"Windows 9" "Threshold" and "Windows" Rumors thread

[eschwartz]

Originally Posted by HarryT

... except that you don't test for Windows versions like that. You do it by calling the "GetVersion" or "GetVersionEx" Windows APIs, and they return Windows versions as a "major" and "minor" version number, not a string. There's no possible way to confuse Windows 95 with Windows 9. Windows 95 returns a version of "4.0", Windows 98 is "4.1".

Clearly some people don't do the right thing...

[MikeB1972]

Originally Posted by HarryT

... except that you don't test for Windows versions like that. You do it by calling the "GetVersion" or "GetVersionEx" Windows APIs, and they return Windows versions as a "major" and "minor" version number, not a string. There's no possible way to confuse Windows 95 with Windows 9. Windows 95 returns a version of "4.0", Windows 98 is "4.1".

In windows programming, yes, I'd agree.

But there are issues in some languages.
e.g java has a popular function
System.getProperty("os.name")
that does return Windows 95 etc as a string

[DMcCunney]

Originally Posted by MikeB1972

In windows programming, yes, I'd agree.

But there are issues in some languages.
e.g java has a popular function
System.getProperty("os.name")
that does return Windows 95 etc as a string

Sure, but it will be an unambiguous string. You won't be able to confuse Windows versions unless your routine to parse the returned value is brain-dead and not looking at the whole string. That's your error, not an OS issue.
______
Dennis

[eschwartz]

Originally Posted by DMcCunney

Sure, but it will be an unambiguous string You won't be able to confuse Windows versions unless your routine to parse the returned value is brain-dead and not looking at the whole string. That's your error, not an OS issue.
______
Dennis

Yep, basically:

"Some people do stupid things to check the version, so we are going to skip a version number to cater to them" rather than, say, allowing the developers of a brain-dead program to fix the issue.

I'm not sure I buy this explanation, although I have heard it explained as a way to ensure backwards compatibility.

[DMcCunney]

Originally Posted by eschwartz

Yep, basically:

"Some people do stupid things to check the version, so we are going to skip a version number to cater to them" rather than, say, allowing the developers of a brain-dead program to fix the issue.

I'm not sure I buy this explanation, although I have heard it explained as a way to ensure backwards compatibility.

I don't buy it. Since there are clearly defined Windows APIs for doing things like checking what Windows version the user has that won't be subject to the sort of confusion mentioned, it's reasonable to assume developers will use them.

And what's in it for MS to handle edge-cases where where developers are ignorant and don't use them? That's the dev's problem, and they need to learn better or find something else to do with their time that doesn't involve writing bad code.
______
Dennis

[fjtorres]

Originally Posted by HarryT

... except that you don't test for Windows versions like that. You do it by calling the "GetVersion" or "GetVersionEx" Windows APIs, and they return Windows versions as a "major" and "minor" version number, not a string. There's no possible way to confuse Windows 95 with Windows 9. Windows 95 returns a version of "4.0", Windows 98 is "4.1".

You may not, but the report clearly states that such code can be found in use.

[eschwartz]

Originally Posted by DMcCunney

And what's in it for MS to handle edge-cases where where developers are ignorant and don't use them? That's the dev's problem, and they need to learn better or find something else to do with their time that doesn't involve writing bad code.
______
Dennis

Kind of what I was thinking.

Backwards compatibility should mean that well-written code does not break simply because of updates to the APIs. Not bending over backwards to account for rough equivalents to the Y2K bug.

[HarryT]

Originally Posted by eschwartz

Kind of what I was thinking.

Backwards compatibility should mean that well-written code does not break simply because of updates to the APIs. Not bending over backwards to account for rough equivalents to the Y2K bug.

Precisely. The Windows 95 version of my astronomy software (highly complex architecturally, multi-threaded, etc) runs absolutely fine nearly 20 years later on Windows 8.1. Why? Because I wrote it properly and followed all the rules.

[eschwartz]

Originally Posted by HarryT

Precisely. The Windows 95 version of my astronomy software (highly complex architecturally, multi-threaded, etc) runs absolutely fine nearly 20 years later on Windows 8.1. Why? Because I wrote it properly and followed all the rules.

I agree, I am just saying, "don't" is not the same as "shouldn't".

Originally Posted by HarryT

... except that you don't test for Windows versions like that. You do it by calling the "GetVersion" or "GetVersionEx" Windows APIs, and they return Windows versions as a "major" and "minor" version number, not a string. There's no possible way to confuse Windows 95 with Windows 9. Windows 95 returns a version of "4.0", Windows 98 is "4.1".

[Glorfindel]

Originally Posted by eschwartz

Name me the unix-based terminal that doesn't do that, bundled or not...
(I use mintty with cygwin. Also did this for years. )
Now if only they would make some useful functionality available via the command-line, so I don't have to install cygwin in order to copy files around.
Bonus points for tab-completion.

Default terminal in Slacko doesn't.

[Glorfindel]

I have installed the Technical Preview, works like a charm runs faster than lubuntu on my box. Specs are 1GB RAM, 3GHZ P4 cpu. 500MB RAM usage with a few programs up, most used so far was 700MB.

[Jack Torrance]

im in the works of doing a fresh install of my home machine and decided for kicks to load the Tech preview on another drive and set them up the same way.

Admittitly my tine spent with 8/8.1 was pretty short but I found it to be off putting. The lack of the start menu while probably something to work around and get used to was very annoying especially with 20+ years of muscle memory built in. While change is nice at times you want to see a functional reason for it as well and as far as I've read that was elusive.

This new version feels more like a progression and seemed to be able to do what I wanted without having to relearn everything. So far so good.

[Glorfindel]

Yeah, I thought that the Start Menu combined 7 with 8 very well.

[fjtorres]

Originally Posted by Glorfindel

I have installed the Technical Preview, works like a charm runs faster than lubuntu on my box. Specs are 1GB RAM, 3GHZ P4 cpu. 500MB RAM usage with a few programs up, most used so far was 700MB.

Good to hear.
Among other things, it bodes well for the cheapy WinTabs.
Full Calibre on a reader-sized tablet should be fun.

[Glorfindel]

Especially with calibre for ARM devices seeming to be never coming.