By Harry McCracken | Tuesday, July 14, 2009 at 2:17 am
In theory, software version numbers should be about as scintillating as as serial numbers, house numbers, channel numbers, or Vehicle Identification Numbers. You don’t get much more mundane than the practice of keeping track of a software package’s major and minor editions by assigning decimal numbers to them.
Except…version numbers long ago stopped being version numbers. Software companies started using them as marketing weapons. They tried varying methods of assigning identities to applications, such as naming them after years. They decided that numbers were too dry and substituted letters and words that were meant to be more evocative. I’m not embarrassed to admit I find ‘em interesting enough to write this article.
I cheerfully admit to using the broadest possible definition of version number in this story–hey, I’m going to discuss names that don’t involve numbers at all. I know that developers still use more formal, traditional software versioning naming conventions behind the scenes. (Windows Vista, for instance, is officially version 6.0 of Windows; Technologizer is on version 20593, but don’t ask me to explain why.)
For no particular reason, I’m going to write this as a FAQ. Even though there’s an awful lot about this topic which I just don’t know…
I wish I could tell you–actually, I’m hoping that someone reading this will be able to. I do know that the FORTRAN II programming language came along in 1958, so software developers having been using numbers to keep track of software versions for more than half a century. By the early 1970s, Unix’s developers were keeping track of its evolution using “edition” numbers–second edition, third edition, and so on–which referred to the revisions to the Unix manual.
I’m not sure if there was a period when version numbers weren’t widely used for personal computer software, but if so, it wasn’t long. When I got into computers in 1978, standard version numbers and point releases (such as Radio Shack’s famously unusable TRS-DOS 2.1) were already part of the culture.
I’m sure that someone will top this within moments of the time I click “Publish” to push this story onto the site, but the highest version number I know of belongs to Broderbund’s The Print Shop 23, a product which (A) has been around for 27 years; (B) has received almost-yearly updates; and (C) uses a traditional version number instead of a year or other identifier.
The modern version of the Broderbund company seems to specialize in elder-statesman software: It also publishes PrintMaster Platinum 18, Mavis Beacon Teaches Typing 20, and Calendar Creator 12. It also offers something called ClickArt 1.2 Million, but in that case I’m pretty sure that the number doesn’t refer to the version.
Honorable runner-up to The Print Shop for highest version number: the legendary text editor Emacs, which has reached version 22.3.
That’s easy–at least if you ask me. It’s 5.1. Which is, of course, the version number of WordPerfect 5.1–a DOS word processor so famous, respected, and durable that it’s identifiable by its version number alone two decades after it first appeared…and people still want to use it. I don’t know of a single person who remembers it who considers it to have anything but extremely positive connotations. (WordPerfect acknowledged it was a magical number by releasing a Windows version of WordPerfect 5.1, too–even though it doesn’t make a whole lot of sense to apply a point-release version number from a product for one platform to a version for another.)
Runner up for best version number: 3.11. As in 1993’s Windows for Workgroups 3,11, one of the best versions of any operating system ever released. (I’m convinced that it’s the supposedly minor versions of Windows that are the most significant–any edition with anything after the decimal point shows that it benefited from bug fixes.) The 3.11 in WfW 3.11 must be the most famous two-decimal-place version number ever–and yes, I miss the era when Microsoft (or any software company) would put mighty marketing muscle behind version 3.11 of anything .
Nope–it’s a far older conceit than that. The earliest example I’m aware of is Fortran 66–no relation that I know of to Sergio Mendes and Brasil 66–an update to what was even then a venerable programming language. One of the first operating systems I ever used was an alternative TRS-80 OS called NewDOS/80. Adobe released Illustrator 88. (There was a WordStar 2000, but it probably doesn’t count–it was released in 1985, and presumably intended to sound futuristic.)
Still, Windows 95 remains the most famous product with a year in its name, and it started a trend which continues to this day, But it fell victim to multiple delays that left it being released on August 24th, 1995, well over halfway through the year. With the exception of tax packages–whose version names refer to the tax year in question–I can’t imagine anyone releasing a product today named after a year that’s already winding down. In fact, I’m surprised that Microsoft didn’t rename the OS as Windows 96 before shipping it.
For obvious reasons, the industry seems to have settled on using years only for products that are updated annually or nearly so. I acknowledge that year-based names helpfully tell you roughly when a product was released–but they still make me think of Airport 1975.
(Illustrator 88 box borrowed from Publicious.)
Back in 2001, Microsoft said the XP stood for “Experience,” and that it symbolized “the rich and extended user experiences Windows and Office can offer by embracing Web services that span a broad range of devices.” Pretty awe-inspiring! I’m not privy to the discussions that led to the moniker, but I suspect that the company thought that both traditional version numbers and years felt dry and stale. Maybe it was motivated by some of the same thinking that led Intel to replace a processor called the 80486 with one known as the Pentium.
The XP name proved only moderately influential. In 2002, Macromedia released a product called Studio MX, which was replaced by a product with the belt-and-suspenders name Studio MX 2004. AMD released the Athlon XP CPU, which struck me as a naming faux pas at the time, since it suggested a link with Windows XP that didn’t exist sounded like AMD was saying the chip wasn’t a good fit for other OSes. And today, Adobe offers [a product which many people, including me, mistakenly believe is called] Creative Suite CS4 (the “CS” stands for “Creative Suite” and is therefore wholly unneccessary).
My only objection to names like Windows XP is that they don’t give you a clue about a product’s relation to its predecessors and descendants. In the last eleven years, Microsoft has released Windows 98, Windows 2000, Windows Me, Windows XP, and Windows Vista, and is about to ship Windows 7. Thousands of years from now, technology historians are going to have trouble figuring out what was released when, and may even believe that Windows 98 was the ninety-first release following Windows 7.
As with Windows XP, Microsoft went for a version identifier that was meant to evoke an emotional response rather than keep track of its relationship to previous versions. The “Vista” in “Windows Vista” meant to convey that the OS was “Clear, Confident, Connected,” Microsoft explained.
Setting aside the question of whether Vista was clear, confident, and connected, the name is as far as Microsoft or anyone else has gotten from old-style version identifiers. It’s more Pepsi Jazz than Coke Zero, and I’m not so sure that its expansive-but-elusive quality wasn’t a minor contributing factor to Vista’s poor reputation.
Well, in principle, giving a software product a version number that really is a version number isn’t a big decision–it’s just a statement of fact. But considering that Windows 7 is the first version of Windows with a straightforward version number since Windows NT 4.0 back in 1996, the moniker is a meaningful statement of some sort. I’m not sure if Microsoft has articulated publicly why it chose to drop the non-number naming convention it used for Windows XP and Vista, but the obvious explanation would be that it’s trying to position Win 7 as a low-key, high-value OS that delivers more than it promises–a sort of anti-Vista. The lucky connotations of the number 7 probably don’t hurt either.
Also, “Windows Vista II” wouldn’t have made anybody happy.
According to Microsoft’s accounting, these have been the major versions of Windows:
Windows 95, 98, SE, and Me–all of which count as Windows 4.0
Windows 2000–which was Windows 5.0
Windows Vista–which was Windows 6.0
Tell me about it. Did I mention that the official version number of Windows 7 is…Windows 6.1?
Its numbering is remarkably conservative, and I’m not sure why. OS X 10.0 (the successor to System 9, but based on NeXTStep) came out in 2001, and in the eight years since, the OS has only wriggled its way up to OS X 10.5. (Version 10.6 is due in September.) It’s true that Apple releases fairly frequent upgrades with healthy-but-not-overwhelming quantities of new features, but just about any other software company on the planet would have gotten to version 14 or so by now. It leaves me wondering what, if anything, would get Apple to roll over the version number to 11–and, if it did, whether it would call the product OS X 11.0 or OS XI 11.0.
Of course, most Mac fans refer to OS X iterations not by version numbers but by their cat-themed codenames–OS X 10.5 is Leopard, and OS X 10.6 will be Snow Leopard. Apple started making the codenames into official monikers for the shipping versions in 2002 with OS X 10.2, also known as Jaguar. It’s an interesting approach, but the company is running out of wild cats to name versions after. I’m not sure if it’ll ever release a Cougar (it may not like the alternative definition) or Lion (which may sound too much like Lyin’). But OS X Lynx still seems like a good possibility, and I’m personally rooting for OS X Ocelot.
OS X is also noteworthy for turning the version number into the very name of the product. I’m not sure if anyone’s done that before.
Open-source projects, more often than not, are admirably restrained–maybe because they’re not under the same marketing pressures as commercial software. The Linux kernel has been around for nineteen years and has only reached version 220.127.116.11; the HandBrake video transcoder is at version 0.9.3 after more than six years.
On the other hand, I’m still confused by the way the same basic office suite is marketed both as StarOffice 9.0 and OpenOffice.org 3.1. And Google’s Chrome browser is an open-source project, but a mere five months after version 1.0 left beta, the company released a version 2.0 that other developers might have called a 1.1 or 1.5.
Yes. Or at least a funniest numbering system. The one used by Donald Knuth’s TeX typesetting language. To quote Wikipedia:
Since version 3, TeX has used an idiosyncratic version numbering system, where updates have been indicated by adding an extra digit at the end of the decimal, so that the version number asymptotically approaches π. This is a reflection of the fact that TeX is now very stable, and only minor updates are anticipated. The current version of TeX is 3.1415926; it was last updated in March 2008. The design was frozen after version 3.0, and no new feature or fundamental change will be added, so all newer versions will contain only bug fixes. Even though Donald Knuth himself has suggested a few areas in which TeX could have been improved, he indicated that he firmly believes that having an unchanged system that will produce the same output now and in the future is more important than introducing new features. For this reason, he has stated that the “absolutely final change (to be made after my death)” will be to change the version number to π, at which point all remaining bugs will become features.
That’s nerd humor, I know, but version numbers are inherently nerdy, so it’s only appropriate.
I’m not sure if we know–has anyone ever been brave enough to sell a version 13 of anything? Relatively few applications have reached version 12 and therefore been faced with the question of what to call the next upgrade. Microsoft’s code-name for what became Office 2007 was Office 12; it didn’t startle anyone in the least when the code-name for the next version was Office 14.
Corel has come closer to honesty here than any other company I know–it called the thirteenth editions of CorelDraw and WordPerfect “X3,” for 10+3, then continued the approach with CorelDraw X4 and WordPerfect X4. May both products flourish long enough that Corel has to figure out what to call the 666th versions.
Yes. The sluggish and buggy Windows Millennium Edition (aka Windows Me) turned out to be the least-loved version of Microsoft’s OS. (At least until Vista came along–although Vista has defenders and WinMe, as far as I know, doesn’t.) And Lotus SmartSuite Millennium Edition turned out to be the last real upgrade to a venerable product. (IBM still sells it as SmartSuite 9.8, but is too embarrassed to call it Millennium Edition.) On the other hand, Windows 2000 and Microsoft Office 2000 didn’t seem to be jinxed. Lesson: The whole idea of the millennium being exciting got old quick–and the cuter the version number, the more dangerous.
Sure–or they’ve been known to stretch the truth, at least. For instance, back in 1993, Microsoft Word for Windows made a great leap forward from version 2.0 to 6.0–which conveniently put it at numeric parity with archrival WordPerfect 6.0. In 2002, Netscape went directly from 4.8 from 6.0–coincidentally matching archrival Internet Explorer 6.0. Digital Reaearch’s DOS clone DR-DOS premiered with version 3.41 (making it sound more advanced than MS-DOS 3.3) and skipped 4.0 because MS-DOS 4.0 was a flop.
Not in terms of their use by software developers; it’s simply mandatory that programmers keep track of each version of an application, and there’s no rational way to do it except to assign each one a number. But as traditional software gives way to Web-based services that roll out new features continuously rather than in occasional outbursts, version numbers just aren’t what they used to be. I have no idea what to call the current version of Gmail, for instance.
So I’m guessing that version numbers will eventually go the way of dot-matrix printers–they’re too useful to face true extinction, but most of us will sort of forget they still exist. I’d love to be proven wrong, though. I may not be around in 2086, but I’d like to think that Broderbund will be–and that it’ll be releasing The Print Shop 100…