Creating a website for Danmakufu

[Helepolis]

Helepolis:
Me want own domain. D:
And also wouldn't dnh.shrinemaiden.org get kinda confused with dmf.shrinemainden.org?

Silly me forgot.

* Helepolis bashes head

[Blargel]

Okay better idea for the multiple versions of the same script organization problem. You create projects on the site that has a title, description, category, engine, and tags. After the page is created, you upload scripts to that page and they show up in a list with the newest one on top and in bold. The project's overall rating is the best rated upload in that project. If people only want the newest version to be available, they can just delete previous uploads.

So the set-up is: A user has many projects. A project has many uploads. An upload has many replays. A project will also have many screen shots.

[ChaoStar]

Don't forget the rating, commenting, and Hearting features~ *nyan*

[Blargel]

Okay, I just realized that my approach to this is all wrong (hey, I'm learning something, yay). I need to plan out the whole basic structure first before I can start building it, but I'm having trouble with that because I don't know what's a good way to organize it.

So far this is what I got.

User
  - Has and can create many projects.
  - Has and can upload many versions to his/her own projects
  - Has and can upload many replays to any project.
  - Has and can post many comments on other users, projects, and versions.
  - Has many comments on his/her own user page
  - Can rate many versions, but only once per version.
  - Can watch many projects and users (pretty much favoriting/subscribing)
  - Has many watchers

Project
  - Has many versions.
  - Has many comments.
  - Has many users watching it.

Version
  - Has one file upload.
  - Has many replays.
  - Has many comments.
  - Has many ratings.


That's about the gist of it. There's gonna be a lot more going on in the background in the actual code, but you don't need to know about that.

NOW, the reason I bring this up is for two reasons:
1.) You'll notice you rate the VERSIONS, and not the projects themselves. I was thinking that the projects would have two rating displays: Total rating (which is an average of ratings across all versions) and rating for the current version (which is the latest version's rating).
2.) If there's versions on this website, it'll increase the amount of hosting space needed.

Basically, a versioning feature on the site would make project rating organization kinda awkward AND may cause me to spend more money sooner. Does anyone have a better idea? Or perhaps this actually sounds like a good idea and I should start working on it?

For now, I'm gonna work on user authentication and logging in until we get a solid design.

[Zengar Zombolt]

I don't think hosting all the versions should be a good idea, actually. After all the latest one is supposed to be the best one, right?

[Stuffman]

Yeah I don't see any reason to host outdated versions. If you update a project, maybe do something to indicate that were made on an older version or something?

[Blargel]

The weird thing about updating a project without previous versions is this:

Let's assume SuperAwesomeScript has 150 downloads (boy, I wish my scripts got downloaded that much), 10 comments, and an average rating of 4.7/5.0 from 25 total ratings. When SuperAwesomeScript gets updated, obviously the comments can be carried over. Carrying over downloads isn't too strange either.

However, should you carry over ratings too? What if the first version had a few issues that caused people to rate it lower that the newer version doesn't have? The newer version would be rated higher than the earlier version, but the people who voted already can't change their rating because they already rated it. If you don't carry over ratings though, it would be strange to see a new version of a script suddenly only have very few ratings.

Carrying over ratings and allowing users to edit their ratings would somewhat solve the problem, but if a user rates it with a 0 and never checks it again because he totally didn't like it for whatever reason, that 0 is gonna be skewing the average even if the new version is good enough for that user to give it a better score.

EDIT: Oh also, bulletforge.org or shotforge.org? :V

[Stuffman]

There's really no good solution. I would reset ratings when a script is updated (and allow people to rate it again).

bulletforge.org might be the coolest sounding website I've ever heard.

[Lishy1]

The problem why the community is so small is that it is hard to jump in.
Unlike Lunar Magic and SMW which is mostly self explanatory and has tons of tutorials, danmakufu is just really hard to grasp at first.

I propose we get better documentation and promote newcomers first before we make a website. However, if you insist to make a site as soon as possible, I agree that bulletforge.org is awesome

[Prime 2.0]

EDIT: Oh also, bulletforge.org or shotforge.org? :V

Bulletforge, no question. You'd better register the domain before someone else snatches it up.

[Suikama]

Hopefully no one will confuse it with http://www.bulletforge.com/

[helvetica]

If your code is decent enough I would be more than happy to host it here under dmf.shrinemaiden.org or bulletforge.shrinemaiden.org or whatever you want to do (like if you get your own domain).  Triela has enough space and enough bandwidth to handle it (and I'm a competent admin o/)

[Blargel]

The problem why the community is so small is that it is hard to jump in.
Unlike Lunar Magic and SMW which is mostly self explanatory and has tons of tutorials, danmakufu is just really hard to grasp at first.

I propose we get better documentation and promote newcomers first before we make a website. However, if you insist to make a site as soon as possible, I agree that bulletforge.org is awesome

The thing about LunarMagic and similar things like Warcraft's World Editor is that there is a graphical user interface. That makes a huge difference about how easy things are to learn. However, a graphical user interface also tends to be more limiting than pure code which I guess the maker of Danmakufu was trying to avoid. That or since it was version 0.12, he didn't get that far yet. Anyway my point is, don't we actually already have a lot of people trying to learn considering how hard it is to learn already? I think a website like this would probably gather more attention and make the community grow even more.

Hopefully no one will confuse it with http://www.bulletforge.com/

Lol, damn it. I'm still gonna use it though <_<

If your code is decent enough I would be more than happy to host it here under dmf.shrinemaiden.org or bulletforge.shrinemaiden.org or whatever you want to do (like if you get your own domain).  Triela has enough space and enough bandwidth to handle it (and I'm a competent admin \o/)

I'm pretty sure the code is going to be at least "decent". It's going to be a Rails application though which I believe takes more resources to run than a Merb or Sinatra application. But anyway, my brother already has a service that hardly gets used (it only has two blogs that don't get that many views) that he said I can put it on for now.

bulletforge.shrinemaiden.org might be better than bulletforge.org too since it's less in danger of being mistaken with the .com and more obviously related to Touhou. Well first I gotta make it.

[CK Crash]

Possible system:

[Project: Game]
      [Scripts] (optional, in the event that a single script relies on many outside files)
            [Screenshots]
            [Ratings]
            [Comments]
            [Replays]
      [Screenshots]
      [Ratings]
      [Comments]
      [Reviews]
      [Replays]

[Project: Boss]
      [Screenshots]
      [Ratings]
      [Comments]
      [Replays]

[Project: Resource]
      [Screenshots]
      [Ratings]
      [Comments]

Users reuploading a script can choose whether everything related is retained. That way, if a replay desynching change is made, then the user can just decide to carry over everything else but the replays.

[helvetica]

I'm pretty sure the code is going to be at least "decent". It's going to be a Rails application though which I believe takes more resources to run than a Merb or Sinatra application. But anyway, my brother already has a service that hardly gets used (it only has two blogs that don't get that many views) that he said I can put it on for now.

If it's going to be Rails I'm already using LigHTTPd on Triela so it'd be like nothing for me to slap it on later if you want, up to you of course.  Advantages of us hosting include me as an admin, full SQL support from PostgreSQL, and the site can easily accommodate the CPU/bandwidth load.

Whatever you do please don't code the DB backend to MySQL because MySQL is garbage and horrible.  Flatfiles would be more reliable than MySQL honestly.

bulletforge.shrinemaiden.org might be better than bulletforge.org too since it's less in danger of being mistaken with the .com and more obviously related to Touhou. Well first I gotta make it.

If we host it or you find another host, let me know and I'll gladly setup the DNS entries and stuff for it.

[Blargel]

Okay, well I have no idea how to use LigHTTPd as a proxy to Rails. I used to use nginx instead so you'll probably have to help me with deployment. Other than that, yeah I guess I'll be taking you up on your offer then. And yes, PostgreSQL is awesome.

Oh yeah, how many Mongrel instances can you afford? I was hoping for two to four to start with. :V

[helvetica]

You don't need Mongrel, just some sort of FastCGI dispatcher for Rails like dispatch.fcgi.  Mongrel is for servers that don't support FastCGI or have crappy support (like Apache), as it provides a whole HTTP server stack you just proxy to.

Since I use LigHTTPd on Triela FastCGI works beautifully, you just need to use a dispatcher instead of proxying to Mongrel.

[Blargel]

You don't need Mongrel, just some sort of FastCGI dispatcher for Rails like dispatch.fcgi.  Mongrel is for servers that don't support FastCGI or have crappy support (like Apache), as it provides a whole HTTP server stack you just proxy to.

Since I use LigHTTPd on Triela FastCGI works beautifully, you just need to use a dispatcher instead of proxying to Mongrel.

Not too experienced in that, but it can't be that hard to figure out. :V (Technically I'm not experienced in deploying websites in general since I've only done it once with the help of someone else.)



Okay now for a progress update:

User login/logout is working.
Project creation is working.
Versioning is working. In fact, I force people to version because it helps with organization.
Script uploading is working.
Project/version commenting is working.
User commenting is working.

Project categories are not implemented yet. (Single, Plural, Player, etc.)
Project tags are not implemented yet. (contest, joke, brofist, etc.)
Project searching is not implemented yet.
Project/user watching is not implemented yet. (Similar to favoriting and subscribing on YouTube)
Project ratings are not implemented yet. (Need help on this, see below)
Design is not implemented yet. (Need help on this too, see below)


About project ratings: Making people rate with different categories is troublesome for both me as the developer because it takes some thinking to code and for the users because you have to click more than once or twice and possibly even type before you submit your rating. For the users, I think it might actually discourage people from voting just because it takes too long. For me, that's too much of a pain in the ass for something that will actually discourage users from using it. That said, I propose one of these three things as the rating system:
1.) Thumbs up/thumbs down
2.) A number from 1 to 5 or 1 to 10
3.) Let the download counter do the talking on popularity

What do you think?

About design: I need someone with a relatively good sense of design to draw some mock ups of the pages of the site (in Illustrator or whatever). Currently, everything is in plain html with no CSS and it looks like shit. Just post in here if you're interested or PM me and I'll give some more details on it.

[Drake]

Download counter + ＃/10 rating system.

[Blargel]

Feels like not many people care about this. :V

If only one person is gonna leave an opinion about the voting system, I might as well just do whatever the heck I want with this. Which is to voting up or down on something. Simplicity is always better in my opinion. Does it really matter if there are some unfair votes in there? Even if something gets voted up because of a silly biased reason, how many of the other voters will have the same opinion? If most people vote down, then it probably sucks and isn't worth your time. If most people vote up, then it's probably good and you can take a look.

Also, this is the easiest method to implement besides just letting the download counter tell you how popular it is.

[Naut]

You know that RG is slow :V

Download counter + ＃/10 rating system.

^, unless you really want to do it the easy way.

[Blargel]

Naut, that short bit of IRC conversation we had about the review plugin that I found is now void. That plugin is the buggiest thing I've seen in a while and the documentation sucks. No way in hell is this thing going to work correctly. That said, I'm afraid I'll have to actually do the voting up and down plugin because it's the only one that's not overly complex or limiting. Guess I should've researched my options better.

Actually, before that, I can always write my own rating system thingy instead of installing plugins... Hmm... I think I'll wait till there's a few more opinions in here on which rating system I should put in.

[helvetica]

Thumb ups/down or a simple plus/minus system is all you really need anyways.  Tack on a way for people to add comments/reviews for detailed explanations or questions about it.

Honestly a 1-5 kind of system is pretty useless as most people tend to vote extremes.  Plus law of averages tends to make more popular things seem worse than they really are, and less popular things tend to have overinflated ratings.  Compare a script that has an average rating of 3 or so over 100+ votes to a script with an average rating of 4.5 over 10 or so votes and chances are the 3 star one is better.  You could do weighted averages based on number of downloads to adjust for inflation but then it just gets too complicated.

Just do what Danbooru does and have a score for each script.  Vote up or favorite adds 1 to the score and vote down subtracts one.  Gamereplays also does something similar with a "wub" system where if you like the replay you vote it up and if you don't then let it be.  Restrict voting to only people who have downloaded and registered to prevent abusing the system.  I'll gladly figure out a way to link into the login DB for the forums, saving people the trouble having to reregister and remember another login.

[Blargel]

Yeah, the reasons you give for the voting up/down system and against the 1-5/1-10 scale is quite similar to mine. I did notice people tended to vote in extremes because most people just think "I like it" or "I don't like it". Most don't really wanna think about how much they like it. The overall trend of the internet is moving towards the usage the voting system over the rating system too anyway.

About restricting voting to people who downloaded the script and are registered, I can limit it to registered users, but the only way to detect if someone downloaded something is to see if they clicked on the download link. That's not exactly gonna enforce that the person clicked Save, or if he did, that he played it so that feature is pretty much worthless. We'll have to go with the honor system for that one.

As for linking the login databases... yeah, that's not worth the effort in my opinion. It's only gonna take 3 fields to register for BulletForge anyway: Username, password, and password confirmation. Not even a captcha (that the user is gonna see anyway). If you insist though, I wish you good luck. It seriously probably isn't worth the trouble considering how many people use Danmakufu out of all of MotK.

EDIT: I have the greatest layout ever!

...Seriously, anyone good at design? D:

EDIT2: Hey, I know this is just a visual thing, but apparently implementing an upload bar that shows how much has been uploaded makes no sense to me is kinda hard to do? Or I just don't know what the hell this page is talking about. If it's too out there, it doesn't really matter. We don't really need an upload bar when most scripts are gonna be under 1 MB. Only the people uploading big games with lots of music are gonna be bothered by a lack of progress bar.

EDIT AGAIN: Nevermind, apparently it's pretty easy if you look at this. I just need you to configure Lighttpd for it to work.

[ChaoStar]

Lol. "how to print screen on mac" Congratz.

[Blargel]

Lol. "how to print screen on mac" Congratz.

Who the hell thought that Command+Shift+3 was easier to use than a single Print Screen button? I thought Macs were supposed to be user friendly. 

[Blargel]

Megabump!

So yeah... pretty much everything that's needed is added and working, but the site is ugly as hell. I'm trying to group information together in a coherent and intuitive way, but it's still gonna be a really plain looking website until someone can give me a decent design. So lemme repeat that:

I STILL NEED SOMEONE TO HELP WITH DESIGN

I'll be more detailed in a private message once someone actually steps up to help me.

[shadowbringer]

just popped in to say that it would be nice if the users (submitters) chose to provide screenshots (amount undecided) for their scripts, not just the scripts themselves for download. That way, people can "judge the book for its cover", and the submitters would have to have an interesting enough content to show.

Personally, I've had many trial/demo software before, that's what (I think that) left me skeptical about just downloading things (fangames, fan music, etc.) that I can't see first. (I don't mind watching videos, though)

[Blargel]

Already done. Just gimme a design.

[shadowbringer]

personally, I have no experience in creating designs for webpages, however, if you have, you can use these two sites for inspiration (more exactly, when I read this thread, I thought about something similar to the second link):

http://shmupthecore.blogspot.com/
http://web.archive.org/web/20080205203904/http://shootthecore.moonpod.com/

(RIP shoot the core )