Development tool life cycle is different from the OS

So as to not hijack the %of OS thread I started a new thread to continue the conversation that rab36 brought up (http://channel9.msdn.com/Forums/Coffeehouse/-of-OS-being-used-by-business/7725765952bf46a8910c9fa000d3ee29).

We saw similar behavior with VS2010 dropping mobile capabilities.  VS.Next dropping pre-vista seems like a disconnect.

Isn't it fair that the development tool business spreads across the OS life cycles rather than stepping with it.  What is going on?

For our customer base we support Windows 2000 and up through Windows 7 and Windows 2008 R2.  This year is the 1st year where we might decide to drop support for Windows 2000.  The Microsoft ecosystem has always been about backwards compatibility so the customer base can transition in a non-jarring way.

This is bigger than the discussion surrounding Windows 8.  This is about the ecosystem being jarred which then leads to the end consumer and businesses being jarred.

Well the alternative is that Microsoft needs to spend more and more time testing Visual Studio for every new OS rather than spending time improving it, because the current version of VS needs to work on XP, Vista, 7 and 8 and the next version will need to work on Vista, 7, 8 and 9.

If you bear in mind that Microsoft are also breaking support for XP imminently and that the majority of customers have now moved away from it, isn't it about time that we freed the VS team to improve VS rather than just supporting crusty old unsupported OSes?

@evildictaitor: I need to read some more on this.  The major point I thought was not that VS had to run on older OSes but that the binaries VS builds could run on the wider spectrum of OSes.

I felt the conversation around that would take away from the original poster's thread about % of OS being used by business.

I'm not sold that VS.Next would build binaries that would not run on XP.  However, the Connect responses seemed valid enough to warrant discussion.

@davewill: But why? Anyone still actively developing for XP can continue using the older versions of Visual Studio, whilst anyone else can take greater advantage of features of newer operating systems can use the latest version. Continuing to support XP means that things like the CRT can't take full advantage of security features like ASLR, not to mention CPU features which might not be supported on XP machines (but would on Vista+ due to the system requirements).

There really does have to be a point when people just accept that XP is going away, expecting developers (whether in Microsoft or elsewhere) to keep compensating for it and maintaining applications on it is just going to become an increasingly costly option.

@AndyC: The whole point of the Wndows ecosystem is to create software that targets a large footprint which allows the costs to be spread over a larger base.  One concern is the fragmentation of the footprint which will mean less span over which to spread the costs.  The other concern is increased costs due to the added complexity of dealing with builds tied to a specific VS and/or OS set.  These increased costs would need to be spread out but if the footprint is also shrinking at the same time it becomes a perfect storm for price increases.

@davewill:

That's sad, but, I think Microsoft BC support is good enough. Also, too much BC is bad for many many reasons, and I can go on and on and on, and you can reply back and back and back. In the end, you have to understand simple fact, Microsoft only support BC up to certain years, not forever, not 30 years. If you have a problem with that, Microsoft products are not good enough for you, meaning it is time for reality check and find a product that will match you expected lifespan.

@davewill:Thank you for moving this to its own thread - I should have done this by myself in the first place.

@magicalclick: No I do not agree that BC is good enough when VS 11 can't create programs for Windows XP while Windows XP is still a supported prodcut. VS 11 launch in 2012, XP support ends in 2014.

@AndyC: this is state-of-the-art in VS 2010: you can choose the right #define WINVER and then you can either use Vista or Win 7 features and loose compatibility with Win XP or you choose not to do so. What seems to be the case in VS 11 is, that the CRT and MFC depend on Vista features and there is no way around it. I want to choose by myself and I have a strong feeling that dropping the Windows XP support is too early.

Maybe there is a very good technical reason for dropping XP support, that can be discussed, but then Microsoft should start this discussion early (now).

E.g. we are selling maintenance agreements for our software to our customers and I have to know now what OSes I have to / can support in 3 years time.

Also there is a lot of good new C++ stuff in VS 11 (C++ 11 standard compatible libraries, updated PPL / ConCRT with task based concurrency, C++ AMP) and it would be a big disappointment not being able to use this until 2014.

Bernd

7 hours ago, rab36 wrote

E.g. we are selling maintenance agreements for our software to our customers and I have to know now what OSes I have to / can support in 3 years time.

Well that's easy, XP support dies in 2014 so if you're selling 3 year maintenance for XP today, you're already creating a potential problem for yourself. How you choose to deal with it is up to you, personally I'm not sure that roadmapping support beyond the lifetime of the OS is particularly helpful for you or your customers, but I guess it depends upon how critical your application is to the businesses you supply.

12 hours ago, AndyC wrote

*snip*

Well that's easy, XP support dies in 2014 so if you're selling 3 year maintenance for XP today, you're already creating a potential problem for yourself. How you choose to deal with it is up to you, personally I'm not sure that roadmapping support beyond the lifetime of the OS is particularly helpful for you or your customers, but I guess it depends upon how critical your application is to the businesses you supply.

++

Visual Studio won't drop support for XP until Microsoft does, so the next version of Visual Studio will still allow you to make XP binaries. The version after that probably won't, and the version after won't let you do Vista and so on.

Why do people think that it is Microsoft's duty to help people stay on old and insecure software rather than helping them upgrade?

@evildictaitor: Where do you have this info from? Have you read http://connect.microsoft.com/VisualStudio/feedback/details/690617?

1 hour ago, rab36 wrote

@evildictaitor: Where do you have this info from? Have you read http://connect.microsoft.com/VisualStudio/feedback/details/690617?

That's interesting - I knew the frontend wasn't going to work on XP, but I didn't know the CRT was ditching it too.

The .NET outputted binaries will still work on XP (they use the framework CRT, not the VS CRT) and you'll still be able to use old versions of Visual Studio both on and for XP. If you're targetting XP why do you need the latest and greatest compiler? VS2010 end of lifes a long way into the future.

Also the end of life for XP has already passed - mainstream support for XP expired on April 14th 2009, so if you're still using XP, expect it to deteriorate and for Microsoft to not support it. The 2014 date is the end of extended support.

2 days ago, evildictaitor wrote

the majority of customers have now moved away from [XP].

That is the opposite of what I see.

My company of 1200 employees is still a mix of NT4, 2000, and XP.

The local hospital of over 15 000 employees is still XP. I know for a fact that they use a .NET-based electronic medical records system.

Most non-geek consumers who haven't bought a new PC in the last five years are still on XP.

Do I think we should move forward and drop support? Of course; as a geek, I love the latest thing. But I don't make those decisions. Companies will continue to use XP (by buying Win7 licenses and downgrading them) until the hardware is no longer supported.

1 hour ago, spivonious wrote

*snip*

Do I think we should move forward and drop support? Of course; as a geek, I love the latest thing. But I don't make those decisions. Companies will continue to use XP (by buying Win7 licenses and downgrading them) until the hardware is no longer supported.

Geek arguments about the latest thing aside, having NT4 boxes in production usage should highlight deep concerns about how well the IT department is doing it's job. They must surely represent a significant liability that really ought to be resolved.

@AndyC: There are multiple factors that contribute to why the latest and greatest isn't always in production.  Budgets and expectations are probably the 2 biggest.  If it isn't broke don't fix it is the 3rd.  Budget contraints manifest themselves as lack of man hours or lack of hard cash to implement an upgraded OS.  Having a Windows OS last 5 or even 10 years is not unusual to hear on the expectations side.  Inevitably upper management errs on the side of caution with the if it isn't broke logic.

I don't like any of the contributors but it is what it is and the IT department reflects that regardless of their desires otherwise.  I can see why in a hospital environment a known entity that is working and has a known set of controllable characteristics would not be touched if it need not be.

1 hour ago, AndyC wrote

*snip*

Geek arguments about the latest thing aside, having NT4 boxes in production usage should highlight deep concerns about how well the IT department is doing it's job. They must surely represent a significant liability that really ought to be resolved.

The machines still work and the software on them is most likely not updated to run on anything else. If the PC dies, they'd get a new one with XP on it. The NT4 machines aren't connected to the Internet, and most are just interfaces for laboratory equipment, so there's not really a security concern.

It's just not in the budget to replace PCs for no reason.

7 hours ago, evildictait​or wrote

*snip*

The .NET outputted binaries will still work on XP (they use the framework CRT, not the VS CRT) and you'll still be able to use old versions of Visual Studio both on and for XP. If you're targetting XP why do you need the latest and greatest compiler? VS2010 end of lifes a long way into the future.

I doubt that the .NET Framework 4.5 will support Windows XP when the VS 11 CRT will not. Admittedly, this is pure speculation.

There are a couple of reasons to use the latest compiler: C++ 11 (at least the library stuff has been implemented), C++ AMP etc. For .NET 4.5: async / await is an exciting new feature, that surely everybody wants to use.

Why should I be forced to wait for another two years before I can use these features in my products?