For example, Microsoft repeated the current industry meme that its own particular cloud will be for everyone from two-person start ups to enterprises.
To do this, though, Microsoft will need raw capacity backed by some rigorous SLAs and reliable management systems to ensure everyone gets exactly what they want when they want it.
Early attempts to push Azure at Microsoft's two shows this week proved there was a long way to go. Attempts to install applications on Azure hit delays, as you'd expect with early software: It wasn't clear whether these were down to network latency of slowness of Azure.
You won't be able to test applications online, instead you'll do it on the machine were you've got Visual Studio and an instance of Azure running. Also, it seems, you won't get any information on the performance of Microsoft hardware fabric.
On storage, Azure is a structured database but it's not relational. There's no enforced referential integrity - you have to write your own - and you don't get views or triggers. Links between tables will be different to what you're familiar with in SQL Server.
One VSLive delegate conceptualized the Azure storage system by comparing it IBM's old Indexed Sequential Access Method (ISAM), where an application would use an API to search indexes and find data files - different to relational data. Hence: mainframe in the sky.
And yet, just as things looked like they were becoming tangible, it seems they are set to change. Microsoft told The Reg Monday Azure's going relational, as it's putting as many features as possible from its SQL Server database in to Azure as quickly as possible in response to partners' feedback.
Azure is a cloud indeed for Microsoft: It's difficult to contain and harness, and its shape is constantly shifting.
In a sign of the work ahead in conceptualizing the cloud, nailing down the features and then delivering, it was telling that while Microsoft evangelized Azure in San Francisco, up in Redmond Microsoft it announced the formation of the Cloud Computing Futures group.
This is a research and development unit that will partner with others to "explore hardware innovations while simultaneously building the software stack to exploit new hardware designs," Microsoft said. The initial target seems to be to find ways of reducing power consumption in the data center.
Microsoft is in the same boat as software, hardware and services companies when it comes to the cloud: This is a new model, with a relative meaning, subject to crushing levels of marketing spin, and that owes more than a passing resemblance to models of computing past.
Expect things to change on Azure, as they should with any new software or service. Just let Microsoft be the one that takes the risks when it comes to making these changes.
Microsoft will, as it did this week, try to persuade you with tales of early adopters Energizer, Ingersoll Rand, or - that evergreen of Microsoft customer references - Coke. But beware. Jump too soon, before the details are finished and the services packaged, and you risk getting left high and dry when things change.
If you want an example of how uncertain things are inside Microsoft's online services world right now, look no further than its online Office bundle Equipt: launched with a flourish last July - and finally killed last week. ®
You forgot point E : any version after #3 is on an irreversible trend of bloat and cruft that will render the product more and more unusable and full of complex vulnerabilities that will be utter bliss for all the skript kiddies and malware authors.
Can't argue with a and b.
I can argue with c though. They only listen to their customers when it suits them, or when to do otherwise will significantly affect their earnings. The problem is that they are too big and are controlled too much by one sector of the business which often leads them into strife. While they have come up with good things over the years, the whole business is geared towards the domination of the market at whatever cost.
As for d, the usual way of putting this, as has been said over so many years, is "Never touch a Microsoft product until the first service pack comes out".
The Truth Will Out
All of us who have used Microsoft development tools for many years know that:
a) There is always a lot of hype
b) Version 1 usually leaves a lot to be desired
c) The listen to their customers and adapt well
d) Version 3 is usually great and significantly better than what competitors offer and it is usually fairly (as defined by the market) priced
People will always bash whomever is on top. Now it is MS. Someday it might be someone else. Someone telling me constantly that the sun rises in the west does not change what I know to be true.