Hacker News Reader

Comment by icco | Hacker News Reader

Being a long time Linux user, I am very aware of the "I don't want you as my customer" feeling.

Being a software developer, I have never understood the decision on why to essentially limit the number of customers you can have. I think the only thing worse than not supporting the majority of platforms is not making them have feature parity.

View on HN

Fairly obvious reasons.

-price,time,skills

They do not do it intentionally. It just is not worth it, or possible.

While I can understand not porting e.g. a DirectX game to Linux/OSX or any other program which heavily relies on OS specific libraries/features, I could not yet understand why not to port it when it is possible with little effort. I'd really like to hear an explanation from someone with knowledge in that matter, unfortunately the only example I can come up with at the moment from the top of my head is WoW/Warcraft/Starcraft. It runs on Win/OSX but has no Linux client. I do not understand why Blizzard would not release a Linux client for which no support will be available. A small team could probably achieve a port in reasonable time and create the update infrastructure around it.

How ironic that Blizzard, by far the brightest star in releasing simultaneous Mac AAA titles, is being singled out here.

I'm going to pull a justification out of my ass: if you support things halfway and do things halfway the brand suffers. Before they'd make a linux client they'd do X-Box then PS3 then Wii for sheer ROI reasons. They aren't doing those either though even with larger consumer bases.

Yes I agree with you in regard to Bilzzard.

To be more precise, my questions is: You have an OpenGL based game running smoothly on OSX, you decide not to port it to Linux. Why?

1. ROI

2. "Have to" offer support for Linux to maintain brand image - very hard

3. hurts game image "game for geeks"

Obviously, it is hugely dependent on the piece of software you have.

ISTR that they have some fairly sophisticated anti-cheat stuff running (esp. for WoW & Starcraft) Warden, I think it's called.

Essentially, a sort of background process monitor/snooper that tries to detect cheating behaviour and block or report it. I have no idea if that could be ported to Linux, and whether in doing so they'd:

1) Get grief for requiring it to run as root

2) Fail to detect all sorts of 'easy' cheats, ruining the online services and leaderboards for everyone

3) Interact horribly with the general fludity/flexibility of a Linux system.

4) All or none of the above.

I'm told that WoW works pretty well in Wine, so I wouldn't think those are valid reasons. Warden only has to inspect the running game process, which does not require root. It also looks at the list of running processes, which doesn't require root.

AFAIK it doesn't require root/administrator on Windows and Mac either.

2. "Have to" offer support for Linux to maintain brand image - very hard

good example of this is with HoN (Heroes of Newerth) where I believe the guy that looked after the Linux port left, leaving S2 having to outsource all their updates every week. The community is fairly good at hacking their way round day to day issues, but it's far hassle free and I'm sure is costing S2 more money than it's making them.

http://braid-game.com/news/2008/08/misc-linux-questions/

That was 4 years ago. Braid finally got a port in 2010 for a previous Humble Bundle. Braid is just a small Indie game. Imagine trying to port some of the big budget productions we have now.

There was Loki back in the late 90s early 2000s that did Linux ports, but they didn't last long.

That was back when most games were developed for single platform though, namely Windows for most/all titles Loki was porting. Today, almost all AAA titles are multi-platform with a console being the lead platform; and lots of indie games are developed somewhat platform agnostic, as well, in order to support OS X or mobile operating systems.

If I had to take a guess, it would be that it probably has something to do with the wild constellation of libraries any given Linux machine could be using, not to mention version of those libraries. Do you build your game to support ALSA, or OSS, or Pulse Audio, or all three? (I'm not a Linux dev so these examples might not be technically right but it's in the spirit of my point.) Do you use the dependencies present on the system or include your own? Does the user have good drivers for their hardware? Do you target OSS or proprietary drivers, or both? Often proprietary drivers have showstopper bugs on Linux that take forever to get fixed, if they ever do.

When your Linux customer complains to you that your game doesn't work on their version of the ATI proprietary driver, what do you tell them? "Sorry, wait till ATI gets off their ass and fixes their driver, otherwise, you can try entering this arcane set of CLI commands to see if it helps, cross your fingers you don't bork your existing driver in the process, and if not here's a refund?"

It's probably just not worth the support headache. After fighting with Linux driver and dependency hell more than once I don't blame them.

Edit: even some of the past bundles have had Linux problems. I think only half of the games in the last bundle would even install on my stock Ubuntu laptop. More than one of the games--the two Shadowgrounds, I think--in an earlier bundle needed some arcane hacks to get sound working. It ain't easy supporting Linux.

If there is one thing the Humble Bundles have proven, however, its that you CAN port your game to Linux and make money. Linux users have traditionally paid more for the bundle, and there are very few good reasons why a game cannot be ported to Linux. So many great games in the Bundle series are available on Linux, so clearly its possible - and clearly there is money to be made.

So its really just a matter of groupthink about "Linux isn't viable" thats the block, here.

Reply on news.ycombinator.com