Is it me,or Safari,or are iPads really a nightmare to use on forums?

[hendorog]

Unfortunately not. They're tied into periodic activation via the web. When you sign into O365, you're signing into your subscription account as well.

Welcome to the world where software isn't a product any more, but a service. Unfortunately, Microsoft jumped on that band wagon when Satya Nadella took over. Maybe it is prudent to explore that path, as the other giants like Facebook and Google are highly successful and the traditional market has been slumping, but in true Microsoft fashion they go all-in and leave people dependent on their software with little choice. People build their companies and infrastructure around these products and if the rug suddenly gets pulled from under them, it's going to hurt in major ways whatever you choose to do next.

That is a shame too, as their current range of products really is great in corporate environments. It's just that how it's structured and licensed can make it hard to swallow.

I figured as much. Having worked for a software company, I really, really understand why developers want everyone to always use the latest version, and the motivation to change from yearly spurts of upgrade income to an ongoing revenue stream. The objection I have, which I suspect is precisely the kind of thing you run into, is the oft unreliable activation systems themselves.

From the client's perspective, it's not an obvious choice at all. You make your business fully dependent on the continued services of another. If that were to be the case with one of the packages essential to your operation, that might be doable. The problem is that everyone is jumping in and that you now need to manage a whole herd of services and associated risks. Especially offline applications can steadily run for years without trouble if it is a product and not a service. I have trouble understanding why any firm would want all their CAD projects, and the infrastructure to access them, hosted elsewhere. If that is your bread and butter, it seems a huge risk to take, both from a perspective of protecting your IP and service availability.

Added to that is that your environment is continuously changing. Testing becomes a nightmare and, in practice, sometimes happens in production and employees get confused or even need to be retrained. Simple things like changing the position or availability of buttons can be a real hassle. Even though you can defer upgrades and all that, it's not quite the same as testing and using a known platform.

You surrender an essential bit of control to parties outside of your control. In the case of Microsoft, that party is big enough to be completely impartial to the demise of your company. I am the first to admit that self hosted applications also require updates and carry other risks, but they seem to be a different order of magnitude.

Funnily enough I happened to visit an IT consulting company yesterday who run _entirely_ from the cloud on a collection of services.
They have zero onsite servers and no IT support staff.
They have about 100 employees spread over two offices. Not judging that as right or wrong but that is what they do and they are happy with it and believe they are spending less than they would be in a traditional model.

It is a trade-off. On site has advantages, cloud has advantages. Both have disadvantages and both have risks. I don't believe on-site is inherently less risky than a cloud server.

Both have costs. Maybe you could gold plate your on-site stuff so it is super coolio awesome rock-solid no matter what happens, but how much would it cost and how long would it take and what could you have done instead with that time and money?

Both types of environments change underneath you without notice - assuming you are in a decent size organisation. Been there, done that got the tee shirt.

One thing that is very clear to me is that physical boxes take significantly longer to provision. Sometimes time is money and that one area where cloud services win hands down.

[CatalinaWOW]

Isn't it funny how we have come full circle.

Back in the 60s and 70s the arguments were about having your own computer in house (either a mainframe or the smaller "mini" computers from the likes of DEC) or renting computing from a timeshare service.

For large corporate users today the arguments are essentially the same, and come down to cost efficiency vs control.

In the interim the PC revolution temporarily freed many of us from having to rent computing, finding a one time payment far easier to swallow than an endless revenue outflow.  This approach caused such a leap in capability that the HW and SW leaked back into the corporate world.  The free ride is ending everywhere but the Linux world as software developers move to subscription models.  This conflict between users and developers will continue and force both sides to examine their real needs vs their wants.  The frightening thing is that the only thing I can think of that is a true ongoing need for users is protection from various malware threats.  Developers face a strong temptation to do a less than perfect job of fulfilling this need so that the need remains strong.

[tooki]

I figured as much. Having worked for a software company, I really, really understand why developers want everyone to always use the latest version, and the motivation to change from yearly spurts of upgrade income to an ongoing revenue stream. The objection I have, which I suspect is precisely the kind of thing you run into, is the oft unreliable activation systems themselves.

From the client's perspective, it's not an obvious choice at all. You make your business fully dependent on the continued services of another. If that were to be the case with one of the packages essential to your operation, that might be doable. The problem is that everyone is jumping in and that you now need to manage a whole herd of services and associated risks. Especially offline applications can steadily run for years without trouble if it is a product and not a service. I have trouble understanding why any firm would want all their CAD projects, and the infrastructure to access them, hosted elsewhere. If that is your bread and butter, it seems a huge risk to take, both from a perspective of protecting your IP and service availability.

To be clear, I think you're confounding cloud computing and subscription pricing, which are very diffferent things, despite having a few similarities. We were discussing subscription pricing primarily, and the unfortunate issues surrounding the activation. It's still, however, software running locally with local files.

I'm not sure what exactly you are claiming isn't obvious to a client -- I didn't claim anything was obvious to the client. I said that it's very clear why software companies want to move to subscription pricing. Customers frequently fail to understand it and feel very attacked by subscription pricing, but having seen the nightmare that is customers who refuse to upgrade, and thus continue to cause a support burden because they want to run ancient software whose bugs were long ago fixed in newer versions, or because they want to run ancient app versions on new OSes, or your latest app release on a prehistoric OS, I am ultimately in favor of licensing that encourages or requires users to upgrade constantly.

Added to that is that your environment is continuously changing. Testing becomes a nightmare and, in practice, sometimes happens in production and employees get confused or even need to be retrained. Simple things like changing the position or availability of buttons can be a real hassle. Even though you can defer upgrades and all that, it's not quite the same as testing and using a known platform.

I'm not sure whether you're arguing for or against change. I maintain that resisting change means building up an ever-growing technical (and user-training) debt, which eventually will need to be repaid -- with interest.

You surrender an essential bit of control to parties outside of your control. In the case of Microsoft, that party is big enough to be completely impartial to the demise of your company. I am the first to admit that self hosted applications also require updates and carry other risks, but they seem to be a different order of magnitude.

The risks are so different as to be difficult to compare (there's tons of risk in using an ancient third-party application that's long been abandoned by its developer). But for sure, the single-source nature of both cloud computing and subscription-software activation are a huge risk, and one that the software vendors, frankly, have yet to address in an adequate manner IMHO. (I've read about graphic designers being unable to work for a week at a time because their Adobe suite refused to activate properly, despite having a valid subscription. That's absolutely inexcusable IMHO.)

[Mr. Scram]

Funnily enough I happened to visit an IT consulting company yesterday who run _entirely_ from the cloud on a collection of services.
They have zero onsite servers and no IT support staff.
They have about 100 employees spread over two offices. Not judging that as right or wrong but that is what they do and they are happy with it and believe they are spending less than they would be in a traditional model.

It is a trade-off. On site has advantages, cloud has advantages. Both have disadvantages and both have risks. I don't believe on-site is inherently less risky than a cloud server.

Both have costs. Maybe you could gold plate your on-site stuff so it is super coolio awesome rock-solid no matter what happens, but how much would it cost and how long would it take and what could you have done instead with that time and money?

Both types of environments change underneath you without notice - assuming you are in a decent size organisation. Been there, done that got the tee shirt.

One thing that is very clear to me is that physical boxes take significantly longer to provision. Sometimes time is money and that one area where cloud services win hands down.

If you read one of my previous comments, you'll see that I've seen multiple companies switch back to on premise hosting after the migration to Office 365 didn't work out for them because they experienced too many problems with performance, outages and other issues. Such a migration is an involved and expensive procedure in companies the size they were. Not something you do unless you have to. I will not claim this represents most users or companies, but it does illustrate that it is not for everyone and that theory and practice do not always line up as expected.

Please note that I have also stated that there is a time and a place for cloud solutions. It is a valuable tool that can be massively advantageous in the right circumstances. However, everything seems to get forced into the cloud model, whether it is appropriate or not, and companies are made to depend entirely on cloud services when the need seems absent. For instance, I cannot imagine being an engineering firm and having all your and your client's IP tied up in a cloud service. Even with a good, independent backup, you won't have the infrastructure or training required to continue business if anything were to happen to that cloud service. How much of an impact even a short outage can have was demonstrated by the Blackberry server issues of 2012, with many companies switching platforms because of it.

How can a company of that size have zero support staff? Even if the servers and services are fully operated by another party, there is still the need for local support and maintenance of client workstations and the network. With around 100 users, that is not a negligible task.

To be clear, I think you're confounding cloud computing and subscription pricing, which are very diffferent things, despite having a few similarities. We were discussing subscription pricing primarily, and the unfortunate issues surrounding the activation. It's still, however, software running locally with local files.

I'm not sure what exactly you are claiming isn't obvious to a client -- I didn't claim anything was obvious to the client. I said that it's very clear why software companies want to move to subscription pricing. Customers frequently fail to understand it and feel very attacked by subscription pricing, but having seen the nightmare that is customers who refuse to upgrade, and thus continue to cause a support burden because they want to run ancient software whose bugs were long ago fixed in newer versions, or because they want to run ancient app versions on new OSes, or your latest app release on a prehistoric OS, I am ultimately in favor of licensing that encourages or requires users to upgrade constantly.

Companies are currently pushing hard to tie subscriptions to the cloud. That is their way of combating piracy and increasing their control over the end user. Local software is rapidly becoming just a front end for cloud services. Autodesk with its Fusion 360 software is a good example of this. But of course you are right that they are two distinct things.

I wasn't disagreeing with you regarding the desire of companies to push for subscription pricing. However, I do think the reasoning is too self-centred. A company that's only focussed on improving things for itself will not survive in the long run. A mutually beneficial relationship with the customer is what forges long term relationships.

I also feel there is a middle ground between running ancient applications on a prehistoric OS and constantly forcing a user to update to new versions. Installing security updates is good, but continuously forcing users to migrate to yet another version is less ideal. I think communicating the product lifecycle to your customers is good. They know what to expect and can plan budget and training accordingly. This is what Microsoft did in the Gates and Ballmer era. Now, there are no versions as such and things are much more fluid. You get whatever you get, each release gets one year of support and you just have to deal with it. Added to that, the mechanism has already been abused a few times to push the Microsoft agenda, and the end user is left with a sour taste.

Don't get me wrong, it's not all bad. Microsoft does some pretty amazing things and the effectiveness of much of the software is undeniable. It's just that I think that this push for the cloud without reasonable alternatives is a mistake. Inevitably, a new balance will be reached. Hopefully this will happen without too much friction along the way.

[hendorog]

Isn't it funny how we have come full circle.

Exactly