Backwards Compatibility in Hindsight

Matt Barton's picture

Backwards compatibility is a complicated, multi-faceted problem. The issues are different for players, developers, and engineers. Each has his or her own reasons to wish to extend or shorten the longevity of software designed for the previous generation’s hardware. Ultimately, though, backwards compatibility is a bad short-term solution to a big long-term problem. It’s hedging the bet on a new platform—and it lowers the stakes and thus the potential winnings offered by that platform. While there are certainly some situations where backwards compatibility is arguably very necessary, it quickly becomes a self-defeating activity. A nice, clean break with the last generation’s hardware and system software improves the odds that the new platform really will be something special. Of course, it could also end up six months later as the most expensive doorstop you’ve ever owned, but in the long term, it’s worth the risk—anything else stifles progress and limits the horizon for future gaming. If we want to move beyond present technology, we’ve got to be willing to take those risks.

Commodore 128: I can do whatever your C-64 does, but I'm more expensive.Commodore 128: I can do whatever your C-64 does, but I'm more expensive.Have you ever owned, or even seen, a Commodore 128 (C-128)? If so, I’m pleasantly surprised. Although Commodore’s 64 (C-64) computer is widely credited as the best-selling personal computer of all time, its follow-up failed to make an impression on the market. Why not? It had a great deal going for it, including the fact that it was “three computers in one.” By entering simple commands, users could switch between Commodore 64, CP/M, and Commodore 128 modes. It also had twice the memory and an 80-column display option.

While there are many possible explanations for the C-128’s failure (it was basically a “tweener” system in a market that didn’t demand one), I like to think of it as an example of what happens when a new platform is so backwards compatible with an existing system that everyone simply ignores its new, platform-specific features. Indeed, many people who owned the 128 never spent more than a few moments exploring these features; they typed in the cheer “GO 64” and hit return, which switched the computer into its near-perfect backwards compatibility mode with the far more popular (and nearly ubiquitous) Commodore 64. Even though the Commodore 128 had some impressive and intriguing features with definite potential, it never saw more than a few dozen games (see this list). A great “what if” to ponder is what would have happened if the 128 hadn’t been at all compatible with the 64? What if it had been a completely new design that paid zero homage to its predecessor? For that matter, how would Commodore’s unreleased C-65 (C-64DX), a suped-up 8-bit with limited C-64 compatibility, have fared on the market?

Unfortunately, there’s no way we will ever know. Perhaps a 128 that was incompatible with the 64 would have been an even bigger disaster, like Commodore’s Plus/4 or C-16 computers. But one thing’s for sure—the 128’s excellent backwards compatibility with the 64 became its Achilles heel. Game developers knew better than to develop games for the 128 when the C-64 was such a booming market. Gamers with C-128’s may have desired more games that took better advantage of their new hardware, but they were ultimately satisfied to play new games in C-64 mode. Though there were certainly other factors involved in Commodore’s failure with the C-128, “GO 64” was a great feature with a bad future.

Backwards compatibility isn’t a pretty term. We use the word “backwards” to describe something that’s not only obsolete, but obviously and painfully obsolete. Generally, it’s applied to people who just aren’t “with it,” who haven’t caught up to the times—the senile old man who stands out in his yard every morning waving to passerbyers on the interstate, or the folks who won’t fly because they weren’t born with God-given wings. In short, people we call “backwards” are not the type we want to be “compatible with.” If we did, we’d find a more flattering way to describe them.

This article is a brief tour through the myriad of issues surrounding backwards compatibility and an appeal to gamers and game developers to consider its implications when thinking about new platforms. We need to send a clear signal to the big companies that we’re willing to compromise our demand for backwards compatibility for the sake of fostering innovation.

Gamers, Developers, and Engineers

If you hear the term “backwards compatibility” in the company of fellow gamers, the context is probably that Joe has a huge and very expensive collection of games for his current game console, and he’s not sure what to do now that it’s about to become “obsolete.” The new console is sure to have games with better graphics and speed, but there won’t be many to choose from at first—and, if the console isn’t a hit on the market, there might not be many games for it ever. A typical console launches with only a dozen or so titles, and many of those might be crapshoots by opportunistic developers hoping to make fast cash. In short, if Joe puts his current console up on eBay to defray the costs of the new one—an all too common practice—he’s taking a risk: He may end up with a great console, but without good games, it’s a worthless hunk of metal and plastic.

Wouldn’t it be great for Joe if his new system could play all his old games on the new system? If the new console were “backwards compatible,” then suddenly the risks would seem less dramatic. Even if the new system never saw more than a few good titles, at least he could fall back on his old collection, and there’s a possibility that those older games would run better or smoother on the new system. Furthermore, as the new console gained in popularity, he’d be able to pick up plenty of games for the older system at bargain bin rates.

But then Joe’s friend Vince points out a few obvious problems with this plan. First off, why would anyone want to play old games in a new console? After all, those older games wouldn’t take advantage of all the enhancements that supposedly justify buying a new console in the first place. If he’s content with what he’s got, why not just stick with it? Secondly, if he sold his old games along with his old console, he could afford more of those $60 launch titles. Besides, it’s precisely people like Joe—who won’t “let go” of their old games—that may very well keep the new system from achieving success in the market.

The discussion doesn’t end there. Carla is a game developer, and her issues are more serious than Joe’s. As a game developer trying to earn a buck in a hotly competitive market, her obvious goal is to sell as many games as possible. That means reaching out to a wide a target audience as she can. Should she develop for the older, well-established platform, or the new, more exciting platform? The advantage of the former is that there are plenty of console owners who might be willing to buy her game. However, all the media’s attention is going to be on games for the new platform, and the people willing to spring for the cost of the console probably also have the desire to acquire games for it—any games, no matter how shoddy, as long as they’re designed for their system. Still, there is a risk—what if the hardware just doesn’t catch on? Throughout gaming history, there have been plenty of new platforms that fizzled out as rapidly as they were hyped up. The safest bet for a game developer is to target the platform with the largest base. Ports, or “translated” versions for incompatible systems, are mostly an afterthought.

A clear example of this dilemma is represented by the poor state of gaming on Apple’s powerful Macintosh platform. Although the Macintosh has historically been and continues to be a potential powerhouse for gaming, computer game developers have stuck diligently to Windows, only porting over their biggest hits—and then often a year or more after the Windows release. Of course, now that Apple has released its Boot Camp software, which allows the new Intel-based Macintosh computers to install Windows on a dual partition, this situation may change. Or will it? After all, Macintosh owners will still be playing games intended for the Windows platform; Boot Camp only extends the Windows game market into new realms of profit. We could perhaps draw an analogy here between Boot Camp-installed Macintoshes with the aforementioned Commodore 128. Perhaps Apple’s “boot camp” will become its “GO 64”—Mac users will get so accustomed to running Windows on their machines that they will begin to neglect Mac-only features and software.

Apple Boot Camp: Install at Apple's peril.Apple Boot Camp: Install at Apple's peril.Let’s bring one more character into our discussion: Rodney the engineer. For Rodney, backwards compatibility is an albatross around his neck. He wants to look forward, not backward, when designing the hardware and system software for the new console. Ensuring that the new system can properly emulate or execute old code is a hindrance and a nuisance. It’d be much better for him if he could start with a clean slate, designing only for optimum performance. Michael Brundage, a software design engineer for the Xbox 360, says that maintaining backwards compatibility with the original Xbox “will be the hardest technical challenge of my career -- I can't imagine what could possibly top it in terms of sheer technical difficulty” (see Xbox Backwards Compatibility). Another developer named Itai Shirav calls backwards compatibility a “pain in the ass”:

In a perfect world (at least from a developer's point of view), every version of the software is completely free from the shackles of the previous versions. It can be deployed without any regard to the established data, APIs and protocols. (see Preparing for Backwards Compatibility)

Ultimately, what’s bad for the engineers is bad for game developers and gamers. Innovations in hardware don’t always lead to innovations in gaming, but they sometimes do. Innovations like the motion-sensing controller for Nintendo’s Wii, for instance, or the split-screen interface of the Nintendo DS, do challenge game developers to find ways to integrate these assets into their games. After all, Nintendo DS owners aren’t likely to be pleased with a game that makes no use of the dual screen feature. On the other hand, games that rely too heavily on those dual screens will be difficult to port to Nintendo’s Gameboy Advance (GBA), which has only one screen—but a well-established user base.

A Brief History of Backwards Compatibility

Backwards compatibility has been an issue since the dawn of computing. One of the earliest examples occurred in 1964 when IBM released a new family of mainframe machines, the System/360 (yes, another 360!). Since IBM feared alienating previous users of its 1400 and 7094 computers, it ensured that its new line was compatible with these older, less powerful machines. On the early 360 models, this could only be accomplished by flipping a switch and restarting the machine. The Wikipedia entry for the System/360 notes that one aspect of this backwards compatibility was IBM’s adoption of EBCDIC over ASCII, mostly for the sake of preserving compatibility with punch cards! Maintaining this backwards compatibility turned out to be a substantial problem (and, arguably a triumph) for IBM—as evidenced by this technical document about the system. According to the authors, achieving “intermodel” compatibility required a “new concept and mode of thought to make the compatibility objective even conceivable.” It was a concept that cost IBM over $5 billion to implement. IBM had bet the farm.

Fortunately for IBM, the System/360 is a success story in the history of commercial mainframes. It quickly become a hot seller and established precedents that are still being followed today. Interestingly, one of the reasons for its success might be due to backwards compatibility, but not in the sense that folks buying or leasing these machines really didn’t care to tap their true potential. Rather, IBM’s decision to maintain backwards compatibility led to one of the most important developments in computer engineering—the separation of architecture from implementation. Essentially what this boiled down to is the creation of computing standards that would ease the transition between generations of hardware, allowing for “families” of computers rather than autonomous, totally independent platforms.

The early history of personal computers and game consoles is essentially a tale of compatibility and incompatibility. Indeed, what we now call the “PC” was once always labeled IBM Compatible, one of a million “clones” of IBM’s personal computer running Microsoft’s DOS. Folks new to computers might be surprised to learn about the myriad of incompatible personal computers so prominent in the 1980s and 90s. Software designed for one of these computers wouldn’t run on competitor’s computers, or even on other computers from the same manufacturer! For instance, software designed for the Commodore Plus/4 computer was useless on Commodore’s Vic-20, which was also incompatible with its C-64. Software for all three of these machines was useless on an Apple, Atari, or Texas Instruments computer, all of which were common in those days.
ColecoVision Adapter: Want to play Atari 2600 games on your ColecoVision?ColecoVision Adapter: Want to play Atari 2600 games on your ColecoVision?

The 16-bit era of personal computing saw much of the same. Commodore never made any real effort towards backwards compatibility between its Amiga computers and the venerable C-64 (though imperfect solutions did exist). As everyone familiar with the Amiga line is painfully aware, the incompatibilities between the Amiga 500, 1000, 2000, and 3000 can be both subtle and show-stopping, as is the variances between releases of the Amiga operating system (1.0, 1.3, 2.x, and 3.x). The incompatibility issues are particularly prevalent in software that tries to “hack” the unique hardware of each computer or operating system, such as some graphics routines and copy protection schemes. When Commodore released its AGA line (which included the Amiga 1200 and 4000 as well as the CD32 game console), it had lost the marketing momentum necessary to secure sufficient game development. Although the AGA chipset in these newer machines offered significant advances in graphics, the lack of a CD-ROM drive on the computers (yet the dependence on one by the game console) confused consumers and divided loyalties among the Amiga community. In short, though the AGA Amiga family was mostly backwards compatible with the older machines, there are too many other factors involved to leave room to speculate about its effect. If Commodore had foregone backwards compatibility completely with the AGA lineup, it’s doubtful they would have sold more than a few dozen units. To put it bluntly, Commodore had made so many mistakes by 1993 that disaster was inevitable, backwards compatibility or not (Amiga 600, anyone?)

Backwards compatibility is a marketing term, nothing more.

Matt Paprocki

Console history is a bit less cut and dry in terms of backwards and what we might term “lateral” compatibility. One of the more dramatic events was Coleco’s announcement that with the purchase of a simple add-on, its powerful ColecoVision system could run any of the games designed for the massively popular Atari’s 2600. This fact gave the ColecoVision quite an edge in terms of software, and despite consistent legal attempts, Atari was never able to eliminate the competition. However, Coleco’s success with unauthorized compatibility seems to be an isolated occurrence. Nintendo’s NES, SNES, 64, and GameCube were not only incompatible with each other, but vigilantly protected from the type of competition Coleco had brought to bear on its rival.

As far as authorized backwards compatibility goes, one oft-cited example is Atari’s 7800, released in 1986. This unit ostensibly sported better graphics and was compatible with the 2600. Unfortunately, a series of unrelated setbacks (including issues with the sound) prevented the 7800 from really challenging Nintendo’s juggernaut, so it’s hard to determine the effect that backwards compatibility had on its sales. Sega’s Mega Drive (more popularly known as the Genesis in the US) did feature backwards compatibility with the older Sega Master System (SMS), albeit via an adapter called the “Power Base Converter” that compensated for the difference in cartridge shapes. Furthermore, Sega’s controller inputs for its systems were the same as those on the Atari and Commodore computers, so that controllers for all systems were in some degree interoperable. This was a decisively different strategy than Nintendo’s, which always favored proprietary inputs for its controllers. For Nintendo, if something could be made proprietary, it was. Sega’s more open-standards approach may have helped it stay in the ring as long as it did, but, again, it’s difficult to determine to what extent it mattered. Sega’s Master System was never as successful in the US as its later Genesis system, so it’s doubtful that most Genesis owners really cared about these older games—except, perhaps, insofar as they expanded the library of available games during the launch of the Genesis. Certainly, Sega’s Dreamcast wasn’t directly compatible with the earlier systems (probably owing to the move from cartridge to disc storage format).

One of the best known contemporary examples, of course, is Sony’s PlayStation 2 (PS2), which is directly backwards-compatible with the earlier PlayStation. Sony’s original console was a runaway success, and when the PS2 arrived in 2000, five years after the debut of the PlayStation, there were already thousands of titles available for it—and plenty of classics like Tomb Raider and Silent Hill. Indeed, I continue to find Sony’s original PlayStation for sale in Target and Wal-Mart, still priced at $50 or more. Undoubtedly, the PS2’s backwards compatibility helped ameliorate the problem posed by its unimpressive launch titles. Although the PS2 would soon have plenty of outstanding games for its system, it’s still possible today to find games for the PS1 available at major stores, though the market for new PS1 games has declined to a trickle.

Certainly, one of the reasons that I was plenty eager to stick with the PlayStation 2 over the GameCube and Xbox was that I could keep playing my two dozen PSOne games on the new system.

Curmudgeon Gamer

Of course, the big question here is what effect the PS2’s compatibility with the PS1 had on the success of the console. Clearly, the PS2 hasn’t become another Commodore 128. Perhaps the key here is that the PS2 had enough truly worthwhile advancements to make clinging to the old system undesirable. Likewise, developers seemed keener to produce games for the PS2 than they did for the C-128—obviously, they sensed a greatly expanding user base and wanted to leap on the bandwagon before it was too late. Sony, unlike Commodore, did an excellent job of marketing its new system and cajoling major developers into exclusive deals for big titles.

However, Microsoft launched its eminently successful Xbox console a year later, which represented a “clean break” from previous systems. Incredibly successful titles like Halo no doubt helped Microsoft establish its presence in a tight market, but I can’t help but wonder if the lack of interoperability made the Xbox seem somehow more revolutionary than the PS2.

Perhaps in the next few months we’ll see the pattern repeat itself. Microsoft opted for backwards compatibility with its new 360 console and Sony plans the same for its PS3. Nintendo has announced that its Wii system will be backwards compatible with its GameCube—a somewhat unprecedented move given Nintendo’s history of “clean breaks” between consoles (though this hasn’t been observed in its portables). Of course, it is unlikely that any two systems will ever truly be 100% compatible; subtle but pesky problems and bugs are the bane of such operations. In other words, if you really cherish your current lineup of software, you’d do well to buy a few surplus last-generation systems and hold on to them. There is also no telling how many compromises were made in order for these machines to retain backwards compatibility. All of that looking backward may have prevented them from moving very far forward.

Concluding Thoughts

Unfortunately, we seem to have no reliable data to indicate whether backwards compatibility really makes a difference on the success of a new platform. The Sega Genesis and the PS2 featured backwards compatibility with earlier platforms and were both successful on the market. Meanwhile, Nintendo initially pursued and later abandoned NES backwards compatibility in its SNES, which may have resulted in the SNES’ one true weakness—its slow processor. Indeed, there’s little doubt in my mind that doggedly pursuing a compatibility strategy would have made little difference either way, and may have actually hurt sales if it meant a less efficient system. Meanwhile, it seems silly to suggest that Sega’s failure with the Dreamcast or Commodore’s with the Amiga can be traced back to lack of backwards compatibility with their previous generation.

Every new platform has to start from scratch. All the folks who yelped for backwards compatibility in Microsoft’s 360 seemed to have had no problem with the original Xbox’s total lack of such. If I were one of Sony’s CEOs who had decide whether to insist on backwards compatibility for the PS3, I would have asked myself one all-important question: Is the new system really a technological achievement over the old one? If the answer to that question was yes, then I’d urge against looking backwards. It’s not worth compromising the new system for the sake of the old. If the new system is a breakthrough, it’s sure to be a success with proper marketing and developer wooing. However, if the new system was truly only an “upgrade” or small step forward, I’d fight heavily for backwards compatibility. If all else failed, and the new system was a disappointment, perhaps sales of the old system (and software for it) might help keep things going until another attempt can be made when the next generation arrives.

In other words, when a company starts touting “backwards compatibility” heavily in its marketing, buyer beware. At best, it’s a bullet-point on a box that you’ll probably forget about in a few months after purchasing the system. At worst, someone in authority has doubts about the longevity of the new platform and has insisted (at considerable cost to development) on hedging the bets. The very fact that a system is backwards compatible represents a series of compromises, some of which we may never know about. It’s like going into a marriage with a pre-nuptial agreement. Sure, you have the security of knowing that, if all else fails, you’ll be keeping your house. If the marriage works, nothing lost. It also means that you lack trust in your new spouse, and that’s a damn good reason to wait until something better comes along.

However, there are other ways to view the future of backwards compatibility, and I’d like to end by mentioning “Procedural Synthesis,” a technique that amounts to engineering today’s games with tomorrow’s hardware in mind. Although I don’t have time to go into the details, I encourage you to visit About.com and learn about it. Techniques like these could really turn the tables on backwards compatibility. Why? Because the folks who really need advanced hardware (i.e., game developers), could use procedural synthesis techniques to guide the development of new hardware. They could design the games they really want to make, but allow them to be run in reduced mode to compensate for the current generation’s limitations. When the next generation hardware was being developed, the engineers would have a clearer sense of direction—they’d just need to create hardware that could run these existing games without limitations. It’d be a case of the dog wagging the tail rather than the old tradition of software development trudging along behind hardware, and companies like Microsoft and Sony attempting to woo developers with promises to better meet their needs.

Postscript
Please note that the above article was written last summer, and much has changed in the interval. The PS3 and the Wii have been released, for one thing. Still, even though the article may be a bit "dated" now, I thought I'd take it out of cold storage along with the other articles from the long-overdue Issue #8.

Comments

Matt Barton
Matt Barton's picture
Offline
Joined: 01/16/2006
Crappy PS3 backwards

Crappy PS3 backwards compatibility...

Observe:

n/a
Matt Barton
Matt Barton's picture
Offline
Joined: 01/16/2006
DARN, that mysql problem is

DARN, that mysql problem is flaring up today. I can barely post!

Anyway, let me retype my message...Basically, if you put the words quote between square brackets, you can use the quote anywhere.

Example:

Bill Loguidice wrote:

Hmm, we seem to have lost the "quote" option. I'm doing it manually.

We need to get the button showing up on pages and forum posts...Maybe we could request it in the next update.

n/a
Bill Loguidice
Bill Loguidice's picture
Offline
Joined: 12/31/1969
By the way, here's the

By the way, here's the official Xbox 360 backwards compatibility list from Microsoft. Some of my students are reporting that some games offer enhancements when played on the 360. The site notes that software emulators are required for each game, so I'm not clear on how well they run--are some games even slower on the 360 than the original Xbox?

If you have more specific info regarding this, please let me know!

Hmm, we seem to have lost the "quote" option. I'm doing it manually.

Anyway, yes, the Xbox 360 uses "software emulation", which does cause some issues with frame rate and online stability in certain games or certain instances. However, overall, games seem to run well. As for enhancements, the only enhancement is upscaling to 720p, otherwise the games aren't touched. This upscaling is nice, as if you have a widescreen TV and the game was originally 4:3, it adds the black borders for you.

======================================
Bill Loguidice, Managing Director
Armchair Arcade, Inc.
(A PC Magazine Top 100 Website)
======================================

n/a
Mark Vergeer
Mark Vergeer's picture
Offline
Joined: 01/16/2006
Bear in mind that X360

Bear in mind that X360 backwards compatibility does differ from NTSC to PAL consoles.
On PAL consoles there is a snag when wanting to use a VGA cable to have something close to HD gaming. Most compatible PAL xbox games I tried won't work with the VGA cable as their refresh rate isn't compatible with the one used by the PAL X360 and the VGA cable. I am sure Microsoft could fix this without too much work by making the games run a little faster (60Hz instead of 50Hz) as on the web there are tons of reports of people with hacked regular xbox'es that are able to run PAL and NTSC games on the same machine in either framerate regardless of origin. On the x360 all is done in software so if hackers are able to do it on the old machine why can't Microsoft fix this woe for PAL users?

Like it is now, PAL country is left in the Dark when it comes to backwards compatibility. As a whole lot of top titles seem to be PAL-50 only and refuse to function on VGA cables.

Here's a snippet from the original Microsoft website

The VGA standard cable does not support PAL-50. Therefore, if you use a VGA cable, you must configure the Xbox 360 console to use PAL-60.

However, some original Xbox games support only PAL-50 and not PAL-60. Therefore, you will not be able to play original Xbox games that support only PAL-50 if you use a VGA cable.

For more information about how to change your PAL settings, click the following article number to view the article in the Microsoft Knowledge Base:
917304 Xbox 360: Your Xbox 360 console is set to PAL 50 Hz and you cannot play an Xbox 360 game on your console

========================
Mark Vergeer - Editor / Pixelator
Armchair Arcade, Inc.
Xboxlive gametag
========================

n/a
Matt Barton
Matt Barton's picture
Offline
Joined: 01/16/2006
By the way, here's the

By the way, here's the official Xbox 360 backwards compatibility list from Microsoft. Some of my students are reporting that some games offer enhancements when played on the 360. The site notes that software emulators are required for each game, so I'm not clear on how well they run--are some games even slower on the 360 than the original Xbox?

If you have more specific info regarding this, please let me know!

n/a
forcefield58
Offline
Joined: 05/19/2006
By the way, excellent

By the way, excellent article!!!

I never had a C64 and got my C128 while overseas. I bought a bunch of games for the 128 and used it's own word processor. I stayed with it until the Amiga's came out as they promised more and better games. That lasted a few years and I was forced to go Windows, yuck!! I never did anything in C64 mode, except type in some of those small programs in DOS that were listed in some of the early gaming mags, just to see if I could "program", hahaha. Cheers

n/a
dragon57
dragon57's picture
Offline
Joined: 05/18/2006
Matt, Great post. I have a

Matt,
Great post. I have a couple of historical observations related to items you brought up.

I was a personal friend of a couple of computer store owners back when the C64 was big. When the C128 came out, the local market was pretty much saturated and they had no one to sell the C128 to. In those days people didn't see upgrading as something that was required or desirable. The sales people also hadn't started pushing 'upgrades' like the industry pushes today. Of the few C128's that were sold in the local stores, they were sold because they were backward compatible. In the end, the C128 failed in my local market because people were not buying any computer in sufficient quantity to justify staying with the CV line or any line of computers for that matter. My local market really went through an up and down market for a number of years around the time of the C128. I suspect other areas of the country did as well.

For me personally, I originally bought the ColecoVision console because it promised 2600 backward compatibility. I owned a 2600 and a huge 2600 library of games and when the CV came out I convinced myself I had to have it. The only way I could afford the CV was to sell the 2600 and a few games. Most people at that time snapped up any 2600 that was put up for sell not caring about the new 'CV' that was coming out. I sold my 2600, bought my CV and played 2600 games for a long time in addition to all the cool CV games that came out. As a matter of fact, I still have my original CV and 2600 adapter! :)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.