. 'Cause syncing RDB stores usefully on S3 is a hassle (see slides 10-14 here).

Or, you can also buy all the missing pieces of the Amazon stack from RightScale.

Then there's Passenger (mod_rails) for Apache, and a Glassfish gem, which both aim to improve Rails deployability and incrementally bump performance.

But all this got me thinking about an alternate universe, one of those hindsight 20/20 things, that's interesting to imagine:

For Sun and Microsoft, I believe the missed-opportunity-of-the-decade was not open sourcing the JVM and CLR respectively (and their core libraries and infrastructure), from around 2000.

In the late 90s, Java was a beautiful thing, and aside from a few anti-OO diehards, it was clear that if the perf issues and minor hassles could be overcome, virtualized apps in an elegant environment (Smalltalk, oh wait, did I say that out loud?) were the future.

By 2000, Java legitimized VMs for server apps. As for GUI apps, graphics, printing, language neutral JIT, all manner of easy native interop, module loading, and the other parts Sun didn't get to ... Microsoft had those problems licked even if CLR was still in beta.

If all this IP had gone open at that time, a few major projects -- maybe 2 or 3 -- would probably have come out of it, and these projects would be the de facto runtimes for everything we want to run today. These few projects would spawn native bindings for all manner of OSes according to the existing APIs, so it would be even easier to really write apps that run on all kinds of devices, or in a cloud facility, or in on local elastic virtual infrastructure.

With the languages separate from the VMs, but with well known interfaces, a lot of debates (JavaScript 2 for example) would be less necessary. Meanwhile, Apple could have contributed whatever changes were necessary to get keep Cocoa APIs in sync, instead of going halfway down the Java road and then getting stuck, which would vastly improve the reach of that platform (their mission being more about user experience than about Objective C).

Performance, security, deployability, management -- these problems wouldn't magically go away, but with a more modest universe of options, they become infinitely more tractable.

In hindsight it seems that keeping these technologies proprietary (Sun's moving along; you can see Microsoft's code now, but you can't change it and re-release your own variant) was about fear and ego and control, defensive maneuvering.

I find it hard to see how either of these players would have made a dime less money in the scenario here; actually they would probably have made more money and have even more mindshare than they do now.

Application Models, Part 2

A little while back I described the idea of an "application model."

The application model is a key concept that keeps getting muddled and confused with partially overlapping concepts like languages (e.g., Paul Graham's discussions of "languages for smart people" vs. "languages for the masses") and platforms (a legacy platform vs. a "great new" platform, etc.)

So besides being a useful term for which I'm not sure there is an equivalent, what is the role of application models?

First, application models -- not languages or platforms or vendors -- argue, persuade, and sometimes win.

Whether it's Visual Basic (the original), or C++ and OpenGL, or Rails, the question is "How well does the application model solve the world's (ok, domain's) problems for the folks who actually do the solving?"

And the cool thing is that application models are very malleable in a way that languages and platforms aren't -- because the they include (well, really their complements or negations include) all kinds of APIs and libs that people cook up.

Huh? Like this: the application model for JSP/Servlet apps (including "model 2") stunk. So folks created all sorts of libs (Struts, Tapestry, etc.) and templating engines (FreeMarker, Velocity, etc.) in order to create a better-shaped "hole" -- a better application model for what were still fundamentally Java Servlet apps. Still a hassle. Nice underpinnings, great performance, needs some work. We got IoC containers and the application model became friendlier. For small, quick apps, Rails had a lot of appeal, so we got Groovy and Grails as options. In the .Net world, we have C# adopting elements of Ruby (yes, they're really Smalltalk or Lisp elements) and ASP.Net offering the "MVC Framework" which -- with LINQ -- creates a flavor of Rails.

There is no intellectual property violation going on here. Platforms and languages are adapting around good application models. Some of this adaptation is vendor driven, some is community driven, some involves new languages, some involves new tools, and some involves APIs. But the common thread is a desire to alter an application model in a particular direction.

Application models are what get mindshare and what make provide for the continuity of commercial application development patterns over time. More on this point later.

Before I get there, it's worth pointing out that the application model also bears heavily on the aspects of application development apart from core feature development. When we get into these aspects, non-functional requirements, or orthogonal (really non-collinear) concerns, the application model determines (or at least posits an optimal or easiest path for) integration. How the UI attaches, how app-level security is managed, how different services are integrated -- these all fall under the sway of the app model, and if we want to make these things better, the app model is what we should be talking about.

Over time, the adaptation of application models (by the all of the parties mentioned above) is the story of the evolution of commercial software. Commercial software evolution is often described as a sequence of machines, or OSes, or platforms, or languages. It most definitely is not these things, any more than it is a story of vendors fighting and dying and rising. The thread of continuity, and the real underlying story, is about people adapting application models, piece by piece, over the years.

Application Model, Part 1

As programmers, we participate in an evolving discourse around the tools, languages, platforms, and APIs that we can use. Sometimes this discourse is purely academic or reflects personal preferences; often, though, the goals are broad: higher quality software, developed more efficiently, with more reliable outcomes than what we do today.

But one critical and useful piece of terminology seems to be missing from the conversation. I propose the term 'Application Model' to describe the essential 'shape' of how a basic program would be arranged, given a particular platform, libraries, tools, and possibly even architecture.

One way to think about this is to imagine a full stack, with a non-trivial Hello World program. The non-trivial Hello World program is a basic program that exercises the core purpose of the stack in the way that the stack intends. So, for example, a simple blogging or e-commerce app might serve as the non-trivial Hello World for a web application programming stack. For a big, high-level graphics library (like WPF), maybe rendering Hello World in a specific font, extruding, rotating to an arbitrary angle, and texture mapping with a streaming video source is the non-trivial Hello World.

Now, remove the platform, libraries, etc., and, assuming it's constructed the way the stack intends, look at the shape of what's left -- of the actual Hello World app. That arrangement or shape of the code is the Application Model.

Another way to think about it in general is to consider the set-wise complements of the API/platform/language stack, in a collection of 'typical applications' running on that stack.

The complement will be a little different for each app, but across the collection, a certain shape should emerge at the boundary.

This boundary, the edge between 'app developer' code and 'stuff that app developer code uses' is different from the formal interface mechanisms (linking/loading; object interfaces) because it is shaped by the practices, patterns, and intents of the platform.

Let's describe a few examples in case it's all too abstract so far:

What is the 'application model' for a C app meant to run in a console? A main function where the execution starts, #includes to access additional platform/API/library capabilities, and a routine to set up signal handling. Compile/link/run. Those characteristics make up the application model. From there on out, it's features and "business logic."

How about a COM Win32 raw (non-MFC) C++ app with a little GUI?  Well, that's a lot more complicated. You need the basic WinMain stuff, the GUI and message loop stuff if you want a real window not just a ShowMessage, then you need the OLE/CoInitialize stuff, and the basic COM dance (CoCreateInstance, QueryInterface) to get the COM object you want. Then you can do myObj -> HelloWorld. Compile/link/run. With the bonus that you don't need to link DLLs (including COM objects).

Win32 via .Net and C#? You're back to void just a main function and a couple of using directives. Compile, don't link, run.

Raw J2EE Hello World? not a ton of code, but you need a lookup service for your other services, you need to do lookups, create bindings, and ask for HelloWorld from whatever service(s) you are demo'ing. Reference other code using import and classpath references. Compile, don't link, but put together descriptors, and subpackages, wrap in an EAR file, deploy, possibly configure app server, run.

J2EE using an IoC container and libraries? access other Java code the same way, but a lot of the other steps drop out. Maybe the packaging work is even automated (and requires less inside anyway), depending on which framework and/or appserver you're using.

The idea is that the 'application model' represents (1) the typical development experience on the platform as well as (2) the core knowledge that a developer will need to have or acquire to be productive.

Different combinations of platform, language, and supporting frameworks create different application models -- so it's a matter of comparing entire stacks and looking at the shape of their photographic negative -- that's the space that will be inhabited by developer work.