Stallman vs. Clouds (Linux Journal)

I don't know if it's that easy. For me, computing services are much more structured than the basic infrastructure service eclectricity where there are no structural variations. There's a nice yardstick to show that: Is there a single measurement that names the amount of service that one gets? In electricity, it's kilowatt per hour, my bill has just that number and that's all. How do you measure the amount of a cloud computing service that you got?

Now, if you talk about specific computing infrastructure services that don't have substructures either (remote storage, CPU time, Internet connectivity, maybe DNS), then I agree with you. Please note that you can measure their usage with one value. Actually, in case of Internet connectivity, we are already there; and Amazon's S3 et.al. show that one can buy such basic services on the fly. I use several of such services and they make much sense.

But the cloud computing apologets usually talk about much more: About the supply of whole application areas in the "cloud" and not just about basic unstructured infrastructure components. Not about Amazon's S3, but about Google Docs. Compared to electricity, that's more like my oven, my washing machine, my stereo, well, the lights in my rooms with their retransformation of power [sorry, probably not the proper English terms]. Comparing highly-structured computing services with unstructured electricity delivery is a nice marketing ploy in the trade press, but it compares apples to oranges. Let's compare washing machines to Google Docs instead -- yes, some people go to wash salons since they don't have an own machine or visit their parents with their dirty stuff if they're young enough, but is this really so typical?

If you consider that level, you might see that some people don't want to get their electric devices from the utility company either, just as they don't want to get their computing applications from the computing utility provider.