A Developer's Perspective On Porting Games To Linux | Hacker News Reader

A Developer's Perspective On Porting Games To Linux

phoronix.com

88 pointsintull30 comments

One slide mentioned having to fix hardcoded backslashes to forward slashes, and using path.join to make it portable. But on Windows, you can use forward slashes as well at the system call level. You only need to use backslashes on the command line, since many programs interpret a forward slash as the beginning of an argument.

Yes, but that doesn't mean the original developers used them :)

Some developers are shockingly Microsoft-centric...

You mean like DirectX developers? Funny that really.

Doesn't have to be on purpose, for instance Visual Studio inserts backslashes in autocompleted C/C++ header paths by default (can be set to forward slash in the settings, but default is backslash). The case-insensitive filesystems on Windows and OSX are usually also a problem when moving games with hundred-thousands of source- and asset files over to a Linux build system.

OSX/HFS has case insensitivity? Since when?

I was baffled too but that's like it is by default, on my MBP #include paths don't have to match case, just like on Windows. Or you can do a 'echo blub >bla.txt' followed by a 'echo blob BLA.txt' and there will only be a single file 'bla.txt' which contains the string 'blob'.

I meant they didn't have to use path.join, they could have changed them all to forward slashes and it would still work on both.

Seeing things like this makes me really happy, even if you are a non gaming linux user, work like this benefits everything else you do and pushes companies like AMD and Nvidia to improve their drivers. I know that since valve started supporting linux a few graphical issues on my fedora laptop have disappeared.

The Mesa improvements mentioned in the slides: http://www.phoronix.com/scan.php?page=news_item&px=MTY4MjA

I hope this at least least will motivate more developers to use GL/GLES for game development since it will be easier to keep it multiplatform (iOS, Android, OSX, Linux, Windows). But still there are a lot of things mentioned there that also applies for general big project development on Linux.

Why do people use the term "postmortem" for retrospectives of projects that didn't fail?

Because people who work from chairs behind desks in climate-controlled buildings get all their excitement from stock phrases and edgy buzzwords?

Why do people say they have a fast-paced work environment when most of the employees move at exactly zero meters per second for hours at a time? It's all about marketing, and the lies we tell ourselves to prop up our self-image.

A rose not named Bloodpetal Olfactoryblastin' Extreme Transcendent Amazing Ultraromance Blossom does not, in fact, smell as sweet.

At least in the games industry this is the typical term for the analysis of a project or process that has been finished, successful or not. Don't know about the origin of the term though.

You've got to love the way language evolves.... I mean, who was the first person to write every word in the sentences you and I are using right now! Who first said "you" ? Who first used the contractive "you've" ? Who first used the word "got" ? :)

Postmortem doesn't (to me at least) imply failure. Specifically for games it's just a critical analysis of what worked and what didn't on a project for the purposes of learning for the next one.

I don't know if the phrase was originated by Game Developer Magazine, but it certainly popularised it's use within the games industry. The GD Mag postmortems are generally exceptionally candid and a great series to learn things from. See http://www.gamasutra.com/features/postmortem/ and the general archives at http://www.gdcvault.com/gdmag (I'm not sure if the website has all the postmortems that featured in the, sadly defunct, printed magazine).

you know 'mortem' means 'death', right?

Not all deaths are failures. Sometimes things just end.

It's only a game. You may restart it using a new life.

you know 'post' here means 'after', right?

Being around after death sounds pretty successful to me. Especially if it involves lots of quaffing and hairy Vikings.

Is there something else you'd rather they call it?

"Postmortem" refers to the project, not the product.

The project has ended, but the output from it may live on.

Microsoft seems to be rapidly losing mindshare.

Not yet, developers are still in the "toying with Linux" phase. Microsoft will be losing mindshare when people install Linux for reasons other than "I don't like Microsoft" in great numbers.

They might be toying with Linux, but they are shipping for Android. Every new game faces 'do we want a piece of 1 billion devices pie?' magic question.

While I don't disagree with your point, developing for Android isn't too close to developing for Linux, and the games available are very different at the moment.

They have 89% desktop market share, and it has been this way for almost two decades. I don't think "mindshare" really means much.

Growth of it does.

Based on what, actually? Did I miss something in the linked article?

Based on some pseudorandom guy's slides about porting to Linux getting posted on Phoronix? Really? I expect that kind of reasoning from stock analysts, not HN posters.

Microsoft has indeed lost mind-share among game developers in the past years, pretty much all smaller and medium publishers which kept themselves afloat with smaller PC- and Wii- titles have now moved on to mobile, which means the dev teams have to move as well, and for many of them, OpenGLES2 instead of D3D9 will be their new main rendering API (since there are still a lot of WindowsXP machines in the world, it didn't make much sense for casual game devs to move on to D3D11).

I wrote a blog post a while ago why I basically ditched DX:

http://flohofwoe.blogspot.de/2013/10/farewell-directx.html

Reply on news.ycombinator.com