If someone only wants to make money, cloning is a risk-free option. Look at Jabong et al. They are shameless clones, but they are getting traction.

Java is so massive in terms of world wide community. It will go on for a long time — or up to just before the Singularity.

What is interesting to watch are the young social, mobile and gaming companies — what technologies new and old are they using. Facebook with PHP compiled to C++ for example. Infrastructure has become so inexpensive that now it’s energy costs that are the focus to reduce cost. Which languages fit that. I’m starting to see a resurgence in Silicon Valley where the backend is written in a scripting language and when speed is required — C/C++

curl http://npmjs.org/install.sh | sudo sh

http://www.youtube.com/watch?v=bzkRVzciAZg&feature=player_embedded&hd=1

Re: Node.js and threads. Don’t get me wrong, I’m firmly in the Erlang camp here. I don’t think node needs threads per se, but it does need a lightweight, pre-emptive computing unit. As far as I can tell, that doesn’t exist yet. Node.js can certainly handle a large class of problems today, but there is a huge class of problems that it doesn’t address at all.

Re: HTML apps are hard. I wasn’t saying that you can’t make rich web apps. There’s plenty of evidence that it’s possible. But it’s much harder than it should be. My point is really more that it’s easier to get a great experience in a native app than in a mobile web app. This might not be relevant in iOS, where apps are essentially isolated silos. In Android, apps mingle more freely. They augment and extend each other. I don’t see how to integrate a web app into that model.

In any case, thanks for the list. Predictions are always fun.

This post would have been too verbose (and less fun) if I reasoned and argued for every point.

I think every point can have arguments pro and against it, I will leave that to future focused blog posts. :-)

—

@Shawn, I agree with you on middleware. The web / browser itself is a middleware. I agree with your summary of Apple’s selfish tendencies.

—

@Jim O’Flaherty,

You make excellent points. Yep, languages once adopted will not disappear. And we are definitely talking many years and not few months or couple of years.

I also understand that my predictions can be way off, and I hope to be the first to admit when one proves to be wrong.

Thanks for the pseudo-congratulations! :-)

—

@Dan,

Reasoning would have made this post too verbose, less fun and deprived me of material for focused blog posts and discussions.

Node.js & threads, nope. But Node.js and processes, yes. Processes are shared-nothing which makes sense for scaling web apps in more ways than one.

Python, very relevant. Especially when your app runs into performance issues. PyPy does the trick. The web frameworks eco-system is leading the pack, some interesting names – Twisted, Tornadio.

You can already make very rich GUI apps with HTML on existing browsers. I can’t agree with your assessment that “it can’t be done” on the browser.

Looking forward to see Scala 2.10. I would love to see it improve and make a serious dent on Java user base. :-)

> Scala (and Java) will cede the web / mobile app space to the likes of Node.js, Python.

Does node.js have a good solution to the multithreading problem yet? There had been talk about adopting web workers, but I don’t know if that ever happened. That, and maturity, seem like the two biggest blockers to broader adoption.

Is Python still relevant? I figured that everybody who might build a Python web app had already moved on to Ruby. That could be reversed, but it will take a lot of effort.

> Bonus non-Scala prediction – Native mobile apps are stupid. They’ll eventually become irrelevant as browser based web / mobile apps consolidate and take over.

You don’t use your phone on the subway, do you?

Over the next decade the mobile browser will mature as a platform. But HTML was never designed for building rich GUI applications. Even though we’ve been pounding our square peg into its round hole for, what, at least a decade, it’s still painfully hard to do. Worse, I don’t see the W3 making any serious attempt to fix that problem. And as distasteful as it may seem, canvas-based applications might be the bright future.

> If Java implements functions as first class members tomorrow, half of Scala’s oomph will be lost.

I think you’re probably right, but let me throw out another prediction. When Project Kepler is released (which I believe will be with Scala 2.10, though I haven’t been watching the news), Scala’s appeal will quadruple overnight. Recently, I started to experiment with compile-time type gymnastics in Scala. It has been very interesting (very mind-bending), and I have learned a lot. But really, most uses of crazy type system techniques are a band-aid over the lack of a good macro system.

A) Once any language/library ends up having even a small amount of success, it will be around forever. Now, the utilization might drop down to very small levels, but it will remain around. To pick a non-COBOL example, just look at Forth. {smirk} So, Ruby and related libraries will continue to be here for decades. However, and to your point, they might fade to relative insignificance over the next decade (but it will be a full decade, not just a couple of years).

B) I think attempting to make effective predictions in any complex dynamic system is fraught with too many unknowns. We can hardly predict the weather (or economics, or social fads, etc.) effectively, upon what basis do you think you can effectively speculate about something which is moving and adapting as rapidly as Ruby, Scala, Groovy, Java (yes, it’s a stretch to include it here, but there is finally progress here, too), etc.? You can’t. What you can do is make a stab in the dark here with your predictions. And then, you can claim “rightness” in the future when some “variation” of some of your predictions land closely enough to what actually happens.

So, I predict that you will end up using your strong confirmation bias to find evidence to support your slightly adjusted via reworded interpretations of your predictions. So, I’ll be the first to pseudo-congratulate you on your successful prognostication! {smirk}

Nope. With browser based apps, you are reliant on browser implementations, which will _never_ reach an acceptable quality across the board. HTML is designed by spec and implemented through interpretation… it’s doomed from the start, it can never truly be consistent.

Middleware is the answer. Adobe AIR, Corona, Titanium, take your pick, middleware which compiles to multiple platforms, with consistent rendering and performance, that’s the future.

To think web apps take over, is to ignore market realities. Apple won’t let this happen, neither will MS. App “lockin” is what drives Apple’s cyclical hardware sales, they can’t afford to have their core apps be web based.

Bonus prediction for you: As soon as WebApps begin encroaching on the apple ecosystem, Apple will just stop adding required API’s for Safari Mobile. (Already you can see this as they cripple the sound api on mobile safari… I wonder why)

{I used profane language here that was cleaned up} I won’t contradict you or propose my bogus “predictions” too. Perhaps, if you mind leaving some arguments, we could have some discussion.

PS: I don’t think that the “native apps” on “Desktop / PC / Mac / Linux space” are disappearing. But I am not a business analyst…

o_O

The song has been played. This has happened in the Desktop / PC / Mac / Linux space.

It’s only a matter of time when technologies consolidate and a single technology is prevalent across platforms.

My bet is on browser based apps.