Sun Offers Solution for "v1 Uncertainty Theorem" - Screw Version Numbers

What is the "v1 Uncertainty Theorem". It's a term I've created to explain a common phenomenon in the sofware industry. It states "The rate of adoption for any software product will be minimal as long as it is still the first major release." In a nutshell, it means that harly anyone uses v1.0 of a product, no matter how cool it is.

So how did Sun combat this? They said "to hell with the versioning system", and changed J2SE 1.5 to J2SE 5.0. [via Slashdot]

In response, Interscape has decided to rename the upcoming ScrollingGrid 2.0 release to ScrollingGrid 9.0. Just kidding.

This is ridiculous. Who taught Scott McNealy how to count? Oh yeah, these are the same guys who lied on the Java benchmarks. Right, my bad, I almost forgot. 

11 Comments

  • Okay, but don't forget that Visual Studio went from 6 to 2002 in one jump. <grin>

  • and Access 3.0 errrrrr 7.0 I mean. Along with Outlook 2000 errr 2002 errr XP err 2003. Damn, he must have learned how to count from Bill Gates. How about .NET 2.0 errrrr 1.2 with C# 1.2 errrr 2.0 in it?

  • And for that matter the very first version of Office was 4.3.

  • This Theory doesn't seem to apply to OneNote, it's extremely popular, especially on tablets. Everyone who uses it seems to become quite attached to it, myself included.



    Also, the "Year Model" approach is directly related to this whole marketing fear, if it's based on a year, the general population doesn't see any version numbers just people who dig into the properties to find out the build numbers.

  • Version numbering in shrinkwrap software is a function of marketing, not development. So don't go looking for sense there. Having said that, I think no matter how you look at it, J2SE can easily justify calling itself a 5.0. The incremental nature of version number changes of Java runtime has always been ludicrous if you consider the actual functional change from point release to point release.

  • Visual Studio is still referenced as Version 7 for 2002, v7.1 for 2003, and v8 for 2005.



    Office is still referred to as Office 10 for XP and v11 for 2003.



    Scott: The next revision of .NET was originally versioned 1.2. Then the decided it would be a major revision not a minor one, so they bumped it to 2.0. C# is now V2, but you also have VB.NET 2.0 (which is really VB8).



    The year-based approach is fine. Call it J2SE 2004 for all I care. But jumping 4 version numbers for no apparent reason? Come on.

  • Netscape tried this (going from 4.x to 7.x) and it didn't help them. Mozilla and Firefox stick to a normal versioning system (even though I'm waiting to see if they'll go to 1.10 after 1.9, or bump up to 2.x) and are increasingly popular. It's not a matter of version numbers, it's a matter of quality. If your code/product sucks it doesn't matter what you call it, and higher version numbers don't mean better quality (Windows ME or Visual Studio 7 are prime examples of crappy softwares with high version numbers).

  • The first version of Exchange Server was 4.0. It was following on from MS Mail which was at 3.x. Completely different products, different code base, etc.

  • Although I agree with the premise, calling it an "uncertainty theorem" seems an egregious misuse of the term to elicit familiarity due to Heisenburg's Uncertainty Principle. It's especially so considering your statement of the theorem is with considerable certainty; well, there's at least no inherent uncertainty in the premise to warrant the description of being uncertain.



  • Didn't Microsoft acquire the technologies for Office, mail, and Exchange separately? and wasn't MSMail a part of Windows 3.0, which would then explain the version number there?

  • You should check out the Solaris numbering scheme when you've got some sparet time. It's a real mess.



    You have to give points to Sun for being consistently inconsistent.

Comments have been disabled for this content.