Support us on Patreon
PayPal
On this day, last year, Valve released Steam Machines onto the world, after the typical Valve delays. While the state of the Linux desktop regarding gaming has improved, Steam Machines have not taken off as a platform, and SteamOS remains stagnant. What happened with these projects from Valve? Why were they created, why did they fail, and what could have been done to make them succeed?
Context
In 2012, when Windows 8 released, it included an app store, much like iOS and Android. With the new touch-friendly user interface Microsoft debuted, there was a new set of APIs available called “WinRT,” for creating these immersive touch-friendly applications in the UI language called “Metro.” Applications created with this new API, however, could only be distributed via the Windows Store, with Microsoft taking out a 30% cut, just like the other stores. To Gabe Newell, CEO of Valve, this was unacceptable, and he saw the risks of Microsoft using its position to push the Windows Store and Metro applications to crush Valve, like what they had done to Netscape using Internet Explorer.
To Valve, the strength of the PC running Windows was it that was an open platform, where anyone could run whatever they want without control over the operating system or hardware vendor. The alternative to these proprietary platforms closing in on third-party application stores like Steam was to push a truly open platform that grants freedoms to change, to everyone – Linux. Linux is just a kernel, but you can easily create an operating system with it and other software like the GNU core utilities and Gnome, such as Ubuntu. While pushing Ubuntu and other Linux distributions would allow Valve a sanctuary platform in case Microsoft or Apple turned hostile, Linux gave them possibilities to create a new platform.
Conception
Valve seemed to have found an opportunity in the console space, if we can call Steam Machines consoles. To achieve the user interface expectations of a console, being used on a large screen television from afar, Big Picture Mode was created. A core principle of the machines was openness; the software was able to be swapped out for Windows, as an example, and the CAD designs for the controller are available for people’s projects.
Originally, Valve had planned to create their own box as a “flagship” machine. However, these only shipped as prototypes to testers in 2013. They would also let other OEMs like Dell create their own Steam Machines as well, and allow a variety of pricing and specification options. A company called “Xi3” showed their small box, small enough to fit into a palm, as a possible candidate to become a premiere Steam Machine, which created more hype around Steam Machines. Ultimately, Valve decided to only go with OEM partners to make and advertise Steam Machines, rather than doing it themselves.
More “out there” ideas were considered. Biometrics, gaze tracking, and motion controllers were considered for the controller. Of them, the released Steam Controller had a gyroscope, and the HTC Vive controllers had various tracking and motion features that may have been originally intended for the original controller concepts. The controller was also originally intended to be more radical in its approach, with a touchscreen in the middle that had customizable, context-sensitive actions. Ultimately, the launch controller was more conservative, but still had features like the dual trackpads and advanced software that gave it flexibility. Valve had also considered making a version of Steam Machines and SteamOS for smaller hardware like laptops. This ultimately never bore any fruit, though the “Smach Z” handheld could be compared to this.
In September 2013, Valve had announced Steam Machines and SteamOS, with an expected release in the middle of 2014. The aforementioned 300 prototype machines were released to testers in December, and in January, 2000 more machines were provided to developers. SteamOS was released for testers experienced with Linux to try out. With the feedback given, Valve had decided to delay the release until November 2015.
The late launch caused problems with partners; Dell’s Steam Machine was launched a year early running Windows as the Alienware Alpha, with extra software to improve usability with a controller.
Launch
With the launch, Valve and their OEM partners released their machines, and Valve also released the Steam Controller and the Steam Link. A retail presence was established with GameStop and other brick and mortar stores providing space. Before release, some OEMs pulled out of the launch; Origin PC and Falcon Northwest, two high-end boutique builders. They had claimed performance issues and limitations had made them decide not to ship SteamOS.
The machines had launched to mixed reviews. The Steam Link was praised and many had considered buying one for their existing PC instead of buying a Steam Machine for the living room. The Steam Controller reception was muddled, due to its rich feature set but high learning curve. The Steam Machines themselves ultimately launched to the muddiest reception, however. Reviewers like LinusTechTips noticed glaring defects with the SteamOS software, including performance issues. Many of the machines were criticized for their high price point and poor value, especially when compared to the option of building a PC from the perspective of a PC gamer, or the price in comparison to other consoles. The use of SteamOS was criticized over compatibility, bugs, and lower performance than Windows. Of the available options, the Alienware Steam Machine was considered to be the most interesting option due to its value relative to other machines and small form factor.
By using Debian Linux as the base, Valve had many “launch titles” for the platform, as they had a library of pre-existing Linux titles. The initial availability of games was seen as favourable over other consoles. However, many titles originally announced for the platform never came out, or came out much later. Rocket League and Mad Max only came out only recently after the initial announcements a year ago, and titles like The Witcher 3 and Batman: Arkham Knight never came for the platform, despite initial promises from Valve or publishers. In the case of The Witcher 3, the developer, CD Projekt Red, had denied they ever announced a port, despite their game appearing in a list of titles on sale that had or were announced to have Linux and SteamOS support. In addition, many “AAA” titles have not been ported; though this situation continues to improve over time.
Neglect
With the Steam Machines launched, developers at Valve had moved on to other projects. Of the projects being worked on, virtual reality was seen as the most important, with about a third of employees working on it as of June. Valve had seen virtual reality as something to develop, and Steam as the prime ecosystem for delivering VR. Using HTC to manufacture, they had designed their own virtual reality headset and controllers, and would continue to develop new revisions. However, Linux and Steam Machines had fallen to the wayside with this focus. SteamVR, until recently, did not support Linux (it's still not public yet, but it was shown off at SteamDevDays on Linux), which put into question Valve’s commitments to Steam Machines and Linux as an open platform with a future.
There has been little development to SteamOS itself. The last major update, SteamOS 2.0 was mostly synchronizing with upstream Debian and required a reinstallation, and continued patches simply continue synchronizing with upstream sources. While Valve has made improvements to projects like Mesa, which have improved performance for many users, it has done little with Steam Machines as a product.
Many features continue to go undeveloped. Steam’s built in functionality like chat and broadcasting continue to be weak, but this affects all platforms that Steam runs on. More pressingly, services like Netflix, Twitch, and Spotify are not integrated into the interface like most major consoles, but accessing them requires using the browser, which can be slow and clunky, if it even achieves what is wanted, or to bring in software from third-party sources, which requires using the terminal, and the software might not be very usable using a controller –this is a poor UX for what’s considered to be an appliance.
Valve put little effort into marketing the platform, preferring to leave this to OEMs. However, most OEMs were either boutique builders or makers or barebones builders. Of the OEMs, only Dell was the major player in the PC market, and the only one who pushed Steam Machines with advertisements.
Sales were not strong. With 500,000 controllers sold 7 months on (stated in June 2016), including those bundled with a Steam Machine. This puts retail Steam Machines, not counting machines people have installed SteamOS on, in the low hundred thousand mark. Compared to the existing PC and console install bases, this is low.
Post-mortem thoughts
So, with the story of what happened, can we identify why Steam Machines failed, and ways they could succeed in the future?
Vision and purpose
Steam Machines did not make clear what they were in the market, nor did any advantages particularly stand out. On the PC flank, building PCs had become popular and is a cheaper option with better upgrade and flexibility options. On the console flank, they were outflanked by consoles with low initial investment, despite a possibly higher TCO with game prices, and a far simpler user experience.
With PCs, flexibility is seen as a core asset, with users being able to use their machines beyond gaming, doing work and other tasks. While Steam Machines were just PCs running SteamOS with no restrictions, the SteamOS software and marketing had solidified their view as consoles to PC gamers, compounded by the price and lower flexibility in hardware with some options. In the living room where these machines could have made sense to PC gamers, the Steam Link offered a way to access content on a PC in another room, and small form factor hardware like NUCs and Mini-ITX motherboards allowed for custom built PCs that are more socially acceptable in living rooms. The SteamOS software was also available to “convert” these PCs into Steam Machines, but people seeking flexibility and compatibility often opted for a Linux or Windows desktop. Recent strides in Windows and desktop Linux have simplified maintenance tasks associated with desktop-experience computers, automating most of it.
With consoles, simplicity is a virtue. Even as they have expanded in their roles, with media features often a priority, they are still a plug and play experience where compatibility and experience are guaranteed, with a low barrier of entry. Consoles also have long life cycles, ranging from four to seven years, and the fixed hardware during this life cycle allow developers to target and optimize especially for their specifications and features. New mid-life upgrades like “Scorpio” and PlayStation 4 Pro may change the unified experience previously shared by users, but manufactures are requiring games to work on the original model consoles to avoid the most problematic aspects. To keep users attached to the systems, social networks and exclusive games are used. Games also come on discs that can be freely reused and resold, which is a positive for retailers and users. Steam Machines have none of these guarantees; they carry PC complexity and higher initial prices despite a living room friendly exterior.
Reconciliation
With this, Steam Machines could be seen as a “worst of both worlds” product, carrying the burdens of both kinds of product, without showing clearly as one or the other, or some kind of new product category. There also exist many deficiencies that neither party experiences, like lack of AAA titles that appear on consoles and Windows PCs, and lack of clients for services like Netflix. Despite this, Valve has shown little effort into improving the product or even trying to resolve the seemingly contradictory goals like the mutual distrust of PC and console gaming.
Some things may make it impossible to reconcile the two concepts into one category or the other, though. Things like graphics settings and mods may make it hard to create a foolproof experience, and the complexity of the underlying system appears from time to time.
One of the most complex parts is the concept of having a lineup – users need to evaluate not only the costs and specifications of a system, but its value and value relative to other systems. You need some way for the user to know that their hardware can run any given game, either by some automated benchmark system with comparison, or a grading system, though these need to be simple and need to support (almost) every game in the library. In addition, you also need to worry about how these systems and grades will age – what does a “2016 Grade A” machine mean three years from now?
Valve, effort, and organization
Valve’s organizational structure may be detrimental to creating platforms like Steam Machines, let alone maintaining services like Steam. Their mostly leaderless structure with people supposedly moving their desks to ad-hoc units working on projects that they alone decide to work on can be great for creative endeavours, as well as research and development. It’s said Valve only hires what they consider to be the “cream of the crop,” with very strict standards, tying them to what they deem more "worthy" work. This view may be inaccurate though; as cliques often exist, the word of Gabe Newell is more important than the “leaked” employee handbook lets on, and people hired and then fired as needed, as a form of contractor working on certain aspects.
However, this leaves projects that aren’t glamorous or interesting, but need persistent and often mundane work, to wither on the vine. Customer support for Valve has been a constant headache, with enraged users felt ignored, and Valve sometimes only acting when legally required to do so, like with the automated refund system that was forced into action by Australian and European legislation, or the Counter-Strike: Global Offensive item gambling site controversy involving the gambling commission of Washington State that’s still ongoing.
This has affected Steam Machines as a result. With the launch delayed by a year, some partners’ hands were forced, by Dell launching the Alienware Steam Machine a year earlier as the Alienware Alpha – causing the hardware to be outdated on launch. These delays may have also affected game availability as well. The opinions of developers and hardware partners as a result of the delayed and non-monumental launch are not clear. Valve’s platform for virtual reality simply wasn’t available on Linux, and as such, SteamOS, until recently, even as SteamVR was receiving significant developer effort.
The “long game”
Valve is seen as playing a “long game” with Steam Machines and SteamOS, though it appears as if there is no roadmap. An example of Valve aiming for the long term is with Steam, from its humble and initially reviled beginnings as a patching platform for their games to the popular distribution and social network it is today. It also helped that Steam was required to play Valve’s games like Half-Life 2 and Counter-Strike 1.6. However, it doesn’t seem as if Valve is putting in the effort to Steam Machines as they did with Steam before. There is also entrenched competition that Steam in the early days never really dealt with. Their competition includes arguably Valve itself, with Steam on Windows.
Gambit
With the lack of developments in Steam Machines, one wonders if the platform was a bargaining chip of sorts. Steam Machines had been originally started over Valve’s Linux efforts took fruit because of concerns that Microsoft and Apple would have pushed them out of the market with native app stores, and Steam Machines grew so Valve would have a safe haven in case this happened, and a bargaining chip so Valve can remind the developers of its host platforms of possible independence. When these turned out to be non-threatening, Valve slowed down development. I don’t see this however; Valve has expended a lot of goodwill with hardware partners and developers trying to push this, only to halt it. You could say both Microsoft and Valve called each other’s bluffs – Microsoft with a locked-down Windows 8, and Valve’s capability as an independent player.
Even then, who is to say developers wouldn’t follow Microsoft with a locked-in platform, if they can offer superior deals to publishers, or better customer relationships? In addition, now Microsoft is pushing Xbox on Windows integration with cross-buy, Xbox Live integration, and Xbox exclusive games on Windows, all while preserving Windows as an open platform – arguably more a threat to Steam.
Another point you could argue is that all of this with Steam Machines was simply to push Linux adoption with PC gaming, and Steam Machines were simply to make it more palatable to publishers and developers by implying a large push and continued support. However, this made it an awfully expensive gambit, and developers continued to support Linux before and after Steam Machines, and could have backfired with developers pulling out of Linux due to the lack of the Promised Land of SteamOS coming.
My opinions on what could have been done
I think there’s an interesting product with Steam Machines, and that there is a market for it, but lack of interest and effort, as well as possible confusion in what it should have been has been damaging for it. I see Steam Machines as a way to cut out the complexity of PC gaming of worrying about parts, life cycles, and maintenance; while giving the advantages like cheap games, mods, and an open platform that can be fiddled with if the user desires. However, they need to get core aspects like pricing, marketing, lineup, and software right.
I think Steam Machines can make compromises on things like upgradability (Though it’s possible to preserve this – but it should be done with attention to user experience.) and choices, to reduce friction. PCs would still exist to these options. The paralysis of choice is a real dilemma, and the sheer amount of poorly valued options available with Steam Machines didn't help. Valve needs a flagship machine to lead Steam Machines. Arguably, the Alienware model was close, but it wasn’t made officially so. There’s good industrial design talent in Valve, and if they focused on their own machine, and with effort put in, it might be worth it. A company like Dell or HTC can manufacture for Valve, bringing their experience in. Defining life cycles and only having one or two specifications updated periodically would help, especially if they worked with developers to establish this is a baseline that should be supported. I’m not sure with OEMs; if Valve is putting their effort behind one machine, they might be made redundant and ultimately only hindering development of the platform.
Addressing the software issues is essential. The lack of integration with services like Netflix and Twitch that exist fluidly on console and easily put into place on PC, despite living room user interface issues, are holding Steam Machines back. Although Valve has slowly been acquiring movie licenses for distribution on Steam, people will use existing and trusted streaming sources. This needs to be addressed, especially as people use their consoles as parts of their home theatre. Fixing issues with the Steam client and platform are essential, and feature parity with other platforms is a good idea. Performance issues with Linux and its graphics stack are also a problem, but this is slowly improving. Getting ports of games will also be another issue. Game porting shops like Feral Interactive and Aspyr Media help the library, but they need to be contracted by publishers and developers, and they often use wrappers that add overhead. Valve has helped studios directly with porting, such as with Rocket League, but this has rarely happened and when it did, slowly at the typical Valve pace. The monolith of AAA games can’t be ignored either – the situation has improved dramatically, but studios like Bethesda are still reluctant to port, especially with a small user base, lack of support from Valve with Steam Machines even if Linux is doing relatively well, and the lack of extra DRM like Denuvo.
Valve also needs to put effort into the other bits beyond hardware and software. With one machine, they have an interest and can subsidize the hardware effectively. This would put it into parity with consoles, and possibly cheaper than custom built PCs. Efforts to marketing the product to market segments that would be interested in the machines are essential, whatever they are. (I myself would be interested in the machines. I don’t like the hassle of dealing with PC building or the premium on prebuilt machines, but consoles often lack the games I want to play, and I have an existing library of games on Steam I acquired cheaply.) Retail partners may not be effective, due to their interest in selling and reselling physical copies of games.
Even with my suggestions towards the platform and product, I’m not sure how effective it would be to help Steam Machines achieve their full potential and do well in the marketplace. Ultimately, learning from not just your own mistakes, but the mistakes of previous entrants like 3DO and Pippin who relied on an open platform or were descended from desktop-experience computing, which are relevant to Valve’s current situation, and the future of Nintendo's Switch, which steps into the realm of possible confusion between values.
Note: Clearing up done by liamdawe, all thoughts are from the submitter. Article taken from GamingOnLinux.com.
Context
In 2012, when Windows 8 released, it included an app store, much like iOS and Android. With the new touch-friendly user interface Microsoft debuted, there was a new set of APIs available called “WinRT,” for creating these immersive touch-friendly applications in the UI language called “Metro.” Applications created with this new API, however, could only be distributed via the Windows Store, with Microsoft taking out a 30% cut, just like the other stores. To Gabe Newell, CEO of Valve, this was unacceptable, and he saw the risks of Microsoft using its position to push the Windows Store and Metro applications to crush Valve, like what they had done to Netscape using Internet Explorer.
To Valve, the strength of the PC running Windows was it that was an open platform, where anyone could run whatever they want without control over the operating system or hardware vendor. The alternative to these proprietary platforms closing in on third-party application stores like Steam was to push a truly open platform that grants freedoms to change, to everyone – Linux. Linux is just a kernel, but you can easily create an operating system with it and other software like the GNU core utilities and Gnome, such as Ubuntu. While pushing Ubuntu and other Linux distributions would allow Valve a sanctuary platform in case Microsoft or Apple turned hostile, Linux gave them possibilities to create a new platform.
Conception
Valve seemed to have found an opportunity in the console space, if we can call Steam Machines consoles. To achieve the user interface expectations of a console, being used on a large screen television from afar, Big Picture Mode was created. A core principle of the machines was openness; the software was able to be swapped out for Windows, as an example, and the CAD designs for the controller are available for people’s projects.
Originally, Valve had planned to create their own box as a “flagship” machine. However, these only shipped as prototypes to testers in 2013. They would also let other OEMs like Dell create their own Steam Machines as well, and allow a variety of pricing and specification options. A company called “Xi3” showed their small box, small enough to fit into a palm, as a possible candidate to become a premiere Steam Machine, which created more hype around Steam Machines. Ultimately, Valve decided to only go with OEM partners to make and advertise Steam Machines, rather than doing it themselves.
More “out there” ideas were considered. Biometrics, gaze tracking, and motion controllers were considered for the controller. Of them, the released Steam Controller had a gyroscope, and the HTC Vive controllers had various tracking and motion features that may have been originally intended for the original controller concepts. The controller was also originally intended to be more radical in its approach, with a touchscreen in the middle that had customizable, context-sensitive actions. Ultimately, the launch controller was more conservative, but still had features like the dual trackpads and advanced software that gave it flexibility. Valve had also considered making a version of Steam Machines and SteamOS for smaller hardware like laptops. This ultimately never bore any fruit, though the “Smach Z” handheld could be compared to this.
In September 2013, Valve had announced Steam Machines and SteamOS, with an expected release in the middle of 2014. The aforementioned 300 prototype machines were released to testers in December, and in January, 2000 more machines were provided to developers. SteamOS was released for testers experienced with Linux to try out. With the feedback given, Valve had decided to delay the release until November 2015.
The late launch caused problems with partners; Dell’s Steam Machine was launched a year early running Windows as the Alienware Alpha, with extra software to improve usability with a controller.
Launch
With the launch, Valve and their OEM partners released their machines, and Valve also released the Steam Controller and the Steam Link. A retail presence was established with GameStop and other brick and mortar stores providing space. Before release, some OEMs pulled out of the launch; Origin PC and Falcon Northwest, two high-end boutique builders. They had claimed performance issues and limitations had made them decide not to ship SteamOS.
The machines had launched to mixed reviews. The Steam Link was praised and many had considered buying one for their existing PC instead of buying a Steam Machine for the living room. The Steam Controller reception was muddled, due to its rich feature set but high learning curve. The Steam Machines themselves ultimately launched to the muddiest reception, however. Reviewers like LinusTechTips noticed glaring defects with the SteamOS software, including performance issues. Many of the machines were criticized for their high price point and poor value, especially when compared to the option of building a PC from the perspective of a PC gamer, or the price in comparison to other consoles. The use of SteamOS was criticized over compatibility, bugs, and lower performance than Windows. Of the available options, the Alienware Steam Machine was considered to be the most interesting option due to its value relative to other machines and small form factor.
By using Debian Linux as the base, Valve had many “launch titles” for the platform, as they had a library of pre-existing Linux titles. The initial availability of games was seen as favourable over other consoles. However, many titles originally announced for the platform never came out, or came out much later. Rocket League and Mad Max only came out only recently after the initial announcements a year ago, and titles like The Witcher 3 and Batman: Arkham Knight never came for the platform, despite initial promises from Valve or publishers. In the case of The Witcher 3, the developer, CD Projekt Red, had denied they ever announced a port, despite their game appearing in a list of titles on sale that had or were announced to have Linux and SteamOS support. In addition, many “AAA” titles have not been ported; though this situation continues to improve over time.
Neglect
With the Steam Machines launched, developers at Valve had moved on to other projects. Of the projects being worked on, virtual reality was seen as the most important, with about a third of employees working on it as of June. Valve had seen virtual reality as something to develop, and Steam as the prime ecosystem for delivering VR. Using HTC to manufacture, they had designed their own virtual reality headset and controllers, and would continue to develop new revisions. However, Linux and Steam Machines had fallen to the wayside with this focus. SteamVR, until recently, did not support Linux (it's still not public yet, but it was shown off at SteamDevDays on Linux), which put into question Valve’s commitments to Steam Machines and Linux as an open platform with a future.
There has been little development to SteamOS itself. The last major update, SteamOS 2.0 was mostly synchronizing with upstream Debian and required a reinstallation, and continued patches simply continue synchronizing with upstream sources. While Valve has made improvements to projects like Mesa, which have improved performance for many users, it has done little with Steam Machines as a product.
Many features continue to go undeveloped. Steam’s built in functionality like chat and broadcasting continue to be weak, but this affects all platforms that Steam runs on. More pressingly, services like Netflix, Twitch, and Spotify are not integrated into the interface like most major consoles, but accessing them requires using the browser, which can be slow and clunky, if it even achieves what is wanted, or to bring in software from third-party sources, which requires using the terminal, and the software might not be very usable using a controller –this is a poor UX for what’s considered to be an appliance.
Valve put little effort into marketing the platform, preferring to leave this to OEMs. However, most OEMs were either boutique builders or makers or barebones builders. Of the OEMs, only Dell was the major player in the PC market, and the only one who pushed Steam Machines with advertisements.
Sales were not strong. With 500,000 controllers sold 7 months on (stated in June 2016), including those bundled with a Steam Machine. This puts retail Steam Machines, not counting machines people have installed SteamOS on, in the low hundred thousand mark. Compared to the existing PC and console install bases, this is low.
Post-mortem thoughts
So, with the story of what happened, can we identify why Steam Machines failed, and ways they could succeed in the future?
Vision and purpose
Steam Machines did not make clear what they were in the market, nor did any advantages particularly stand out. On the PC flank, building PCs had become popular and is a cheaper option with better upgrade and flexibility options. On the console flank, they were outflanked by consoles with low initial investment, despite a possibly higher TCO with game prices, and a far simpler user experience.
With PCs, flexibility is seen as a core asset, with users being able to use their machines beyond gaming, doing work and other tasks. While Steam Machines were just PCs running SteamOS with no restrictions, the SteamOS software and marketing had solidified their view as consoles to PC gamers, compounded by the price and lower flexibility in hardware with some options. In the living room where these machines could have made sense to PC gamers, the Steam Link offered a way to access content on a PC in another room, and small form factor hardware like NUCs and Mini-ITX motherboards allowed for custom built PCs that are more socially acceptable in living rooms. The SteamOS software was also available to “convert” these PCs into Steam Machines, but people seeking flexibility and compatibility often opted for a Linux or Windows desktop. Recent strides in Windows and desktop Linux have simplified maintenance tasks associated with desktop-experience computers, automating most of it.
With consoles, simplicity is a virtue. Even as they have expanded in their roles, with media features often a priority, they are still a plug and play experience where compatibility and experience are guaranteed, with a low barrier of entry. Consoles also have long life cycles, ranging from four to seven years, and the fixed hardware during this life cycle allow developers to target and optimize especially for their specifications and features. New mid-life upgrades like “Scorpio” and PlayStation 4 Pro may change the unified experience previously shared by users, but manufactures are requiring games to work on the original model consoles to avoid the most problematic aspects. To keep users attached to the systems, social networks and exclusive games are used. Games also come on discs that can be freely reused and resold, which is a positive for retailers and users. Steam Machines have none of these guarantees; they carry PC complexity and higher initial prices despite a living room friendly exterior.
Reconciliation
With this, Steam Machines could be seen as a “worst of both worlds” product, carrying the burdens of both kinds of product, without showing clearly as one or the other, or some kind of new product category. There also exist many deficiencies that neither party experiences, like lack of AAA titles that appear on consoles and Windows PCs, and lack of clients for services like Netflix. Despite this, Valve has shown little effort into improving the product or even trying to resolve the seemingly contradictory goals like the mutual distrust of PC and console gaming.
Some things may make it impossible to reconcile the two concepts into one category or the other, though. Things like graphics settings and mods may make it hard to create a foolproof experience, and the complexity of the underlying system appears from time to time.
One of the most complex parts is the concept of having a lineup – users need to evaluate not only the costs and specifications of a system, but its value and value relative to other systems. You need some way for the user to know that their hardware can run any given game, either by some automated benchmark system with comparison, or a grading system, though these need to be simple and need to support (almost) every game in the library. In addition, you also need to worry about how these systems and grades will age – what does a “2016 Grade A” machine mean three years from now?
Valve, effort, and organization
Valve’s organizational structure may be detrimental to creating platforms like Steam Machines, let alone maintaining services like Steam. Their mostly leaderless structure with people supposedly moving their desks to ad-hoc units working on projects that they alone decide to work on can be great for creative endeavours, as well as research and development. It’s said Valve only hires what they consider to be the “cream of the crop,” with very strict standards, tying them to what they deem more "worthy" work. This view may be inaccurate though; as cliques often exist, the word of Gabe Newell is more important than the “leaked” employee handbook lets on, and people hired and then fired as needed, as a form of contractor working on certain aspects.
However, this leaves projects that aren’t glamorous or interesting, but need persistent and often mundane work, to wither on the vine. Customer support for Valve has been a constant headache, with enraged users felt ignored, and Valve sometimes only acting when legally required to do so, like with the automated refund system that was forced into action by Australian and European legislation, or the Counter-Strike: Global Offensive item gambling site controversy involving the gambling commission of Washington State that’s still ongoing.
This has affected Steam Machines as a result. With the launch delayed by a year, some partners’ hands were forced, by Dell launching the Alienware Steam Machine a year earlier as the Alienware Alpha – causing the hardware to be outdated on launch. These delays may have also affected game availability as well. The opinions of developers and hardware partners as a result of the delayed and non-monumental launch are not clear. Valve’s platform for virtual reality simply wasn’t available on Linux, and as such, SteamOS, until recently, even as SteamVR was receiving significant developer effort.
The “long game”
Valve is seen as playing a “long game” with Steam Machines and SteamOS, though it appears as if there is no roadmap. An example of Valve aiming for the long term is with Steam, from its humble and initially reviled beginnings as a patching platform for their games to the popular distribution and social network it is today. It also helped that Steam was required to play Valve’s games like Half-Life 2 and Counter-Strike 1.6. However, it doesn’t seem as if Valve is putting in the effort to Steam Machines as they did with Steam before. There is also entrenched competition that Steam in the early days never really dealt with. Their competition includes arguably Valve itself, with Steam on Windows.
Gambit
With the lack of developments in Steam Machines, one wonders if the platform was a bargaining chip of sorts. Steam Machines had been originally started over Valve’s Linux efforts took fruit because of concerns that Microsoft and Apple would have pushed them out of the market with native app stores, and Steam Machines grew so Valve would have a safe haven in case this happened, and a bargaining chip so Valve can remind the developers of its host platforms of possible independence. When these turned out to be non-threatening, Valve slowed down development. I don’t see this however; Valve has expended a lot of goodwill with hardware partners and developers trying to push this, only to halt it. You could say both Microsoft and Valve called each other’s bluffs – Microsoft with a locked-down Windows 8, and Valve’s capability as an independent player.
Even then, who is to say developers wouldn’t follow Microsoft with a locked-in platform, if they can offer superior deals to publishers, or better customer relationships? In addition, now Microsoft is pushing Xbox on Windows integration with cross-buy, Xbox Live integration, and Xbox exclusive games on Windows, all while preserving Windows as an open platform – arguably more a threat to Steam.
Another point you could argue is that all of this with Steam Machines was simply to push Linux adoption with PC gaming, and Steam Machines were simply to make it more palatable to publishers and developers by implying a large push and continued support. However, this made it an awfully expensive gambit, and developers continued to support Linux before and after Steam Machines, and could have backfired with developers pulling out of Linux due to the lack of the Promised Land of SteamOS coming.
My opinions on what could have been done
I think there’s an interesting product with Steam Machines, and that there is a market for it, but lack of interest and effort, as well as possible confusion in what it should have been has been damaging for it. I see Steam Machines as a way to cut out the complexity of PC gaming of worrying about parts, life cycles, and maintenance; while giving the advantages like cheap games, mods, and an open platform that can be fiddled with if the user desires. However, they need to get core aspects like pricing, marketing, lineup, and software right.
I think Steam Machines can make compromises on things like upgradability (Though it’s possible to preserve this – but it should be done with attention to user experience.) and choices, to reduce friction. PCs would still exist to these options. The paralysis of choice is a real dilemma, and the sheer amount of poorly valued options available with Steam Machines didn't help. Valve needs a flagship machine to lead Steam Machines. Arguably, the Alienware model was close, but it wasn’t made officially so. There’s good industrial design talent in Valve, and if they focused on their own machine, and with effort put in, it might be worth it. A company like Dell or HTC can manufacture for Valve, bringing their experience in. Defining life cycles and only having one or two specifications updated periodically would help, especially if they worked with developers to establish this is a baseline that should be supported. I’m not sure with OEMs; if Valve is putting their effort behind one machine, they might be made redundant and ultimately only hindering development of the platform.
Addressing the software issues is essential. The lack of integration with services like Netflix and Twitch that exist fluidly on console and easily put into place on PC, despite living room user interface issues, are holding Steam Machines back. Although Valve has slowly been acquiring movie licenses for distribution on Steam, people will use existing and trusted streaming sources. This needs to be addressed, especially as people use their consoles as parts of their home theatre. Fixing issues with the Steam client and platform are essential, and feature parity with other platforms is a good idea. Performance issues with Linux and its graphics stack are also a problem, but this is slowly improving. Getting ports of games will also be another issue. Game porting shops like Feral Interactive and Aspyr Media help the library, but they need to be contracted by publishers and developers, and they often use wrappers that add overhead. Valve has helped studios directly with porting, such as with Rocket League, but this has rarely happened and when it did, slowly at the typical Valve pace. The monolith of AAA games can’t be ignored either – the situation has improved dramatically, but studios like Bethesda are still reluctant to port, especially with a small user base, lack of support from Valve with Steam Machines even if Linux is doing relatively well, and the lack of extra DRM like Denuvo.
Valve also needs to put effort into the other bits beyond hardware and software. With one machine, they have an interest and can subsidize the hardware effectively. This would put it into parity with consoles, and possibly cheaper than custom built PCs. Efforts to marketing the product to market segments that would be interested in the machines are essential, whatever they are. (I myself would be interested in the machines. I don’t like the hassle of dealing with PC building or the premium on prebuilt machines, but consoles often lack the games I want to play, and I have an existing library of games on Steam I acquired cheaply.) Retail partners may not be effective, due to their interest in selling and reselling physical copies of games.
Even with my suggestions towards the platform and product, I’m not sure how effective it would be to help Steam Machines achieve their full potential and do well in the marketplace. Ultimately, learning from not just your own mistakes, but the mistakes of previous entrants like 3DO and Pippin who relied on an open platform or were descended from desktop-experience computing, which are relevant to Valve’s current situation, and the future of Nintendo's Switch, which steps into the realm of possible confusion between values.
Note: Clearing up done by liamdawe, all thoughts are from the submitter. Article taken from GamingOnLinux.com.
Some you may have missed, popular articles from the last month:
The comments on this article are closed.
All posts need to follow our rules. For users logged in: please hit the Report Flag icon on any post that breaks the rules or contains illegal / harmful content. Guest readers can email us for any issues.
choose ONE third party hardware manufacturer by "generation". then, design and create their own SteamMachine
ie:
Google Nexus One(HTC)---> VALVe SteamBox Alchemist (Dell SteamOS 1)
Google Nexus S (Samsung)---> VALVe SteamBox Brewmaster (ZOTAC SteamOS 2)
Google Nexus 4 (LG)---> VALVe SteamBox ChaosKnight (DRIFT SteamOS 3)
Google Nexus 5 (LG)---> VALVe SteamBox Doom (HP SteamOS 4)
etc..
Industry really needs one point of reference to put things in order... and that order have to make the owner of SteamOS: VALVe.
Except Google gave up on the Nexus strategy and is going for Pixel - Google designed, and HTC is only involved in manufacturing. I think Google wants to rein in the OEMs (they don't like the poor updating and OEM bloat) and has the power - save for Samsung.
Basically, Google thinks a cathedral will work better, over their bazaar.
Last edited by calvin on 10 November 2016 at 8:32 pm UTC
At least that's how I look at it. Game consoles are stagnant hardware platforms. And anymore, they literally are chunks of x86 compatible hardware with a specific OS. Virtualize that OS and you can run any of the 'Exclusives' eventually.
Personally, if I were Valve, and looking at a re-launch of Steam Machines, I would not do it this year. Maybe next year . . . right now, I would work hard on the infrastructure of drivers, Vulkan, game engines, maybe even Wayland to get definitively past X, so that next time around, SteamOS wouldn't be lagging performance-wise. Poorer performance compared to Windows was a major killer in PR terms, and really Linux should be able to win there.
And I'd feel out those multimedia issues quietly, getting things sorted out so that I had working code in-house and contracts with Netflix etc. ready to go.
When the pieces were in position, I'd go for a major launch that Christmas.
Reserving the space for my comment - gonna read now.
Time to read 3,621 words.
Last edited by ElectricPrism on 10 November 2016 at 10:12 pm UTC
Looks like they already were at the time.
:P
1. Not everyone has a smart TV. (I don't)
2. Smart TVs are often quite miserable. Buggy, laggy, and insecure. Sometimes they even inject ads.
3. You might have a Chromecast or a Roku or whatever that does Netflix or whatever, but consoles are often these boxes that provide them.
4. The experience is often better and smoother. You can even add features - the Xbox 360 used to have (though I'm unsure why they removed it) party chat integrated into the client, allowing everyone to watch the same movie.
How did this article get a pass to be featured on GamingOnLinux.com
The author clearly is some college English student who is whiny, angry, and doing his best to tear down Gaming On Linux & Valve.
He is a judger who has already written Steam Machines and SteamOS as FAILURES.
He doesn't understand how PURPOSE, AUTONOMY, and MASTERY are powerful things that Valve Infrastructure weild's to create amazing games and products.
There are numerous spelling mistakes and missing words it's like this article hasn't even gone through a spell checker, but what's worse is the author is missing a few screws too.
Also if you could keep your fucking 6 page essays to to fucking 1,200 words max this article wouldn't come off as such a assault to the eyes.
There are even no fucking images on this post. WHAT THE FUCK?
Genius Simplifies and Ignorance Complicates. This article didn't need to be fucking 3,600 words or 6 fucking pages long in my word processor - I HOPE I NEVER SEE THIS KIND OF TRASH ON GAMINGONLINUX.COM AGAIN.
Last edited by ElectricPrism on 10 November 2016 at 11:21 pm UTC
Just because you learned a few things in college doesn't make you a "know it all", "better than everyone else" smart fuck who immediately knows how to run Valve, Inc better than Valve, Inc.
Rome wasn't built in a day. If you're angry that things have not gone according to your timetable then you may want to analyze that you had inappropriate expectations which led to your disappointment.
Youngling Padawan, you have many years and days to suffer to learn why your ideas about Success, Failure and traditional workplace structure will fail - and I sincerely hope you don't learn the hard way and listen to the smarter people who have researched this shit.
View video on youtube.com
Edit:
☑ Steam & Linux are going to hit nearly 3,000 games
☑ I have more AAA games than I have time to play
☑ I play on X360, PS4, and Steam Controller
☑ my FPS on RX 480 MESA on most games is usually 80 - 300 FPS
☑ I know my games will work for decades on Linux
☑ I have no CD's to Scratch
☑ Updates to the game and bug fixes are free and digital
☑ I don't have to dedicate storage space to a DVD Game Library Bookshelf
☑ My Game Console is Upgradable and Portable because it's self built
☑ I can play my games in Vulkan, the future of cross-platform GL.
☑ Linux is destined to conquer all other systems because of it's versatility, free, and open nature
☑ Linux shares commonality with Android, PlayStation 4 and macOS making game porting between these platforms easier meaning it will eventually win.
☑ New Job listings suggest Nintendo NX is running Linux
Jesus, the list goes on and on and on and on... given some time I could give you the whole shotgun list of reasons but this list should suffice as starting reasons why Linux will conquer and succeed as a permanent installment, and eventual king of the future of gaming.
Last edited by ElectricPrism on 10 November 2016 at 11:36 pm UTC
Perhaps once Valve "proper" has their fun with the initial R&D side of things, they need to hire people outside of their standard flat organisational structure to be the caretakers of these products. Even if it means forming sub companies, sort of like how 2k has sub studios. Another "Valve" Studio could take on maintenance and future features of the project that the core Valve team has got bored of.
Ultimately: Think about Steam Machines as a product. Linux is great; this is why we're on a site talking about that. But that's not why Steam Machines would sell - the advantsges that Linux can bring can sell it, (like performance) but Linux itself will not sell machines. Linux (and other things, like Vulkan) is infrastructure for SteamOS, Arch, Ubuntu, whatever. It won't sell them; it's the things you can make with Linux that sell. As a community, we get caught up in technical minuta only to realize all that are just cogs that make the things on top that we really wanted to run. As such, their improvements benefit the whole system.
tl;dr: It's the gestalt, stupid.
On smaller things:
1. On discs - As stated, discs can be freely traded and resold. This is good for users, and retailers take advantage of this. In addition, for customers with slow internet, having the game on physical media is faster. (I don't think Steam Machines should have optical drives though; I think the advantages of digital are worth it, and so does Valve.)
2. On your comment of "desktop->Debian" - I had excised a footnote explaining the differences between Android and GNU/Linux, that "desktop Linux" implies. The rest of the stack on Android is extremely different, only the kernel is the same. (It irks me when people call Android Linux, implying it means much. RMS is right; GNU/Linux clarifies this big ambiguity. I have no idea if the Switch will be GNU/Linux or Android-like or completely different - if the stack and APIs are different, then Linux is simply a part of infrastructure that matters less than graphical and input APIs)
3. On automatic servicability - It means PCs are becoming more console-like in their simplicity. I admit it's a bit of a non-sequitur the way it is now.
4. On your performance and upgradability - Good for you. Vulkan helps (and should have helped a long time ago) but this isn't your system, is it?
5. On the number of games and AAA titles - As I stated, the situation gets better. However, people have their pet AAA (or not even AAA!) title they want. It doesn't matter how many games you have; if you don't have Fallout 4, you are dead to a large number of consumers. (And discovering and promoting these already existing games is another issue.)
6. On Linux winning with its nature - I covered this in the initial paragraph, but normal people don't care about the OS. For most, it may as well be etched on silicon. What matters if the experience you can create with Linux in your product, wether it be Steam Machines or Ubuntu.
7. On time and success - Valve has taken their time, with little to show for it, beyond what you could consider to be basically a beta launch. Things take time, but they come together in the end. It doesn't look like Steam Machines have. I'm not sure by what metric you can consider Steam Machines a success. They've barely sold and had a poor reception, with little improvements on that front. For GNU/Linux gaming, the situation has improved, but this is about Steam Machines.
ElectricPrism talks about two things: First, what really motivates people. I have for a long time found the presentation linked there very compelling, and for that matter overall I am very much in favour of flatter organizational structures. But that doesn't make the article wrong. The article does not claim that the people at Valve lack motivation or that they are not productive, or even that Valve is not broadly effective at making money (clearly it is). Nor does it claim that flat organizational structures are, overall, rubbish. What he claims is that Valve's specific flat organizational structure, combined with a certain view of the employees as "elite", seems to make Valve much better at doing, basically, new sexy creative stuff, than on following through with boring detail work or other stuff that's more unglamorous slogging. This does not contradict ElectricPrism's views on motivation, nor does it really construe flat non-hierarchical organizations generally as ineffective.
Second, ElectricPrism talks about how awesome Linux is and how well it is moving along as a gaming platform, and like that. Well, Linux is awesome and hopefully it will ultimately conquer. But none of that has anything to do with the article, which is about whether Steam Machines have succeeded. They haven't, and every indication seems to be that Valve's efforts to make them do so have pretty much petered out. Most of the article's take on this seems pretty sound to me.
So yeah, both ElectricPrism's major points aren't really wrong, they just don't represent a critique of the article. They're irrelevant.
Aside from that, I think ElectricPrism's comments, particularly the one with all the digital red crayon, are really bloody rude.
Having lived through the life of the Atari Jaguar (as short as it was) and the fact that it lives off of homebrew titles still, I'd consider that a failure in the mainstream, but people still love it. Steam Machines are still being made, more games are being released for it, and the library is getting larger every day. So really, how are they a failure? Have they sold millions of units? No because they have various versions. We'd have to count every machine that has pre-installed SteamOS to count it, and I don't see Dell, Zotac, Syber, etc releasing sales numbers yet.
Claiming the Steam Machine is successful because "games are being released for it" is akin to claiming Wine is successful because so many games are "released for it".
One can certainly come up with definitions of "failure" by which Steam Machines haven't failed. But gauged as an attempt to shake up the market for PC gaming and/or consoles, or to take serious steps towards making Linux a major and growing platform for gaming as Gabe Newell seemed from some of his statements to intend, the Steam Machine has not been a success and is not currently on course to ever become a success (although I would argue that in many ways, the Steam Controller on the other hand is a success).
Valve could fix some issues, improve their strategy, and relaunch with a harder push and potentially change all that. But they ain't done it so far and have given no indication that they have any plans to.
Last edited by Purple Library Guy on 11 November 2016 at 7:03 am UTC
ok, lets assume that your most profitable game ever made 200 millions in profit, but each new game you made, you sold less than the previous. (eg: 200, 180, 150, 110....) and you reinvest allmost everything that you made in profit to make an new game...
if some one off you 1 billion for exclusivity, would you refuse?
companies that have an exclusivity deal have nothing to lose with the deal, nothing. if they sell 0 copies they already made profit.
if they sell, even better.
companies that make consoles give more money to the game developer than the game will make in profit, why? because they will help sell the console, so the install base of the console grow and they make more money in each game sold for the platform.
you could argue that gamers are the ones who lose, but i will discuss this a bit:
what would happen if nintendo didnt make exclusives for their consoles anymore nor paid for thirdyparty exclusives? sony did the same, microsoft the same with xbox.
what would happen?
nintendo would not sell any console, nor sony neither microsoft.
all games would be done to windows, some to mac, and a few to linux.
we wouldnt have an world without exclusives because windows would still have 7000 exclusive games (only counting steam) 4000 that have an version for win and mac and the 3000 ones left that have an version for linux (mostly indies) would also be avaliable for win/mac so most of people wouldnt consider swich for linux anyway.
as you can see, even on an windows without exclusives, we would still have exclusives due to direcx or marketshare (companys that dont want to support, because the cost of support an platform is greater than the profit from it)
and different from what the stupid people who say: "pc master racers" think, if PC won, they wouldnt be able to play all their games on windows.
why? becasue microsoft would do special versions such as "windows home, premium, ultimate" where only the most expensive can play all the games (they would do exclusivity deals anyway)
or things like "buy the ultimate to unlock the full potential of your computer!" for those who want to play on ultra.
at least with consoles around, microsoft have an reason to keep improving their products instead of let then years without any major evolution while they try to dominate other markets (as they did with the internet explorer)
so an world without exclusives would never exist, if we insist in pursue this what we will got instead is an world with thousands of windows exclusives and no other option to gaming (no nintendo, no sony) if we were lucky enough to microsoft not doing what nintendo did when they were the major gaming platform: if you want to make games to our console, you can only make games for our platform.
this is the only generation that consoles are using x86.
ok, xbox original was x86, but only it.
and even in the current gen, wiiU is still powerpc.
as for perfect emulation, there is no such thing as perfect emulation:
http://arstechnica.com/gaming/2011/08/accuracy-takes-power-one-mans-3ghz-quest-to-build-a-perfect-snes-emulator/
There's no announcement that they have ended the SteamOS project. This will take time. Valve time.
SteamOS haven't even gotten out of beta yet - that's why there's no SteamMachines (with a very few experimental exceptions). It's not officially available yet - you can participate as a beta tester if you like, but it's not meant for public consumption.
One can discuss the period behind us, and by God that's been discussed a lot now, but it's not gone. It's not over. Not yet.
Because this is a blog, not a professional, editorial product. You just did the work of an editor there. And I agree, it would be real great if this site had one.
But view the site as a blog, with the pros and cons that comes with that.