Only large and deep-pocketed users opted to lease a mainframe for their exclusive, local, use.
"Time sharing" is "pay X per month and use up to Y hours".
"Pay per usage" is "pay X per hour of usage, no minimum requirements".
In a time-share it can be very hard to break out of the contract.
The key difference between cloud and the traditional mainframe model is the decoupling between usage and the assets.
Time sharing is buying a share of an asset. Reserved instances is pre-commiting to computing utilization.
Actually, it was identical according to you description. Today, there is the decision of on-prem vs of remote computing, which was the same decision then.
> instead of, "Here, spend a million dollars and get locked into a service contract," it's, "Here, spend $0 up front and just pay for usage, then leave whenever you want."
The huge capex and million dollar service contracts still exist today, for example. There have been articles posted here on HN about companies that moved to the cloud for savings, grew their business, and eventually left their cloud provider due to high costs.
No, it wasn't the same decision. With cloud I can use as much or as little cloud resources as I'd like, and I only pay for what I use. Back in the day, remote computing was not like that.
> The huge capex and million dollar service contracts still exist today, for example.
Right, that's exactly what I was saying. Cloud offers an alternative to huge capex, which is the biggest drawback of mainframes, not counting the fact mainframes are difficult to second-source.
I've never been in a position to find out the hard way, but that's how "the cloud" always struck me. "Have access to the computing power of a multimillion dollar mainframe w/o buying the mainframe!" But if/when you actually need and can afford a multimillion dollar mainframe (or equivalent), cloud pricing might not be much savings.
Already happened, unfortunately. Google's ecosystem, is tuned to constantly nudge you towards the cloud. My pet peeve is their "discontinuing" of Picasa (offline photo manager/editor/viewer) and removing the download links from their websites, and peddling Google Photos (cloud service) as a "replacement". It also takes mental effort to prevent photos from going to the cloud when you get a new phone.
BCL is just one part that's clearly inspired OS/360 et al. (look at JCL).
I have never heard such a succinct reason why Stallman's crusade against proprietary software will not work. I will remember this.
Here's Stallman in 2010: