[General Battuta]

I've said this before but I think the forum metrics from the past year only make it more urgent: we must have an end-user-friendly install solution built to professional standards if we are going to maintain or grow the FreeSpace community.

It must be self-contained. It must not require reference to any outside website or list of instructions. It must be legal. We need to do everything we can to stress-test it and work out any kinks. We need an installer that is useful not only to power users but that works for the evangelizer, the newcomer, the skeptic.

Goob's been working on a solution to the problem, but for whatever reason (and it may be a very good reason), that effort is not complete. So please, let me renew the plea: what can we do to achieve this goal?

I think this is the single most important thing we can do for the future of HLP and FreeSpace. If we cannot achieve it, I think the only alternative is for individual mods to start shipping preconfigured standalones.

[Phantom Hoover]

I'm kind of wondering if Nexus mod manager can provide a suitable solution for this. I know it doesn't fit all your criteria, but it has built up a massive user base despite that and it might be able to deal with all the necessary modifications to retail FS2 with minimal oversight. I meant to look into this further but got sidetracked by Christmas.

[Kopachris]

I'd like to take a look at integrating or cloning innoextract with Python.  If that can be done, it'd be fairly straightforward to create standalone installers that work on Windows, Mac, and Linux.  (Why Python?  No reason, other than that's what I'm most familiar with.  It's pretty easy to make standalone executables that don't need Python already installed.)

I think the biggest aid would be a more centralized repository of mods, SCP builds, tools, and standalone TCs with a small index file the installer could download when run.  I'd be willing to host that myself, since I have a server I'm otherwise not doing much with.

[karajorma]

I still think this functionality should be built into the launcher. That was the plan from day one after all and I believe it's still on the roadmap.

[m!m]

We should also agree on a file format to manage the mod versions so we could host some sort of mod repository with informations like mod version and install commands in one or more file.
Doing this with a widely supported format like JSON or XML will make is easy to integrate it into the launcher. That way we could also add dependency tracking to each mod to make sure everything is installed for the mod to work correctly. Hosting the actual files could be done somewhere else, we would just include the URL of the download so hosting the mod files on a separate server would not be a problem.

[niffiwan]

I think there's value in having a separate Freespace2/retail installer for OSX/Linux (i.e. to mirror the retail installer for Windows), no matter what form the other installer takes (Goober's/wxLauncher/new(?!?).  Hellzed has made some good progress already on this, with the script written in bash it should probably run on OSX as well as Linux. This could even be run by the other installer on Linux/OSX.

Of course, that is a secondary concern, the primary concern is mod downloading/dependency resolution/installation/verification. If wxLauncher was intended to do this then perhaps we could get a "report" from the wxL devs on the current state of wxL mod handling, what the current plans are, and info on how non wxL devs can contribute?

edit: clarify / refine the post

[fightermedic]

Battuta is right, this is the most pressing thing at this point
90% of the people will not install something if it requires them to read more than 2 lines

[Phantom Hoover]

installing fso takes a lot more than '2 lines' and frequently results in users getting bombarded with useless errors we have to decipher for them

[Bobboau]

How is it supposed to be stand alone and legal? At least if it is going to result in a functional game at the end of the process.

[fightermedic]

i don't think somebody is talking about standalone
rather a case of reducing a noobs usual way of installing stuff from:
- install fs2
- read through confusing instructions, get an up to date launcher and copy all the files by hand
- read through confusing instructions, get an up to date version of fso and copy all the files by hand
- read through even more confusing instructions, get the up to date mediavps (several seperate downloads, a noob will have no idea what he needs)
- select the exe and mod you want to use (read: mediavps) in the launcher
- have crashes because you installed freespace into a directory with a too long file path

to:
- install fs2
- install everything fso from one .exe, and be ready to go for playing with the latest stable fso and the latest mediavps
 (the installer should come with a BIG warning that freespace can crash from too long file paths)

[Hellzed]

I agree with General Battuta (and many things said in later posts).

@niffiwan : i really hope my current work *will* be superseded by a graphical tool included in a launcher... I'm not a real programmer, I have close to zero experience, so I'm a bit worried about the quality of what I'm doing. And I need comments, advice, help for testing... Even comments and advice from people who do not use Linux.
I think we really should go for something a bit different from a mod installer linked to a repository : people should have the choice to add mods from other sources and still manage these from the mod installer. I'm not sure about my current implementation, but it could be a starting point for a discussion about new ways to do this.

[deathspeed]

Turey's old installer is was brought me to (and kept me in) the FSO community.  It installed everything I needed, and if I recall correctly, I could even choose which mods to install and/or update.  At the time I first used that, I would have been overly-intimidated to try to follow the current manual procedures, but it was easy enough to use that I got hooked and stuck around.  Yet, as fightermedic noted, I still get confused sometimes about which files I need to download or I am uncertain whether I am using the latest version of something. 

[General Battuta]

I'm an FSO power user, and I find the current install process an active turn-off to getting my friends involved...never mind them getting themselves involved. We need something at least as easy as Turey's.

Let's assume that Goob's installer is a long-term solution that will not be finished in the next couple years (sorry Goob, take it as a testament to your monumental ambition). Does anyone want to put forward a proposal as to what we can do? Could a single huge archive make an effective solution to drop onto an existing FS2 install? It'd be impossible to maintain.

[Bullhorn]

i don't think somebody is talking about standalone
rather a case of reducing a noobs usual way of installing stuff from:
- install fs2
- read through confusing instructions, get an up to date launcher and copy all the files by hand
- read through confusing instructions, get an up to date version of fso and copy all the files by hand
- read through even more confusing instructions, get the up to date mediavps (several seperate downloads, a noob will have no idea what he needs)
- select the exe and mod you want to use (read: mediavps) in the launcher
- have crashes because you installed freespace into a directory with a too long file path

to:
- install fs2
- install everything fso from one .exe, and be ready to go for playing with the latest stable fso and the latest mediavps
 (the installer should come with a BIG warning that freespace can crash from too long file paths)

(1) was **EXACTLY** my experience with my current, 2 month old install.
And I had installed FSO previously! A couple of years ago, but still - I already knew what the MediaVPs were.

To me, this issue takes precedence over nice graphics, scripting, AI, upgraded models, etc.
If we cant get new people playing fairly easily and quickly, then whats the point?

BH

[Bullhorn]

What about focussing on first things first?

ie,
1) FS/FSO installer, with appropriate stable MVPs.

Get a stable, up-to-date installation of FSO running, FIRST.
Never mind mod repositories, just get the damn game and its upgrade installed, FIRST, ASAP.

Hell, a simple batch script, or even NSIS to just GET THE GAME & ITS STABLE UPGRADE ONTO THE COMPUTER AS EASILY & QUICKLY AS POSSIBLE.

THEN move on to forever discuss/sporadically design/roughly build a mod friendly/universal installer.

BH

[Goober5000]

Geez guys, at least wait until after the holiday break before posting this.

[Mongoose]

I will take that statement with guarded optimism.

[Darkmage]

Cross platform modding tools and the WxFred port are more important IMHO. Installer support is nice, but once WxFRED and all modding tools are ported to Linux/OSX you should submit a post to slashdot and other news websites. It'll help promote/grow the community. You'll also be the first 3D space game in Linux history to have a working level editor. Submit screenshots/videos of the engine upgrades along with the news of the editor port in the links you give to phoronix.com and slashdot to get better coverage. Phoronix.com might make a negative comment about the engine's tech level, so if you can, upgrade it so it supports some features from OpenGL 4, even if no one really uses them. That will make the game get more favorable reviews, they will compare the game from 2002 as a game that can go toe to toe with the latest company tech from 2014 as most linux games coming out right now barely support OpenGL3 and there's almost no OpenGL 4 titles.

[General Battuta]

Even as a content creator with a big stake in improved, cross-platform versions of the tools I use, I don't think content creation tools are nearly as important as content consumption tools.

In your specific scenario, we can get all the positive press we want, but when it turns out that the game's an obnoxious bear to install, that press will curdle.

Additionally, getting the installer working is a lot easier than WXFred.

[The E]

As a developer, I would also disagree. wxFRED is an important part of our toolset, yes, and a priority for development. But making it easy for people to actually consume the content we're producing is far more important.