Showing posts with label software. Show all posts
Showing posts with label software. Show all posts

Sunday, March 02, 2008

Software: product or service?

I've always maintained that software is a service, and not a product. After all, the only raw material is man/brain power. The reason it has traditionally been sold as a product is because producers could get away with it: dollops of functionality released periodically as new products are a more secure revenue stream than the perpetual service update. The article below blends Software as a Service (SaaS) into the traditional software stack - quite nicely, I think.

SaaS and the global virtual stack by ZDNet's Phil Wainewright -- SaaS isn't just a deployment option. It's part of a bigger picture that will obsolete conventionally licensed packaged business applications and change the entire framework of how businesses consume computing. Here's what that bigger picture looks like.

Wednesday, March 21, 2007

Skills mapping

The next mainstream OS will not be desktop-based, it will be hosted by central servers. Consumers' security concerns can be alleviated by tight integration with keyfobs, and possibly a light, encrypted local instance of the OS running on linux (maybe even on the keyfob).

The enterprise OS is a different story: the winner there will not necessarily be the most technologically advanced, but just the one who wins the skillset wars.

Where are the tech dev skills? Somebody give me a Google Earth overlay of tech dev skills: geography and skill type. That will lead to the next paradigm.

Thursday, February 15, 2007

The trouble with software...

.. is that it's soft. Malleable, pliable, moveable... wobbly. And, what's more, invisible. This makes it much harder to engineer than, say, a chair or a car.

With physical objects, you get predictable results: the chair stands, the car starts. Even when things go wrong, the symptoms are relatively straightforward, and the causes clear: the car won't start because the starter motor is broken or the battery is flat, probably caused by wear & tear or unusual weather.

This article (BCS) argues that software should be designed based on this dependability: risk-based software design. It's an admirable principle, but where to start...?