Software Versions and Releases

One of the little quirks in the computer software field that bothers me is when people start talking about "versions" and "releases" as if they were interchangeable.  They're not, and it just rubs me the wrong way.  They most definitely are not synonymous.

When it comes to software, a "version" refers to an adaptation to suit a particular computer environment, for example; a single program that works on MS Windows and another version that works on Linux, another on the MAC OS, etc. (distinctly separate operating systems).  Each operating system has its own unique nuances that prohibits a program written for one operating system from operating on another.  Perhaps the best way to think of this is from the old videotape wars, whereby a movie was distributed in "VHS" or "Beta," one did not work in place of the other.

"Versions" have always been a headache for software vendors.  Inevitably, when a program is first written it is done so for a specific platform, normally one that dominates the industry.  It is then converted to other platforms and incorporates their peculiarities.  This of course means there will always be one version released ahead of another.  To get an idea of how pervasive this problem is, see the Adobe Reader download web page.

"Versions" would be an obsolete concept had everyone adopted the Java programming language years ago whereby a single program could be executed on any operating platform, but this never happened as the software industry tends to buck any attempt of standardization.  Plus it would make the operating system a triviality, something the people in Redmond simply wouldn't sit still for.  Oh well.

In contrast, a "release" is just that; an issuance of software to their customers.  Although, it could be numbered sequentially as 1, 2, 3, etc., most software vendors long ago adopted a three positioned numbering convention, such as "9.02.05".  Under this scenario, the first position refers to a major release of the software, usually with some significant changes to the file layouts; the second position represents modifications/improvements added to the major release, and; the third position represents corrections to defects. This numbering convention served the computer field well for a number of years until, unfortunately, it was bastardized by vendors who would increment the initial number as a marketing ploy to indicate they were ahead of their competitors thereby making it meaningless.  Believe me, comparing the numbering conventions of different vendors is like mixing apples with oranges.  It is simply nonsense.

Then along comes our old friend Bill Gates who decides to break with tradition and release his company's products based on a given year; e.g., Windows 95, 98, 2000, 2003, 2007.  MS Office followed suit, as did many other vendors hanging on Microsoft's coattails.  The only problem with associating a year with software is it has a tendency to put pressure on vendors to produce a new major release every year, as in the automotive industry.  Unlike the automotive manufacturers though, software vendors tend to miss delivery dates and, as such, it is not realistic to expect a major new release every year.  Bottom-line, the whole concept of naming releases after specific years is retarded and should be dropped.  Interestingly, it appears Microsoft has done just that as it prepares to release the next generation of their operating system, "Windows 7."

A software release should denote nothing more than a distinctly separate issuance of a product, nothing more, nothing less.  It should definitely not be labeled for marketing or competitive purposes.  More importantly, stop using the words "version" and "release" interchangeably.  It simply doesn't make sense.  Then again, common sense is not very common when it comes to computing.

One last note, do yourself a favor and never experiment with "beta" release software (experimental).  You can get burned and it is simply not worth it.