Hacker News Reader

Comment by b_lindahl | Hacker News Reader

Visual Studio "15" = Visual Studio 2017 = VC 15.0 Crystal clear. Why didn't Microsoft bump up VC version to 17 to sort of make VC version follow VS version. Would clear out some confusion going forward.

View on HN

I am beginning to think that Microsoft's bad verison naming is entirely on-purpose.

Windows has had terrible naming since NT -- back and forth between names/numbers/years (or nonsense like "Anniversary Update" or "R2"), have a completely different scheme for Server vs Desktop, and if you look at the corresponding kernel versions of these releases[0], the major/minor bump in versions seems completely arbitrary.

.NET is even worse. Framework vs CLR is a mess [1]. Core started over at 1.0. .NET Standard brings some consistency to it all, but also introduces another versioning scheme with overlapping numbers that mean different things[2].

But even other divisions do this: Xbox -> Xbox 360 -> Xbox One.

[0] https://en.wikipedia.org/wiki/Windows_NT#Releases

[1] https://en.wikipedia.org/wiki/.NET_Framework_version_history...

[2] https://blogs.msdn.microsoft.com/dotnet/2016/09/26/introduci...

You forgot about Windows 9. ;)

Was meant to have been skipped because it was too similar to Windows 9x (Win95, etc...).

Which wouldn't have been a problem if they followed a consistent naming convention in the first place...

On purpose... but any guess as to why?

Windows 2000 = 5 Windows XP = 5.1 Windows Vista = 6.0 Windows 7 = 6.1 Windows 8 = 6.2 Finally, they have Windows 10 == 10

No, 2017 written as separate numbers are 2 0 1 7 and adding them up gives 10. Half of 10 is 5 (obviously) and that sum is 15. Get it?

or perhaps 0 is some kind of infix flip-subtract operator

a 0 b := b - a

Call in Ramanujan. He'll surely come up with a good explanation.

https://en.wikipedia.org/wiki/Srinivasa_Ramanujan#Hardy.E2.8...

"No! It is a very interesting number. It is the smallest positive integer that Microsoft have not yet used as a version number."

Actually that would be 13.0!

The naming has been awful with this release in particular. Having two Visual Studio (20)15s out at the same time is very confusing. Almost as bad as the Skype vs Skype for Business marketing boondoggle.

There's some sick bastard in Microsoft who really likes these confusing sames. He's probably the same guy who came up with .NET.

Remember when .NET could refer to virtually every product Microsoft made?

I was half expecting them to call it Visual Studio Surface. Then you could run Surface on your Surface on any Surface on the planet.

Indeed, .NET must be the worst, most misleading, least googleable product name ever. I suppose we should be grateful that MS called an earliest technology COM not .COM.

I must say Google handles it well. I'm searching with these terms quite often and Google knows .NET == dotnet ~ C# and finds results without much problems.

Google is terrifyingly good at finding relevant developer documentation. I'm convinced it will be writing software for us in a couple years.

The first result I get googling for ".net" is the Wikipedia article ".NET Framework". The first result I get for "COM" (or "com") is Microsoft's page on "COM: Component Object Model Technologies".

Honestly, I have never had a problem googling .NET and COM-related stuff.

It was mostly a problem when you were having a conversation with someone that went something like...

"So I was installing .NET and it failed because I didn't have the right drivers for my SCSI card."

"What? I've never needed disk drivers to install .NET, which version was it? 1.1?"

"No, it was 2003"

or...

"Can you install .NET on my PC?"

"Sure, done"

"Where's Word and Excel?"

"What?"

This is after it became popular, and Google (I assume) just special-cased it. Back when it first came out, though, typing that into any search engine we had back then would mostly just give you tons of domain names that end with ".net".

Then also, we had Office .NET, Visual Studio .NET (with Visual Studio .NET 2003 added later), and pre-release versions of what was to ship asWindows Server 2003 were known as Windows .NET Server for a while.

Web services (the SOAP and XML kind) were the next big thing back then, at least as far as Microsoft was concerned, and so the entire software stack was promoted as conceptually centered around that - hence the .NET name. Kinda ironic, what with the rise of cloud computing later, when .NET brand was already toned down to refer to the framework specifically, and the consequent need to invent "Azure".

Back then I just used MSDN.

It could have been worse. The original internal name for .NET before the rebranding was COM+ 2.0, if I remember correctly. You can still see some legacy of that in names like COMPLUS_ZAPDISABLE (an environment variable that disables loading of ngen'd assemblies by the CLR).

Somewhat more indirectly, it also surfaced in low-level native APIs used to host CLR inside an app, with names like ICorRuntimeHost - if I remember correctly, COR here stands for Common Object Runtime. Also seen in mscorlib.dll (the assembly containing the most fundamental types, which is the top node in any CLI assembly dependency graph) and mscoree.dll (EE = execution engine).

Oh, and C# was COOL (C-like Object-Oriented Language).

> confusing sames

I can't tell if that's a typo or intentional. I might start calling them that.

I am at my cleverest when it's completely by accident.

On the other hand if you want an SQL server how does MS call it? SQL Server. Unlike for AWS you don't need a glossary to get around in Azure. (At least less so) This is why this disparity in naming thing very simply and straightforward and mind-boggingly badly is incomprehensible for me.

Ah yes Office .Net.

Such confusion. They then went and tagged "Live" on everything.

It's just a product with a name that has a year in it, but the executables inside each have their own different version.

For example for Visual Studio 2015, the visual studio executable is at version 14, and the C++ compiler is at version 19.

So would you also suggest changing the C++ compiler version back to 17 now?

Even the MS installers get this mixed up sometimes. Sp5 to vs 2015 mixed up some paths between vc12 and vc14.

Many software packages have poor or confusing version numbering. Java, IIRC, jumped from Java 1.2 to 2 or 1.4 to 4 or something like that. Windows 3.1 to 95 to 98 to Me to NT (stood for New Technology, ha ha [1]) to 2000 to XP to Vista to 7 to 8 to 10 ... (might have got some of the order wrong there).

[1] New, as in the classic marketing/advertising phrase "new and improved", maybe ... :)

NT was called "New Technology" literally because it was a new technology. It was based on the NT kernel (which is used now) as opposed to the 9x/MS-DOS kernels used in 1.0-Me.

Also, NT came before 95. Specifically, as Windows NT 3.1.

>NT was called "New Technology" literally because it was a new technology. It was based on the NT kernel (which is used now) as opposed to the 9x/MS-DOS kernels used in 1.0-Me.

Thanks; I was aware of that when I wrote my comment that you replied to. I was around in tech when NT was released, so I had read the expansion of the acronym then in the media. In fact I had read the book Inside Windows NT (the one by Helen Custer, a good one) some time after it came out. Remember it talking about Dave Cutler as the main NT architect, his DEC VMS background, etc. And it had a fair amount of interesting content about NT internals. Also remember trying this Linux C utility that I wrote, selpg (which I posted about in this other recent HN thread - https://news.ycombinator.com/item?id=12960242), on Win NT, because I had read that it had a POSIX subsystem (then and for a while). And it worked on it, including the POSIX popen() call (though not to an lp command, since NT didn't have a UNIX-like lp subsystem). My point in mentioning marketing was that even if Windows NT really was new technology (and it was), giving the name "New Technology" was somewhat unusual (after all, any non-trivial product upgrade is usually new tech, unless only cosmetic) , and so I thought it might have been the marketing folks's decision. And we know that marketing people do sometimes push product names for marketing reasons, that tech people might not always agree with. That's all there was to it.

>Also, NT came before 95. Specifically, as Windows NT 3.1.

I did clearly say above:

"(might have got some of the order wrong there)."

This is what happens when marketing determines your version names.

Reply on news.ycombinator.com