What is the correct way to sequence DLL versions

Question

I've always understood that each number in the version (ex: 1.2.3.4) is treated as an regular int value. We are using some third party DLLs that follow a different convention and it seems to be causing problems with our patching process (patch won't overwrite the old DLL with the new one).

The old version of the DLL is 1.38 The new version of the DLL is 1.4

I understand things to be 1.4 < 1.38. I would have personally used 1.40. Their claim is that 1.4 == 1.40.

I hate making mistakes due to ignorance and I also hate blindly following conventions without understanding the meanings behind them. So it's time for me to ask for clarification. Any views on this?

Answer 1

Of course, 1.4 is in no way equal to 1.40.

1.40 is revision fourty, 1.4 is revision four, there are 35 revisions between them.

I can say that Photoshop v.10 is certainly better than Photoshop v.1

Answer 2

The convention is 1.4 != 1.40. Versions are a set of numbers, not a single number. They should have been using 1.3.8 if they wanted to increase to 1.4.

Answer 3

It's up to you, but as you've noticed, it's easy for things to get confusing.

In my personal projects, I always numbered with straight integers, never any dots: v1, v2, v3, etc. Kept things simple.

Answer 4

I agree with your reading on it. "1.4" == "1.04" < "1.38"

Answer 5

Of course it is individual integer values! Think of the nightmare if version numbers were floating point numbers, I mean, with rounding and all that. :P