Skip to content


Semantic Versioning

Today, I read about Semantic Versioning. As a first version, I think this document lays out a great plan for versioning software. One problem that I see is there is more than one great way to version software. X.Y.Z is good for software that you release to others, but for software like websites, X.Y.Z doesn’t work too well.

I’d like to propose another scheme (the name is tentative of course):

Website Semantic Version Specification

When moving a website to production, you must version it, so that you can either refer or rollback to a previous iteration. In order for everyone to be on the same page, I present not one, but two ways to version your website.

The first version is YYYYMMDDII, where YYYY is the four digit year, MM is the two digit month, DD, is the two digit day, and II is the two digit iteration. This allows for 100 unique iterations a day if you work that fast, and it also gives a general time when the website was released.

The second version is YYYYMMDDHHMM, where YYYY is the four digit year, MM is the two digit month, DD, is the two digit day. You also have HHMM which is the current time. The benefit of this method that you can see exactly when the website was released. A negative point in this case is that you will have to agree on the timezone beforehand, because it is always noon somewhere.

Posted in Smarticus Says.

Tagged with .

  • HotFusionMan
    I think UTC is the only sensible time zone to use for timestamping releases. It doesn't observe DST, and everyone knows (or can easily look up) how many minutes their own time zone is offset from it by. So there's no ambiguity about when the release occurred.
  • I actually agree. But some people are silly and like to defy conventions.
blog comments powered by Disqus