Wesnoth Online World?

[moraes]

Since you're starting now, consider using Google Code for the repository hosting. Issue tracker is far more usable, and you can use Mercurial instead of Subversion.

[bonndan]

What's wrong with SourceForge? I've been using it ever since, and there is already a checked SVN repo, with externals merging the Wesnoth SVN repo. Is Mercurial so much better? In principle I would consider changing it, but only if it's really much better.

And besides, this was not the question

[Aethaeryn]

I've heard from frequent users that the Sourceforge website gets progressively worse and user-unfriendly every time it redoes its website's GUI (especially the administrative options), and it frequently reinvents itself. If you prefer long-term stability over change, you might not like Sourceforge that much. Personally, I've had no problems with it since I mostly access it via a Terminal to do SVN downloads/uploads/updates, but the negative reactions I've heard would be enough to make me think twice before willingly choosing to administer a project on Sourceforge.

One additional consideration is that Sourceforge is very popular, which probably means that some competing sites might offer more services in an attempt to attract projects (since it would be harder for them to do so). Often the second-largest website/service/etc. is better than the most popular choice simply because it's trying to take over the number one spot. I'm speaking in general terms because I haven't read enough on the issue to make up my mind on which is the best, but I would recommend researching a bit first. It's more of a hassle to move an existing project once you've made the choice, even if you later find the website frustrating.

I cannot evaluate SVN too effectively, other than to say that it's better than CVS (which it was designed to replace). git is supposed to be superior in just about every way except friendliness-for-Windows-users, but I haven't used git enough to get used to the way it does things (which is different enough from CVS and SVN to be somewhat confusing for me). I'd probably choose Subversion for a project just because I know how it works without having to consult others for help, but it's all up to personal taste. I doubt the choice of revision control systems matters much for small projects, and it's probably more important in projects with lots of activity or complexity.

The bottom line is that choice is good and if you read up on what you want to use before you host the project, you might not have to worry about the hard work of moving the project later on if you make the wrong choice. (It's all about personal taste, but make sure you make a decision that fits your personal tastes that you won't regret.)

[bonndan]

My impression of SF is completely different. It might not offer the best user experience, but imho it has gotten a lot better since when I started using it in 2002, especially in the last year.
I use and like SVN for my private projects and at work, and the one time I tried git it really frustrated me. When I googled for "Mercurial vs SVN" most people wrote that they are both equally good. So, until further decisions the project stays at:

http://sourceforge.net/projects/irdya/

However, this was not the question. I'd like to set up a Wordpress for multiple users to write about plans and development of Irdya. But where should it be done to get as many people as possible on board?

[Zarel]

Since Warzone recently moved from Gna to Sourceforge, I think I may be able to provide some input.

Note that this is only about the SCM and file download hosting; we already use Trac for our issue tracker, development wiki, and source code browser.

The main objections to SourceForge was that its web interface was bad, and its website had tons of ads. But the "bad interface" part is subjective, and personally, I think it gets better with each change.

The main objections to Gna was that it used a self-signed SSL certificate (making browsers pop up a warning along the lines of "THIS WEBSITE IS TRYING TO SCAM YOU GO AWAY"; Firefox being by far the worst offender), that its spam blocking was overzealous (we often had developers who couldn't commit because Gna thought they were spammers).

The main objections to Gitorious is that SVN has far better GUIs, such as TortoiseSVN. We had two developers on Windows at the time (and one or two others on Linux who preferred SVN to Git for unknown other reasons), so Git was unacceptable.

The main objections to Google Code was that they took about a month before they even replied to "Our repository and files go over your limit; can you give us an exception?" and they said "No, we can't handle files that large."

[moraes]

But where should it be done to get as many people as possible on board?

Anywhere. Just let people know how to find it and how to contribute.

The repository thing was just a minor suggestion. I didn't intend to discuss the color of the bike shed.

[bonndan]

But where should it be done to get as many people as possible on board?

Anywhere. Just let people know how to find it and how to contribute.

The repository thing was just a minor suggestion. I didn't intend to discuss the color of the bike shed.

To all: Thanks for the input. This is the outcome:

- Development will be located at SourceForge (stays as is)
- Project public address site will be Wordpress at http://stack76.net (stays as is)
- Bug Tracker + Feature request will be collected at SourceForge https://sourceforge.net/tracker/?group_id=292414
- Forum stays here at Wesnoth for now.

Anyone interesting in contributing is hereby kindly asked to sign up at SourceForge and/or PM me to get a Wordpress account at stack76. I hope the discussion continues here as productive as before.

This thread should maybe closed soon in favor of a new Irdya design thread.

Kind regards

Dan

[ahyangyi]

Since Warzone recently moved from Gna to Sourceforge, I think I may be able to provide some input.

Note that this is only about the SCM and file download hosting; we already use Trac for our issue tracker, development wiki, and source code browser.

The main objections to SourceForge was that its web interface was bad, and its website had tons of ads. But the "bad interface" part is subjective, and personally, I think it gets better with each change.

The main objections to Gna was that it used a self-signed SSL certificate (making browsers pop up a warning along the lines of "THIS WEBSITE IS TRYING TO SCAM YOU GO AWAY"; Firefox being by far the worst offender), that its spam blocking was overzealous (we often had developers who couldn't commit because Gna thought they were spammers).

The main objections to Gitorious is that SVN has far better GUIs, such as TortoiseSVN. We had two developers on Windows at the time (and one or two others on Linux who preferred SVN to Git for unknown other reasons), so Git was unacceptable.

The main objections to Google Code was that they took about a month before they even replied to "Our repository and files go over your limit; can you give us an exception?" and they said "No, we can't handle files that large."

I think there's also something called TortioseGit well, I'm not actually using it...

[Zarel]

I think there's also something called TortioseGit well, I'm not actually using it...

I said "far better GUIs". TortoiseGit is somewhat usable, but it's far behind TortoiseSVN.

[Rakarei55555]

An MMORPG set in Wesnoth would be the coolest thing ever.

It would be set on the eve of Konrad and Lisar's takeover, one of the biggest quests in the game would follow that story. other campaigns could be added later as new big quests.

quests in this game would have to be divided into two categories, normal quests which don't follow the already defined story of wesnoth, and story quests, which would follow the story of wesnoth. there could even be ways to go back in time and have a rise of wesnoth quest.

it would most likely be pretty traditional MMO style, but maybe each person could have one hero and a few units that follow them around. the higher the level the more followers they get.

[dipseydoodle]

Is the game playable yet? I just signed up and logged in and couldn't do anything.

[EvilEarl]

Amazing project, I recommend that you update the first post on this thread so that it looks good at first glance.

Doing browser-based for the early development is good, but you might want to consider it to be downloadable (for at least the graphics) to prevent lag and increase loading speed. If you stick to Wesnoth-grade graphics then a downloadable pack wouldn't be killer in size (please note that I am not considering music files in this system, unless you want to use midi files) and the actual engine probably won't be much either.

I assume this is GPL, and if it is then I don't want to see some system in the final product where you get a special currency for special stuff by donating. I want to see an online game where free as in everything is free to play. You should still have a donate system for server costs (when it comes to that) but otherwise you'll just annoy people and really tick off the people (like me) who are used to either absolutely free or buy-the-whole-thing-at-once systems (an in between doesn't seem right for that.)

For mapping, how about large scale movement on the main map, and then it loads a separate zoomed in map when you enter the "city limits" of an area, and then a smaller map for directly in the area of buildings/ruins/tents. And for teleporting, make it few so that walking isn't obsolete, but make enough the cross-country movement is required. Faction teleporters that are mobile should have some restrictions on them so you can't move whole armies to a spot within 10 seconds (without warning), otherwise I see ambush wars resulting from this.

As for the acronym (WoW), I say keep it for as long as you can and see if you can draw in any World of Warcraft junkies into this to help test/develop (I'm sure some of those people dream of being part of the creation of something in the essence of WoW, and I'm sure some would pay to help if it wasn't free to modify). If you have to change the name so the acronym changes then you should make it something close enough to still draw over WoW junkies.

Before I finish the biggest post I've ever made, the big question I have is continuity with Wesnoth? Will the mapping be as similar as possible in terms of world map? Will there be quests made to replicate some of the campaigns? Will name locations be the same? I understand that the timeline isn't always the same (wasteland in the far future isn't as pretty as the HttT period), but you could probably get away with as sort of "overlapped time" of similar periods to make things work.

Now goodnight, I need to pass out from all of this thinking. I'll give more feedback as it comes to me.

[axyd]

Try github.com (hosting for Git source control) it's exceptionally good. It's free for open source projects.
Though it's not very popular outside of Ruby community.

[Jetrel]

Git/Mercurial/Bazaar (distributed SC) versus SVN (centralized) basically boils down to some preferences about how you want to do source control. I'm still figuring this out myself, so please correct me if I'm saying something horribly inaccurate.

With Git, basically everyone has a copy of the entire source tree and all the history. Instead of there being one single "canonical" version of the source code, all the major developers are living in their own world, per se, and they each have to somehow "post" access to their code out on the net, which places like github help out with.

Since you obviously want all the developer's contributions together in one build, what you do is git lets you "pull" changes from some other developer. For example, if wesnoth was using git, Ivanovic would probably occasionally "pull" from my source tree to get all the latest graphics. If pulling succeeds (which seems to be the black box in git that no one can explain to me, and I'm confused as to how this can work well), then some or all (depending on your choice) of your two source trees are brought in sync.

It seems like part of the dirty trick might be that if someone else's code is incompatible with yours, pulling fails, and you force them to fix their code to match yours, rather than receiving a random patch and having to deal with it yourself. So once you've pulled from most of the main contributors, you've got fairly up to date changes, and can build a binary out of them.


It seems great for playing around with code, and for projects that are forking like crazy (e.g. linux, for which some guy practically makes a new distro each day). It seems awful for projects where you want one, single "canonical" version of the program.

It also seems like a royal pain in the ass to track down other people's changes, since they (AFAIK, could be wrong) can't do the job of committing to the canonical version for you. It's like someone has to take the time to be a full-time manager of finding stuff and putting it in the canonical version, rather than just giving people access, and having them completely doing without any involvement on your part. So I guess it would make sense when a project is big enough that there are people already doing nothing but that.

[pauxlo]

It seems great for playing around with code, and for projects that are forking like crazy (e.g. linux, for which some guy practically makes a new distro each day). It seems awful for projects where you want one, single "canonical" version of the program.

It also seems like a royal pain in the ass to track down other people's changes, since they (AFAIK, could be wrong) can't do the job of committing to the canonical version for you. It's like someone has to take the time to be a full-time manager of finding stuff and putting it in the canonical version, rather than just giving people access, and having them completely doing without any involvement on your part. So I guess it would make sense when a project is big enough that there are people already doing nothing but that.

I'm using git in a project now for some months, and the thing is, you can use it like a centralized VCS too, with the additional advantage that you can have offline commits to your own repository.

In our case, we have a repository on the server, and all 3 developers (5, if you count my two accounts and the production system separately) have push-access to it from their local repositories. Before doing a push, you have to do a pull, merge all changes occurred in the central repository in your repository, so the push automatically works.
It is like doing your work on a (private) branch and then merge it back to trunk.

The main advantage compared to Subversion then is that branching and merging works really good, and you can have offline commits.