August 07 2006 04:28:36 PM
Software Complications
[+][-]
In a rare fit of brevity, I'm going straight to the point.
Software is too complicated.
It seems like no one writes simple programs anymore. Installers, config files, menus, options, blahblahblah, ad nauseum. Whatever happened to the UNIX concept of "do one thing, and do it well"? I'll tell ya what happened. 20 years of Microsoft.
It became obvious to me while installing a popular PHP based photo gallery for a client a few nights ago that it was far more featureful and complex than what was really required to meet the need; which was simply a place to put photos for site members.
I blame commercial software. It's packages like Microsoft Office with their incessant feature creep that have driven the industry to create 'one more must have feature' releases. But it's not just commercial software. Free software is just as bad. OpenOffice.org is just as much of a behemoth as MS Office is. The problem is that in order to compete, even as free software, OpenOffice.org must provide an equitable feature set. It's an arms race. And you as the software user are on the losing end of the war.
August 03 2006 12:12:42 PM
The Interactive Net
[+][-]
As it would happen, I was reading an interesting article today about the new flash player for linux and I had a really interesting idea.
Before I go into the idea too much, I'd like to summarize some of the discussion that spawned it. Basically, the article was written by an Adobe employee who decided to post the technologies that the new flash player for linux is going to use and ask for feedback. He probably got a lot more than he really cared for. Some of the ideas that were put forth were fairly interesting, however, as were his arguments against them.
The problem, in a nutshell, is that truly interactive content is limited a great deal by the browser that it must run in. For example, OpenGL accelleration would take a great deal of processing power off of the CPU and provide a much smoother and immersive end user experience. It'll never happen in a browser though since the 2D APIs that the browser runs on would not mesh well with the 3D rendering provided by OpenGL. [Actually 2D OpenGL accelleration also does not work well with the 2D Apis of say X11]
This got me thinking. What I came up with is why not have an application specifically designed to run in a 3D container such as OpenGL, and provide a framework designed from the ground up to support interactive content, such as flash, across the internet? It makes sense to have an application designed to deliver fully interactive, dynamic, vector based content rather than trying to kludge it into a browser. While we're at it, why not write a network protocol more suited to it? It's amazing what has been done with GET and POST, but HTTP was never really designed to handle advanced user interaction very well.
A fork of the web to handle interactive content might not be as crazy an idea as it sounds. I can see a few very good arguments for it.
