Remember Sun’s vision for Java:  client-side applets, downloaded on the fly, running in/alongside the browser?  It was a great vision, but Microsoft moved to block browser adoption and Sun botched it.  After nearly 15 years, client-side Java never really hit critical mass.

Instead, the original vision has quietly been realized by Adobe Flash.

What happened?  Video.

Site operators (notably, Youtube in 2005) discovered Flash was the best way to get in-browser video.  The growth of Internet video cemented Flash’s position as a ubiquitous browser-plug in.  Then, Adobe started adding Java-like features, notably ActionScript 3, a fairly efficient virtual machine, and a decent set of libraries.

Most of my developer friends reject Flash as “an animation system, with a crappy scripting language“.   They were probably right 3-4 years ago, but not any more.  With the latest Flash players, Adobe’s moved away from the animation-centric design to much more a generalized, programmatic system (that happens to have great animation support).

Just like Java, except with >98% market adoption.

Over creamed chipped beef this morning, I subjected my friend Antonio to my latest rants on challenges and opportunities in the software business.  He helped me crisp up a major theme:  being smart is no longer a barrier to entry.

Software used to be really really hard:  there weren’t a lot of developers (and few that were superstars), languages were primitive, tools were primitive (and expensive), the stack was expensive and buggy, and servers cost real money.  Dev cycles of 12-18 months weren’t crazy, because it sometimes took that long to build something meaningful.  Sometimes merely having shipping software was enough of a barrier.

Today, things are much easier:  dynamic languages are hugely productive, the tools+stack are free, free code libraries are everywhere, and fast servers can be rented by the hour.  Smart developers abound (they may not want to work at your company, but they’re out there competing with you).  Useful functionality can be built in days, weeks or a few months.

So what does this mean?  It means that software efforts doing things that are easy and plentiful generally aren’t going to be worth that much — it’s supply and demand.

For the effort to be truly valuable, it needs some barrier element beyond “we have really smart people with a clever idea”.   There are LOTS of smart people out there that can re-implement your idea, or cherry-pick the essential parts.  If you only spent 2-3 months building it, you don’t have much of a head start.

The key is to attach the software to some non-replicable element.

Last year I wrote about the brilliance of platform APIs for the big guys:  let your app developers figure out what the market wants, and then incorporate those into your system.  Most API terms prohibit you from suing the platform provider, effectively giving them a non-exclusive license to everything you do.

Now, we have examples of Facebook absconding app features.  From VentureBeat:

Facebook has a history of mimicking the functionality of hot services around the web. Your status updates weren’t always there, you used to have to go to Twitter to do that. Then Facebook rolled out that functionality in a way that looked eerily similar to Twitter in some regards. 

I was cleaning last weekend, and came across my college 12Mhz PC/AT. It has a 40Mb Seagate hard drive ($425, new!) and monochrome Hercules graphics. We used that computer into the late 90s; eventually, it’s sole purpose was to run DOS Quicken. We still miss that version — it was fast, minimalist, focused, and did the job very well. If they had added Internet statement downloads, we’d probably still be using it.

Now, Quicken drives me nuts — we’ve “upgraded” a few times, only because Intuit has stopped supporting our old versions. Rarely have we gotten any features we actually want; usually all I get is Kellie’s (justified) complaining about learning a new UI.

This is the core problem with the old software model: the publisher is incented to keep selling you new versions, even if you don’t really need them. I’m still using Office 2000, and in my view, it’s “feature complete”. Office 2008 doesn’t have anything I need or want.

Adobe’s PhotoShop Elements is the worst offender: I bought 3.0, 4.0 and 5.0 an then gave up. In successive versions, there were few new features, lots of gratuitous UI redesign, and in some cases, features taken out!

Subscription models are the future, clearly.

I do have a few Windows machines left around the house doing utility duty, and it seems each update of Firefox runs slower and slower.  Maybe Web browsers are like the government: inevitable bloat. 

I’m trying the beta of Safari (Apple’s Web browser) on Windows.  It’s got some quirks, but it’s quite zippy (esp. on a fiber Internet connection) — very promising.

Way back when, I thought I was a PowerPoint jock. Until I worked on a presentation with Bob Weinberger, and realized how little I really knew.  I wanted to reach back and redo years of presentations.

Like many users, I avoid reading manuals (and fewer apps actually have manuals). I stumble around and figure out enough functionality to get by, and get stuck in that starter “box”, rarely learning new features until some need pushes me.

Until now; with on-line video becoming more widespread, tutorial screencasts are a great way to learn new apps.  I forced myself to go through the tutorials for OmniOutliner and Lightroom, and I’m glad I did — I’m getting much more out of each app than I would have just stumbling around.

For app developers:  if you don’t have tutorial screencasts, you’re missing a huge opportunity to educate your users!  Please keep them short (3-5 minutes each), single topic, and update them when you revise the software.  You don’t even need a camera, just a microphone and screen capture software (I use SnapzProX for the Mac).

Like many users, 99% of my app usage has settled into Firefox+Word+Excel+PowerPoint.  Most of the new “apps” that I use are Web-based.

So, using a new app like OmniOutliner is a rare event, and adopting an entirely new category of app (a dedicated outliner) is an extremely rare event.

At the core, I’m a hierarchical thinker and I love outlining.  I usually start a new project by writing down every bit I can think of as quickly as I can, then going back and organizing & refactoring into something coherent.

Word’s outline mode is great, but OmniOutliner has a few features that make it worth having a dedicated tool.  One of the biggest (for me) is columns:  you can add columns of various types, and have the app automatically total/summarize them for you.  For example, if you are outlining project tasks, you can have a column for “effort” (e.g. “6hr”, “2d”) that OmniOutliner will automatically total for you up the outline hierarchy.  Here’s the video tutorial page with many examples.

Thanks to Pito for the suggestion.

(Sorry, Mac only).

I was having a jam session today with Steve Kane and Keith Bostic (just before the storm), and we got to talking about platforms.

I think most platforms have evolved from a successful app, where someone woke up and said, “wow, let’s generalize this and build some APIs.”  Can you think of a successful platform that didn’t come about this way?

Platform plays are seductive:  “hey, instead of fighting the war, let’s just sell the bullets!”.   The problem is platforms take all of the risks of a startup, and double them.  You have to not only build a successful platform, but you have to convince folks to build (or build yourself) successful apps.  You may enjoy some artificial success at first, but unless you have app builders enjoying sustainable success, your platform won’t enjoy sustainable success.

When I hear someone pitching a platform, I usually try to convince them to build a successful app first.  A platform play is an opportunity you earn the right to have.