Dan Mons

Interview with eXo: PC, MS-DOS and Windows game preservationist

“Preservation” is a term we hear frequently in the retro video game community, and one that is extremely important to the art form of video games.  Despite being such a young industry compared to other art forms such as novels, music or film, video games have arguably been some of the most difficult content to access past their extremely short commercial release window.

Many individuals and groups dedicated themselves to the actual act of preservation itself – the conversion of games from their native formats (cartridges, tapes, floppy disks, optical media, digital distribution etc) to simple files on disk that can be checksummed for “fixity” (a fixed state that can be verified via tools like checksums), and then hopefully distribution to the wider community or hosting on services such as The Internet Archive that are dedicated to being the modern digital world’s equivalent of our vitally important local libraries.

The loss of digital art is as dire as the loss of any other art form.  Preserving these titles often comes with some anger from modern commercial groups, but these items are a part of our culture.  While retro gamers might enjoy simply playing old games for reasons of nostalgia or discovery, ensuring old titles live on is vitally important for the study of those cultural elements – either from a sociological perspective, or a commercial perspective where game developers can look back on, and study old video games in the same way filmmakers, musicians or novelists would with their own art forms.

But more than merely just preserving old video games and getting binary images on disk, accessing them via actually playing them is just as important.  Thankfully in modern times we’ve seen an explosion of quality emulation systems for console based games, both in pure software form and more recently in FPGA, all of which have dramatically improved our ability to make these preserved titles more accessible without the limitations of keeping old hardware alive to play them.

One area where overcoming both the preservation and accessibility challenges of retro gaming is the humble personal computer.  PC gaming has a long and rich history of fantastic games across both the “bedroom hacker” scene, the boom of shareware, physical releases on optical media, and right through to modern commercial Triple-A titles and varied digital distribution methods. But what sets PC gaming apart is the sheer diversity of hardware, operating system software, and drivers to glue these platforms together, together with the sheer breadth and variety of publishing methods. A challenge not nearly as complex or diverse on most console platforms, where official game lists are often a finite number, and distribution was handled directly through the first party console manufacturer.

One individual who aims to tackle this challenge head on is eXo. Beginning in 2007, eXo started meticulously gathering and preserving MS-DOS games, a gargantuan task given not just the number of games, but just even trying to find out how many existed at all due to a completely open marketplace. With no central vendor or distributor, it meant even discovering and cataloguing every title would be challenging. 

Going beyond just collecting games as images on disk, eXo would also embark on an equally arduous task of creating an easy to use launcher to play these games on a modern platform.  The challenge here was enormous in scale – dealing with the sheer breadth of hardware options, video cards, monitors and resolutions, sound systems, hardware varieties (not just different speed CPUs, but cache and RAM options that changed, and sometimes broke game performance).  What has resulted was eXoDOS, an environment that in its current form of eXoDOS version 5.0 offers over 7,200 playable MS-DOS titles, complete with GUI launcher to simplify playing games down to often just a double click.  Along with it, gigabytes of metadata – game flyers and boxart, instruction manuals, information and other tidbits that built a full picture of what PC video gaming was like.

eXo has then continued this work into other areas.  eXoScummVM offers 387 games centred around the SCUMM (Script Creation Utility for Maniac Mansion) and offshoots – famous for extensive point and click adventure style titles on PCs.  And eXoWin3x, now at version 2.0 offering 1,138 games for Windows 3.x (a graphical operating environment that sat on top of MS-DOS, released prior to the complete Windows 95 operating system, released April 6 1992 and 30 years old this week!).

Helped now by an enthusiastic group of volunteers both in the preservation and launcher development sides of these projects, eXo’s important work continues on.

The sheer enormity of this work is quite amazing to consider just in human hours of effort alone, and currently stands as not only the single largest collection of preserved games for the PC platform, but also the fact that these are all offered for use in easy launchers that remove most of the difficulty surrounding actually accessing and playing old games on a modern computer.

eXo has very kindly agreed to a written interview to talk in his own words about these amazing projects.

Dan: eXo, thank you for taking the time to chat. Firstly, what drove you to start a project of this scale?  Was this a project of organic growth, or did you have a strong desire to build the entire game and launcher platform from the beginning?


eXo: It definitely didn’t start with a guy thinking he could single handedly preserve the world of MS-DOS. One night I wanted to play Discworld, so I fired up dosbox and began trying to get it working. I had gotten some simpler games running, but quickly found this game to be problematic. I started stumbling across several threads of other folks having similar issues, and I think that is when I realized the emulation of MS-DOS games itself was a significant barrier. Once I got Discworld running, I decided to get all of my DOS era adventure games running. This led to wanting them all in a nice little menu system, so I grabbed a free front end off the web (MaLa at the time) and started populating it with batch files that would launch games for me.

As I continued the process I started seeking adventure games I hadn’t played yet. After I got to around 200 or so I was under the mistaken impression that I had found “almost every adventure game”, and with that unearned sense of accomplishment I decided it might be fun to tackle another genre, such as RPG games.

Clearly I can not claim to have started all of this with the lofty goal of preservation. It really just boils down to me wanting to play a single game, and that turned into deciding I should be able to play any game I owned without spending an hour (or an evening) fidgeting with config files. Everything changed though as I began to realize that my hodge podge collection of DOS games had become more useful than the existing massive game collections that were floating around at the time – such as DOS Collection or DOS Memories Project. Both were great in the absence of anything else, but they were both also full of half installed games, scene rips, and sometimes just plain corrupt copies. I actually spent 2 years before eXoDOS trying to create a launcher for the DOS Collection. However it became clear to me during that time that the bigger issue was the files themselves.

Compared to these existing mass-dump projects, the games in my set were actually playable and this in turn drove interest in them. Very few other consoles or computers require working knowledge of an operating system and a dictionary full of emulator specific settings just to have a chance of getting a game working. This led to the most important tenet I have regarding this project, Preservation through Playability. It wasn’t enough to have a clean image of a game if most folks couldn’t actually experience it. The structure was all there from the beginning (launcher, config files, game files, etc), but I suppose I didn’t fully grasp what I was building towards for a good year or so. Sometime around the release of volume 3, a strategy collection, I knew I was going to try and capture every game. That is when eXoDOS, as far as what it is today, really took shape.

When you started out, did you have any idea of just how many titles you’d come across during these collections?

I had no idea at all. This is mainly because there will never be a 100% complete list of every released MS-DOS game, and there definitely was not one in 2007. Between dial-up bulletin boards, floppy swaps, and the early internet you have thousands of shareware and freeware titles that were released, some of which are known and still completely lost to time. Even among commercially released games though, major websites like mobygames are missing hundreds of entries. We just located 6 titles last week on a compilation CD that were completely unknown from a developer called Entrex. It seems that just about once a week someone in the group locates a game on eBay or some other source that is missing from all documented sources.

There is a lot of digital archeology involved in not only discovering unknown titles, but also actually finding all of the release information and then of course the actual files themselves. I often reach out to the original developers when I can find contact information for them, and the majority have been extremely receptive to the idea of getting their work archived and back into a playable state. I have been surprised by how often I find a developer who is happy to help simply because they have no way of actually playing their own game anymore. I’ve had original unfolded boxes, master disks, flyers, manuals, and all sorts of other ephemera sent to me by people who understand the value of preservation. Unfortunately, I have also come across developers who lost track of their software over the years and have no idea if a copy still exists. A lot of these guys ended up becoming executives for major companies in the tech industry and they stopped worrying about a game they made in high school years ago. 

One of the most unknown genres is edutainment, especially in the Windows 3.x multimedia era. A few of the better known titles, such as the MS Home series, have made it into online databases. Currently we have over one thousand Win 3.x era software titles that simply aren’t “game” enough to be considered a valid fit for an online games database, and there really aren’t any other databases that are attempting to track this sort of stuff.

Are you able to estimate the number of hours of effort between yourself and your volunteers that have gone into these existing projects?


Prior to 2018 I worked on my projects alone (with suggestions\submissions sourced from users of the project on various forums). At the time, I generally put 20-30 hours per week into eXoDOS and eXoWin3x. Accounting for taking weeks off here and there or getting distracted, I would generally round down and say I put around 10,000 hours or so into the project during that time.

Once I started my discord server however, I very quickly started meeting a lot of great like-minded folks who were more than happy to find their own niche to help me with. This actually increased my involvement with the project, as the overall scope quickly grew and we found ways to accomplish much more sophisticated goals with better tools. While each of my volunteers put their own varying amount of time into the project, I can safely claim that between myself and them we have easily put another 10,000 hours into the projects over the past 4 years. To be quite honest, I think I am under-estimating. I have put another 5,000 hours (at least) into this in the past 4 years, and to suggest my entire team has only put another 5,000 in is a severe disservice to their time and what they have accomplished. We have folks who focus hundreds of hours on finding the best quality manuals, work on helping me add language packs, record video snaps for every single game, bug test all of the games, record music from games to be used in the launcher, write plug-ins to assist the launcher in dealing with the project, seek missing games, and some who do almost all of the above depending on the day.

It is a true labour of love.

What are your thoughts on the current “GLAM” (Galleries, Libraries, Archives, Museums) industries and their efforts around PC preservation?  Has there been a big focus on formal preservation of digital content from those industries compared to other art or content forms?


This is a great question. Right off the bat I would like to say that I truly admire the efforts that go into operations like The Strong Museum and the Video Game History Foundation. That said, I don’t think they go far enough and a lot of that has to do with the stigma and grey area of existing copyright laws. Older art forms do tend to have much more infrastructure in place to preserve them, however, they are also simply easier to preserve in terms of the necessary requirements to appreciate\experience them. A book or a film, for example, can have its original media preserved and then transferred to a modern media. You can read a classic novel on your e-reader or watch a classic film on your phone or television. Computer software is a bit different. Emulation helps us access these original files, but there is so much variability in the *way* you choose to emulate something. The speed of the machine, the video card, the sound card, the control mechanism, and even the interaction with physical inserts such as manuals and copy protection create a scenario in which the game you play today may not feel anything like the one you played thirty years ago. Maybe you played it on a Tandy, but the preserved copy you fire up is set up using CGA with a PC Speaker.

This is where playability becomes such a key factor. Even though many of the museums that catalog these things do allow public access, it would still fall upon the researcher to be able to provide the right hardware. To know how to install and set it up. To be able to navigate the multiple issues with magnetic media. And of course, to physically be able to visit this location in the first place.

I happen to be lucky enough to live relatively close to the National Videogame Museum in Texas. It is a fun place to visit, and they even have a row of vintage computers that visitors are allowed to use. It would be impossible however to expect them, or any other museum, to provide a way to be able to enjoy *every* MS-DOS game. You would need a dozen or so computers that covered everything from 1980 through 1998, and that is without getting into specialised hardware or playing multiplayer games.

This is why it is so important for people to be able to enjoy these games outside of a museum environment. While having a clean boxed copy and original game media in a museum is incredibly important, unlike a film or a novel, it can not be appreciated in that setting.

What are your thoughts on the commercial gaming industry and its own efforts to preserve old PC games – both the offerings of first party rights holders themselves, and various publishers that offer commercial access to retro PC titles?


This is where the industry suffers the most. Of the 7,200 games that are available in the 5th release of eXoDOS, less than 3% are available in any commercial format. In many cases, no one even knows who owns the rights. In other situations, the rights have been absorbed by mega companies who have no interest in these games from a commercial standpoint.

Even among the small percentage of games that are commercially available, they are often stripped down shells that offer no options. Good Old Games and Steam do not provide original game images, they often strip what they consider to be extraneous driver files, and then they also often remove the actual setup files.

Take a well known series such as Sierra’s King’s Quest. The Collections available on both of these sites outright omit the original release of King’s Quest 1 in favor of the later SCI re-release. We are talking a difference of about 250kb (compressed). For the games that are included, they have all been set to run using adlib and either VGA or EGA (depending on whatever the best mode available for that game is). The forums for these sites are full of end-users wondering how to change the graphics to whatever video card they grew up with. Or maybe they are interested in hearing these games with Roland MT-32 audio (which is what they were originally composed on). The solutions require grabbing missing driver files from external sources and hand modifying configuration files due to the setup file missing. Many times the user has to replace the copy of dosbox distributed with the game with a newer copy that actually allows proper emulation of these features.

Not only do I not see a good reason for these choices, but I believe I can make a compelling argument for how this practice actually damages preservation. As these copies become the only commercially available version of the game, this also means they become the primary copies of the game. I have already seen cases where a GOG release becomes so ubiquitous online that the original media is essentially lost. I have had to go out and track down an original copy and dump the files all over again just to find a non-GOG modified version. At some point, the number of people who remember these games were even capable of MT-32 begins to dwindle so low that this fact is lost. It is not as much of a concern with a massive sales success such as King’s Quest (which sold hundreds of thousands of original copies) but it is a major concern for games that are lesser known, where the GOG or Steam release is likely to be a person’s only experience with it.

This is compounded by the fact that none of these commercial companies really seem to actually care about the quality of what they are selling. For years the copy of Ravenloft: Stone Prophet that GOG was selling was bugged to the point that it could not actually be beaten (it also had other silly issues like arrows doing no damage). Triggering the ending cutscene also triggered a catastrophic crash that closed the game. Over the years multiple users complained about this, but nothing was done. Last year a contributor to the project discovered the reasoning for this crash and managed to fix it. Ultimately, it was not a complicated fix…. the cinematic files on the CD image they were using were misnamed using underscores instead of hyphens. Had anyone at GOG actually spent time looking into this, I have to imagine they could have solved it relatively easily. Ironically, just 2 days after we posted the fixed files, the GOG release was updated as well.

While I am proud that work derived from our project was beneficial in this way, I also do not believe it is the job of me or my volunteers to handle GOG’s quality assurance. It does shine a light on the fact though that the folks who are putting these games up on commercial marketplaces have no vested interest in presenting the best possible version of the game. From what I can tell, it is actually rare for rights owners to provide services like GOG with any original files. Rather, they source those files from projects like eXoDOS. We have seen multiple occasions where files that we generate within eXoDOS as name tags have ended up in GOG releases. So, this sort of brings up a chicken and the egg conundrum. In how many instances is the commercial availability of a game only present because the game was first preserved and made available by a non commercial project such as eXoDOS?

It also damages preservation by creating the illusion that these games have been preserved simply because they are available to purchase. Your average person has no reason to believe that the GOG version of a game is lacking files\drivers\setup\etc. It creates a narrative that projects such as eXoDOS are undercutting commercial efforts. But this is like watching a movie studio continually edit and cut down a classic movie, and claiming that because some version of the film is available the entire film has been preserved and no one should bother archiving the original. I imagine Star Wars fans can relate to this a bit.

There’s often great conflict in the world of retro gaming between being able to play a game right now using “high performance but not accurate or portable” tools, and being able to ensure a game can continue to be played on future platforms on tools that may be lower performing now, but ensure long term use.  What are your thoughts and goals with such tools, and why?


I have somewhat different feelings about this depending on if I am discussing consoles or computers. The conformity of consoles makes solutions such as FPGA incredibly attractive. When implemented properly, this is the equivalent of hardware preservation. With computers however, the answer gets more complex.

There is a 486 FPGA core available that many late 80’s and early 90’s games from eXoDOS run decently on. Users have written some very clever utilities to port my work in eXoDOS over to this core, however it is at the cost of choice. The primary benefit of an emulator such as DOSBox is the ability to configure that instance to a wide variety of DOS compatible hardware. If one game written in BASIC requires an 8086 with 256kb of RAM and the next requires a pentium with 64mb of RAM and a 3DFX video card, then the associated configuration files for each game can direct DOSBox to emulate the proper environment.
An FPGA implementation would require a different core for each chipset (8086\88, 286, 386, 486, etc…). It would also require hardware documentation of all components the system uses.

That said, DOSBox is tackling the mammoth task of emulating 2 decades worth of hardware and the entire OS that was sitting on this hardware, so it clearly has limitations as well. My biggest fear is the “romification” of DOS games that some of the more ridiculous DOSBox ports have attempted. Multiple times in the past year I have seen a desire to treat MS-DOS games like cartridge ROMs. Folks want to throw a bunch of DOS games into something like RetroArch, have them all mapped to their controller, and call it a day. This is about as far from preservation as you can get, and I have been very vocal in my opposition to it, at least in terms of how it relates to eXoDOS. For me, preservation is about maintaining the integrity and the options that the original software provided. We use multiple versions of DOSBox within the projects to ensure that games are presented properly. For example, Silmarils games often shunted DAC through the PC Speaker. On original hardware this sounds like a multi-tonal chime, however in vanilla DOSBox it is a fairly obnoxious BEEP. Luckily DOSBox-X has support for this little trick though, and so when played over there you get the original chime.

I do believe that FPGA implementations are the likely future of preservation. Emulators, by their nature, are tied to specific operating systems and current hardware speeds. But I also do not believe we have the necessary existing cores to make any meaningful jump today. Eventually, as more hardware is documented in FPGA and the switching of cores becomes more seamless, I think seeing something like eXoDOS (in its full form, with menus, metadata, and images) sitting in an FPGA environment would be amazing.

PC audio was quite a diverse landscape in the MS-DOS and early Windows eras, with many different ways to get sound and music out of old games such as PC speaker, digital sound and a huge variety of MIDI, FM synthesis and wavetable devices.  Video too – with various colour and resolution modes of titles. How do the various projects aim to tackle challenges like these and present these options to the player, where one game can be experienced in a number of unique ways?


The launch options present in eXoDOS are probably what I am most proud of.

One of DOSBox’s biggest weaknesses is its documentation. There are folks who are wizards at the base features of DOSBox, but have no idea how to use forked builds to enable extras such as general MIDI or 3DFX.

When we decided to add Roland MT-32 support to the 4th release we found just how lacking information was on which game’s actually supported that sound card. Having not grown up with one myself, I wasn’t super familiar with it. So the first time I played a game and heard it (properly working), I was outright amazed. This is another example of a DOSBox feature that isn’t incredibly well documented. I had played with MT-32 sounds before this, but I did not have it set up properly at the time. With no proper sound bank mapping, the game I was trying to play sounded like someone banging pots & pans together. Getting the MT-32 properly setup made me want to play through my old favorites all over again just to hear them “properly”.

For the audio options, we identified every sound card that DOSBox was capable of emulating. We then audited the collection (both by online databases and by searching driver files in the game folders) to find games which supported these sound cards. The sound cards we have enabled support for so far include pc speaker, tandy 3-voice, covox sound thing (Last Half of Darkness in particular has some great speech samples when this is enabled), gameblaster, adlib gold, sound blaster, gravis ultrasound, MT-32, and general midi (via fluidsynth). For cards like the Ultrasound or Adlib Gold, it was not about enabling everything game that had these on the setup menu. In many cases, they push nearly the exact same audio through. However, there were games developed with these cards in mind. Dune, for example, is known for its gorgeous Adlib Gold soundtrack while the games that came out of Epic often took much better advantage of the GUS. We then create playlists within the front end for games that support each sound type. The user can easily browse all games for these sound cards without having to know which games are supported ahead of time.

Video options are handled on more of a case by case basis. Well known games such as the Sierra adventures have had their rarer late release EGA counterparts hunted down, and they are provided as an option. Early games released for the Tandy & PC Jr machines are also often presented with a choice between the higher color Tandy\PCjr graphics and the 4 color CGA palettes. Generally, if there is any question as to what video mode to support, options for all modes are added.

We also have launch options that enable save state transferring (for games that allowed you to continue progress from a previous or related title), IPX\Null modem emulation over IP for multiplayer gaming, 3DFX support, composite video, and modern conveniences like auto mappers and mouse helpers are also options.

What have been some of the biggest technical challenges in preserving and being able to play some of these games? Have there been any stand-out titles that were particularly difficult to locate or emulate?


The number one challenge has simply been identifying what is missing. Once we have a game in our sights, the next challenge is acquiring and dumping it. Magnetic media from that time period is notoriously unreliable. The primary source for most of our games is eBay and the storage of the media over the past few decades is always questionable. I have had brand new shrink wrapped games arrive that refuse to read. Meanwhile, I have had disks that looked like they were used as coasters and smell like cigarettes show up and read just fine.

The process of actually getting the games to work hasn’t been so bad. After a few thousand games you start to learn a lot of the tricks, however every so often a game comes along that I kick out to the community to see if anyone has any ideas. Luckily there are some true experts out there that I can fall back on for the really problematic titles that elude me.

Then there is the challenge of knowing if the game is actually functioning as it should. This one is trickier than it sounds. For example, many games have been dumped incorrectly as ISO files, which strips any potential CD audio tracks. However, it isn’t always clear when a game is supposed to have CD audio, as some would default to MIDI if the CD wasn’t detected in the drive. At one point we essentially went through every commercial game released after the introduction of the CD-ROM, looking for evidence of a version that may have had CD Audio. Another example are games that had unique copy protection that was cracked at some point in the past, but had an adverse effect on gameplay. The games Summer Challenge and Winter Challenge both can detect that they have been altered. Rather than warning the player about this however, they simply make it impossible to ever get a gold medal in any event. In one instance (speed skating), on the last lap it shoves you into the wall….LOL. Catching things like this usually requires someone coming along who owned the game originally and knows how it is supposed to behave. When I am play-testing, I am more likely to assume that I am simply horrible at the game than to think the game is sabotaging me due to a hack that was applied to the files 30 years ago. Often, finding clean files that have not been altered is the final challenge to getting any particular game added. Of course, this means now either finding original documents or coming up with a new hack that preserves the game’s original presentation while not causing unintended side effects. One side project I am proud of is our code wheels project. A member saw an interactive codewheel online and learned how to start building them for us. He then created an exhaustive list of all known codewheels, which we spent the next year sourcing. We carefully disassemble them, scan each layer, and then combine them in html. Now users can check the extras on these games and pop open a fully functional code wheel.

Finally, there is the challenge of simply finding the full capabilities of these games. I’ve lost count of the number of times we found a hidden command line switch 10 years after I thought I already had a game working perfectly. Maybe it enables a hidden audio option. Maybe it enables a high-res mode. Any given month I am going back and reopening 10-15 existing games and either adding new options to them or enabling new ones that we just found.

As far as specifically problematic games go, one thing you start to notice after a while is a correlation between unquestionably bad games and launch problems. Which means the games I have spent the most time working on are often games that very few people will ever actually play.

One title that comes to mind is Backroad Racers. The game was released packed with an actual Revell car model (A BOSS Mustang if I remember correctly) and then released again later as budgetware by Expert Software. The video decoder the game uses absolutely refused to function properly in DOSBox and required the creation of an image that drops to real DOS, loads specialized drivers, and then allows you to watch really silly FMV scenes as you play a game to controls exactly like you would expect for a game packed in with a car model. All in all I spent over a week on this particular game. Ever since I got it working properly I’ve never booted it again and I probably never will.

Another fun story is a silly game called Planet of Lust. The MS-DOS version was ported from the Amiga, and in the process the entire color palette of the game was screwed up. Unfortunately, this is simply how the DOS version has always been, as the game wasn’t exactly successful enough to warrant an update. One of our users however popped the game open and managed to replace the DOS bitmaps with the original Amiga ones.. So, we can proudly say that eXoDOS is the only place to play Planet of Lust in MS-DOS the way is *meant* to be seen, hah!

What are your plans for the future with the existing packs and launchers?

Currently I am a solid year into development of the next eXoDOS release, which will be version 6. On the backend, a lot of changes have taken place that create a much snappier user experience, while also layering in the ability to have global settings that affect all games without having to go in and set each game. We have identified between 3 to 4 thousand shareware\freeware titles that were not in eXoDOS previously, all of which are planned to be added. We will be making use of the playlist features within the front end to allow users to easily sort between the commercial releases and the homebrew stuff in order to make sure this massive influx of games does not bury the quality stuff that most people showed up for in the first place.

Another thing I have begun working on are focused deep dives. For example, any collection of DOS games is going to have DOOM in it. But anyone who grew up playing it in the 90’s knows that the episodes that shipped with the game were just the tip of the iceberg. I’ve created a new game entry that takes the 100 top WAD files and total conversions (as defined by DOOMWorld), and created a batch menu system that allows the user to navigate from 1994 through 2003 and play the top ten WADs released for each year. It also adds 10 of the worst WADs I could find as this more accurately represents the levels most of us ended up playing. There is a third “bonus” category where I have put WAD files such as the recreation of the Stauf Mansion that resides as a bonus on the 11th Hour cd’s or the recreation of the National Videogame Museum WAD that was released recently.

Other games we have done this for include ZZT and Unlimited Adventures, both of which have hundreds of user maps and descriptions available for immediate play through a menu system that removes the need for the user to know anything about loading custom mods or levels. There are several other games that had (or have) active communities behind them that I would like to add support for, however the first challenge is finding lists of the “best of”, which can be surprisingly hard for games that you would have thought there would be lots of discussion on. For example, I have yet to find any definitive Duke 3D lists.

Another huge project is the language packs. For years I have been asked if I will include games that were released in non-English languages. I had avoided this specifically because it is nearly impossible for me to install, configure, and playtest a game if I don’t understand what is going on. It seemed like a recipe for mediocrity if I added anything I could get working without taking the time to verify it. Currently I have two incredibly active staff members who have created a thorough German language pack. The back end will allow this to be easily added to an existing install of eXoDOS, with the ability to plug in other languages as curators for those languages step forward. Work has been put into the French and Polish packs as well, however my hope is that once we release the first pack it will generate interest in languages which we currently do not have a curator for.

I have also begun to try to capture other aspects of the era, such as adding in computer gaming magazines and online e-zines from the era. My hope is to eventually index them so that a user can easily jump straight to a preview, review, or even advertisement within a magazine from that game’s entry.

I am also working on eXoWin3x Version 3. The third release will more than double in size, with a majority of the games being completely undocumented commerical and eductional releases. For any criticisms I have made above about the preservation of MS-DOS software, you can multiply them by 10 for Windows 3.x. Incredibly well known games such as the 16 bit windows releases of Carmen Sandiego and Super Solvers have absolutely zero commercial or preservation representation outside of some files that have been uploaded to sites like archive.org. The most common question I am asked in regards to Win3x is if there are any games worth playing, which in of itself is a really strong indication of how little that platform is understood as there are some remarkable gems hiding in there.

Finally, we have a handful of other projects that are being worked on quietly. MS-DOS and Windows are not the only systems that lack any meaningful preservation currently. With the framework we have built out, we have at least 2 more releases that are coming in the next year or two that I am looking forward to releasing.

What are your plans for the future on more modern PC titles?  Will we see an “eXoWin9x” pack?  And what do you see as the biggest technical challenges for this platform, both in preserving the games themselves as well as running/emulating them?

A 9x pack is my white whale. The primary reason an eXoWin9x pack doesn’t already exist is due to the technical limitations of properly presenting that software within the framework that a project like this demands. To be clear, getting a 9x game running is not necessarily difficult for someone who has the proper tools. The problem is putting together thousands of Win 9x games in a portable manner.

The first challenge is file size in regards to the operating system itself. For Win3x every single game has its own copy of Windows. This adds about 20mb to the zipped file size of each game. Due to the incredibly picky nature of Win 3x though, this was essentially a necessity. Simply changing the color depth for game 2 would make game 1 stop working. God forbid you update your quicktime install for a new game, as the old game will throw a tantrum and refuse to work anymore. It was also the wild west in terms of how software treated the operating system. Games had no qualms about writing their configuration files directly to the C: or the windows folder. For all of these reasons and more, it was imperative to give each game its own operating system.

For Win9x, this isn’t possible. A compressed Win98 install takes approximately 200mb. Granted, they can be stripped down smaller by removing all sorts of dependencies, this isn’t feasible to do for every game (thousands of times) nor does it make sense to do this up front and assume it won’t adversely affect games you add later. Currently, we have at least 10,000 identified Win9x games. This would be 2tb in duplicated OS files alone. Not the full pack…. Just Win98 10,000 times. And that is compressed. So for a Win9x pack to be feasible we need parent child images. Something that would allow us to have a master Win98 install, with each game installing to its own child image. This would allow any changes to the OS folders (or the C: at all) to be captured by the child\diff image. The second primary need is dynamic images. Most image based software uses static images that take up the same size on disk as the total size of the image. So, if you create a 1gb drive image, even if it only contains 100mb of files, it will take up a gigabyte on the disk. A dynamic image will grow as the file size grows. This would allow us to report to the OS that there is plenty of free space available without having to make a custom image for every single game that has exactly the amount of space we need for the installed product.

Ultimately, we need both of these features combined with an emulator that has the horsepower needed to run games released from 1994-2002 (even later honestly). We came close to attempting to use PCem, as it added features very similar to those described above. However shortly after the project went on hiatus for a while and only recently returned. When we are talking about sinking another 10,000+ hours into something, it is key that the software we are relying on doesn’t get pulled out from underneath our feet. It also helps to have a working relationship with the authors, as a project like this is about the biggest stress test you can possibly put an emulator through.

All of that said, we have an eye on the future and we currently have somewhere between 4 and 6 terabytes of Win9x software images preserved and ready to go. So at the very least, we are trying to prevent the data from disappearing while we await the proper technology to make these games playable again.

How can the community help your projects, and what is currently the highest need for the projects?

I spend a lot of time on the discord server just thanking people for showing up and discussing the projects. The insight we get from various folks is always invaluable. Someone who lives and breathes Ultima is going to be able to give me insight that I never had as I never played those games.

Beyond that, when folks ask me how they can help I usually just ask them to hang out and see what is going on. At any given time I have between 10-20 volunteers working on some aspect of the project. Not a single person was asked or assigned their tasks. In all cases, I either put out an open call for help or they just started assisting and before they knew it, they had been recruited.

One fellow, who goes by Python, started out as just feeding me occasional detailed bug reports. I was impressed by how thorough he was. It turns out he had started a similar project before realizing eXoDOS existed, and he was feeding me reports based on his own findings. Soon I found he was putting as much time into the project as I was. These days he is the second in charge on the project, and a great personal friend.

I could sit here all day and brag about how amazing everyone on the staff is. I gush about it sometimes in the discord channels, but having spent 11 years doing this myself I have more appreciation than anyone in regards to what it takes to tackle something like this and stick with it.

I have so much respect for the guys on my team, and I try to show my appreciation for them whenever I can. eXoDOS wouldn’t be what it is today without everyone’s time they have put into it so far.

Really, what we need most are people who are interested and open minded. The number one thing is taking time to understand how the current processes work and then figuring out what you can do to help us improve. The fellow working on our video snaps project got involved after hearing us discuss it one day. He created a replicable system that allows anyone to jump in and create uniform MS-DOS video snaps for all of the games in the pack. In turn, we all have a blast reading through that channel watching him and the guys helping him discover some of the odder games in the pack. The key is really taking the time to understand the project’s needs. A lot of suggestions that are intended to be helpful are non-starters as they don’t fit within the preservation goals of the projects.

We can always use folks who are interested in assisting with language packs, missing manuals, metadata research, or any number of other facets.

What we don’t need is a copy of DOOM. And yes… that has been offered to us before. Multiple times. Like… way more times than you can ever imagine.

Finally, on a personal note, what are a few of your own favourite titles available in your current packs?

It is funny to think I started this because I “didn’t want to waste an evening every time I wanted to play a game”. Thousands of hours later, and I still haven’t taken much time to actually enjoy the project as much as I should.

A lot of my all time favorite games are driven more by nostalgia than actual game quality, however I love it when I come across a game I wasn’t familiar with that still captures my attention decades after it was designed. I grew up on Sierra, Bullfrog, and Apogee as my primary influences, but I almost played everything I could get my hands on. King’s Quest 6 is a high water mark in terms of story, presentation, and puzzles and Blood is my all-time favorite FPS as it has some of the best level design and unique weapons of any other FPS of the era. And while I love so many of the usual suspects (X-Com, Command & Conquer, Duke 3D, etc), I have always had a real soft spot for text adventure games, especially the Legend and Magnetic Scrolls releases. For me, what always made computer games seem so much more interesting than console games growing up (granted, we are talking the NES era and prior) was this belief that you could do almost anything in them. While genres like interactive fiction clearly have severe limits, as a kid it felt like I could try to do absolutely anything… even if almost every action resulted in something like “I don’t understand” being printed back to me from the parsar.

In eXoWin3x The Dark Eye is a gorgeous stop motion animation game based on the works of Edgar Allen Poe, narrated by William S. Burroughs. Timelapse and The Lighthouse (not a windows game, but still worth mentioning) were both Myst clones that ended up bringing unique twists to the first person adventure formula.

All of my packs have a “favorite” list within them that is designed to help a user narrow in on some of the best games ever released for the platform. I find it is a great way to not be overwhelmed by choice and yet still find something worthwhile. I always shove my nostalgic favorites into these lists along with the heavy hitters just to give them a little boost!

I would like to thank eXo for generously giving his time to me for this interview.  You can find more information about the various projects on his main site:


Donations to the project can be made at the site, and physical “big box” collectables are also available from eXo’s Etsy site, with all donations and proceeds from purchases going towards acquiring new games to add to these collections. eXo notes that as little as 10% of software acquisition costs to date have come from donations and purchases, with the other 90% generously coming directly from project participants. As such, any help greatly pushes the project forward: