pouët.net

Go to bottom

Uber scene database, opinions wanted

category: general [glöplog]
Hello. I'm not Optimus, despite what you might think from this post.

For the last few months, I've been privately plotting to take over the demoscene with an enormous new database-driven site, bigger than anything that has gone before. Think Pouet, but more detailed (down to the level of individual author credits), wider scope (not just demos, but standalone graphics and music too) and more complete (full compo results, rather than just the entries people consider worth commenting on). And with Slengpung, demoparty.net, Nectarine and whatever else I can think of thrown in.

Before you start screaming "No! Don't fuck with Pouet!", hear me out on this.

My main reason for wanting to do this is that I don't really have the time to single-handedly maintain zxdemo.org any more (look at it this way, if the stuff I've done on that site were translated to glops, I'd be well into five figures by now), so the obvious answer is to rewrite it into a pouetized collaborative thing, and add all the features that people have requested while I'm at it. But if I'm doing all that work, why should I limit it to the Speccy scene, when I could make it a lot more useful for a lot more people (and attract a lot more contributions) by covering everything?

Besides, the stuff you could do with that data, the questions you could answer, would just be awesome. Things like: when's the next party with a 256b intro compo? Who's the most successful MOD musician by number of compos won? This guy sitting next to me at Assembly, has he worked on any demos that I've heard of? How many sceners live within 30 miles of me (and do they want to meet up for a beer)? Or how about getting a podcast of the most recent tracks to win music compos. Or plotting a map of scene activity to find the demoscene capital of the world.

Yes, it's stupidly ambitious, but I'm determined to succeed or die trying. I've kept this project private up to now, partly so that it wouldn't be labelled vapourware while I have nothing substantial to show, and partly because I had wild hopes of launching a fully working site out of nowhere with huge fanfare, just to show how l33t I am. Finally I've realised that keeping it to myself is a fucking stupid idea, for the following reasons:


  • One person trying to map out the scene is like the proverbial blind geezer fondling an elephant - there's too much of it to understand in one go. If I create this with no input from anyone else, it'll only represent my perspective of the scene.
  • I need some people with a sense of proportion to give me a good kicking when I start obsessing over pointless shit like "ooh, how can I make this 4K procedural graphics compo count as both a graphics compo and a 4K intro compo?".
  • If people know about it and expect it to happen at some point, I'll be more motivated to finish the damn thing.
  • If everyone convinces me now that it's a really, really dumb idea, I can drop it now and not waste so much of my life on it.


I should make it clear that I'm not trying to kill off Pouet (or any other site for that matter) - I don't think Pouet intends to ever become a complete encyclopedia of scene activity, and quite rightly so. I don't expect the entire community to migrate en masse to this new site either - so I'm sure the two sites can coexist happily. Alternatively, if people think it would make more sense to join forces, I have nothing against doing that.

So. Does anyone here share this dream, and have a burning passion to argue over database schemas (and more interesting things, such as coming up with a name for the damn thing for a start)?

Thank you for reading. You may now tell me how great or stupid my idea is, offer to help out, or fill this topic with Japanese goat porn and then amuse yourself at Breakpoint 2012 by going up to me and asking "So, er, how's the scene database thing coming along?" and laughing manically.
added on the 2006-01-12 02:10:00 by gasman gasman
don't forget to come to InerciaDemo 2012 :P

Stupidly big thoughts man, so i commend your bravery.

Initically my reaction is not good.. mainly because um; i don't really think it's neccesary... or rather, i think the decenralised organisation works better. This is because in reality i think that the scene is best described as a whole lot of smaller 'scenes' - like the demo.music scene (nectarine), the techy-conversion dudes at demoscene.tv, news, community, party organizers, etc.

The fact that there are a few ubergods whom take charge at the most of these sites isn't the point - it is that one can pick and choose their specializations, and get involved with whatever they want to, and only that. Hence you get specilized people only doing what they want; and its good.

The other reason why this is headachey, is a physiological one. Think of the stupidly huge amount of resources you would need to host it, let alone keep it online. Sure, you might have something in mind that is similar to DeviantArt.. but lets face it; it's going to need people and probably money.. which means ads and that means lamerscene.

And should all this monstrosity take place, what are you left with? More confusion in where to go for demoinfo, and intense data replication. leave it be, and save some power.

But I am just a minon so you'd have to ask the bigger pouet gods..

added on the 2006-01-12 02:24:14 by Gaia Gaia
Quote:
I don't think Pouet intends to ever become a complete encyclopedia of scene activity, and quite rightly so

But... do YOU?

Quote:
i think the decenralised organisation works better

I agree... Specialized sites seem to work better in this case, if only by having specialized admins.
added on the 2006-01-12 02:37:59 by Gargaj Gargaj
gasman: It's not just the size of this post but the monstrous idea behind, that reminds me :)

And I like it. There are many cool features missing from Pouet database, I doubt we will see (CsDB-like Demos with real credits connected to sceners(users) in the database). I hope you will finish it oneday and more sceners support or approve the idea, cause I don't think much people would suggest monstrous plans that may still make no sense for some.

# If people know about it and expect it to happen at some point, I'll be more motivated to finish the damn thing.

That was my way once, till I felt so strange making so much hype but producing nothing, that I stopped. Instead, I sometimes say "Hey, I have some crazy ideas, but they are just ideas for inspiration! Aparrently I will never find the time to dedicate on such big projects.." (Like my crazy idea for The Sins computer game I once wrote in some older forums here =)

# If everyone convinces me now that it's a really, really dumb idea, I can drop it now and not waste so much of my life on it.

I hope not :)

p.s. Something similar that I thought. I had once the idea to make the ultimate demoscene DVD, it would have a main programm which will be like a database, the DVD would have all demos from 8bits/16bits and several PC demos till a year, the user could add stuff in the database and someone could ask questions in the database parser like "Show me a list with all the demos till 1996 with voxel landscape routines and plasma effects plus Boris Valejo pics and using only GUS" or several other things too. It would also connect to launchers like emulators, DosBox or plain launching of EXEs, reconfiguring memory, etc. Wicked! Also, I was thinking if my database programm could connect to the pouet net and update it's entries somehow? The ultimate demodatabase in one DVD with several demos and the main programm. Users could update the most wicked data for demos (if there are gfx of Jeniffer love Hewitt in that demo, demofuckings, or use of volumetric effects) and the database could be updated with these data coming as files or downloaded from a website. Just a crazy idea I had in one morning and reminded me very much at some points your project :)
added on the 2006-01-12 02:52:32 by Optimus Optimus
iddb or isdb vs Pouet ?


prolly a lot of dupe work, but adding gfx compos to Pouet would be nice :]
added on the 2006-01-12 03:25:21 by Zest Zest
In my opinion, since currently SceneID works rather nice in interlinking persons, Slengpung and DemoPartyNet share parties, Pouet is the definitive site for productions, Scene.org works as a file repository and OJ more or less serves the news, there's no real need to make this one big mess (OJ tried, see the results...)
Covering the graphics/music competitions would be great however, a "SceneMedia" site would be greatly appreciated, imho.
added on the 2006-01-12 03:30:03 by Gargaj Gargaj
gargaj: gfxzone.planet-d.net was supposed to be doing the gfx, and scenemusic.net the sounds. but one is innactive and the other abit skewed from justc overing music compos (which btw lets face it, dont even deserve beeing covered like that couz no one cares about the tracks released at the music compos)

gasman: ever considered joining scene.org and develop pouet.net to be the scene repository a-la zxdemo/csdb that you envision? you say that pouet doesnt intend to be a complete scene database, thats false, atleast when it comes to all prods, we do want to have everything listed, just people dont bother adding certain stuff and we're too little active admins to cover all the holes.. (not to mention that we get called glopwhores by the authors)

i think the main issue here is integration between the diferent databases. we need to find solutions for cross-matching the information to avoid redoing the wheel. if you're willing to work on pouet to extend it to incorporate the zxdemo style of prod association (prod credits and having all sceners despite them having an account or not) it could easily be extended to assimilating the info from csdb the same way, would only need for a system to submit the missing info, which for smaller scenes like speccy and c64 is doable "by hand" but for pc scene is your worst nightmare ever, you can never know _all_ groups and _all_ sceners.. there needs to be a way to let the users associate and revise each others associations... the +zxdemo and +csdb and +slengpung were approaches to unify the data. but there are still shitloads of holes to be filled on pouet..

the short version of what im trying to say is: i have thought about what you're trying to accomplish before gasman, i would love to see it, i have no time to implement it on pouet, if you wish to get any extra info from our database to use on your project or wish to join our dev team to push it more into your idea of a complete database, we're open for suggestions. we just lack time, not interest. :)
added on the 2006-01-12 06:27:31 by psenough psenough
if pouet had a competitor, would that force both sites develop faster? i guess so, temporarily, but the practice in the last 10+ years was that scene takes only one dominant forum at a time. just like vhs vs betamax, it's standardization! ;) product archive was hornet, then flerp, then scene.org, and portal was oj (now dead) then pouet, not to mention previous textmags/diskmags. so such a new portal would either die or kill pouet after a while, unless they split the tasks somehow. but then working together is better.

filling the gap that gfxzone left is a good idea. this cross-linking prods-to-artists-to-groups thing is nice but i miss other features more from the current sites' palette. first we need an online magazine, where you could even publish technical articles (gamasutra-like) - and pain doesn't seem to become a forum for serious articles (diskmags are dead?). the second is a serious discussion forum, where flamewars or pouet-style linking-funny-images would not be tolerated. don't tell me scene.org forum, no one reads that.
added on the 2006-01-12 10:01:55 by Ger Ger
gem: quite the oposite, when a worthy replacement comes we stop feeling the need to develop, couz it's already getting covered properly and we never have time for this stuff anyways, would much rather use that time in making prods instead of databases and php and articles. atleast thats how i feel about it. dunno about the rest.
added on the 2006-01-12 10:21:33 by psenough psenough
I do think its a good idea, but morely content/datastoring when it comes to archiveing partys/results bound to people or sceneids.
added on the 2006-01-12 12:35:02 by Hatikvah Hatikvah
of coz its not a bad idea.
its just plain stupid to try to make another page instead of joining forces and improve the exisiting stuff (aka pouet).
added on the 2006-01-12 12:50:19 by xeNusion xeNusion
pouet could be this indeed, get rid of the stupid glops, put a bit more restriction here and there and you will have a fine site with all scene information needed. Imho
added on the 2006-01-12 13:19:17 by okkie okkie
hmm... tough one this. Its definately a good idea. There's a few things i'd like to be able to do without so much farting around, like being able to see a list of other prods a musician worked on if i really like a tune in a particular demo.

Replicating what's already here would be pretty pointless, and having 2 very similar sites would be too, but there are things that would justify a second site. Eg. the forums - more serious forums with no inline images etc. would be great, but we'd lose a whole lot of fun if that was done on the pouet forums, so that should be kept separate.

How about having 2 separate sites sharing the same database then, with pouet being mostly as it is now, and the new site being a more serious/detailed version?
added on the 2006-01-12 13:54:02 by psonice psonice
On one hand i'm with gargaj on this. A separate 'scenemedia' site with the gfx/music compos would be better than integrating that into pouet. Despite i'm a graphician, i don't really keep track of the gfx-scene, or neither the music-scene. And i guess i'm not alone with this. From time to time i download the top3 compowinners, but it's not like i keep the gfx's and watch them again and again as with demos.
Pouet has been always mostly about demos, so adding all this to it (with all due respect to the graphicians/musicians) would spoil the thing a bit, although you could obviously implement some prodfilters, but i guess the database queries would be hellofalot slower.

On the other hand ps is right, why not join the devteam and help making the todo list smaller? :)
added on the 2006-01-12 14:13:05 by zoom zoom
I would like to remind people here, and I am sorry to have to do it publically that pouet.net is *not* a production of the scene.org staff.
added on the 2006-01-12 14:19:22 by _-_-__ _-_-__
The internet is closed. Please go away.
added on the 2006-01-12 14:27:34 by Jcl Jcl
über database would be great to have. go gasman go!

also, one thing that would be needed is a website for non-sceners who are intrested about scene. i don't think any of the scene-sites are so easy to use if you just happened to hear about scene and then go looking for what is it. though, all non-sceners are lamers so who cares about them.
added on the 2006-01-12 14:50:03 by nosfe nosfe
Quote:
I am sorry to have to do it publically that pouet.net is *not* a production of the scene.org staff.
Only with the hacking by some scene.org members pouet became what it is today. The time when I became a regular user was the time when I saw pouet getting better and better.

gasman: I like your idea, but having an all in one looks like serious overkill. Some API-like interfaces from/for all that scene portals would be cool, your idea could become reality that way.
Sure your idea sounds great. But as previously said it's overkill to begin from scratch. Gargaj and PS suggestions make sense. A SceneMedia site could be really interresting, and you could also join the Pouet dev team to improve its granularity.
added on the 2006-01-12 22:30:32 by p01 p01
sounds great, but i would rather appreciate an up-to-date zxdemos.org.
added on the 2006-01-12 22:56:25 by dipswitch dipswitch
gasman You could start by working on the music/gfx part of things. Once you have that working nicely, you can either work on integrating it into pouet, or you can keep it standalone, or you can decide if you want to continue with the effort and incorporate things that are already served by pouet, oj, etc. Having done the work for gfx/music, you would be a in a much better position to estimate the time required for you to get the ultimate encyclopedia completed.
added on the 2006-01-13 01:22:05 by legalize legalize
Yay! Thanks for the thoughts everyone, really appreciated. I'll be here for ever if I answer everything individually, so I'll try to summarise...

decentralised organisation vs ubersite: Yep, the existing sites have the main bases covered very well, and I wouldn't want to reinvent the wheel. A major purpose of scenedb (to give it a working title) would be to serve as the glue between the current sites. We have a limited amount of 'glue' already, such as sceneID and the zxdemo / csdb / slengpung IDs in Pouet, but I don't think that approach is going to scale upwards - I reckon a unified database is going to be the only way to get the right sort of architecture in place.

To give a specific example - if we want to start adding individual credits to prods (where prods = demos or standalone media) then Nectarine has its own way of doing that already, while Pouet or scenedb will have to start linking Pouet demo IDs to SceneIDs (or something like that). A simple query like 'show me all works by Gasman' involves a load of ID juggling and connecting spaghetti. Load all the prods into a standardised table and query that, and it makes life much easier.

duplication of effort: I plan to import / syndicate data from existing sites, wherever the maintainers will let me, so the important data will still originate from the respective experts in each sub-scene - no need to worry about redoing it all from scratch. As ps says, it's technically quite a messy problem to cross-reference multiple data sets - it's something I've spent a lot of time pondering over lately, and while there's still a lot more pondering and discussion to be done, I'm confident that it can be done in a reasonably automatic way.

joining forces with pouet: Cheers for the positive response on this one! ps, I'll drop you a mail some time soon. At this point I'm not sure whether scenedb could be a direct extension of Pouet - from what can I observe of the Pouet db from the outside, there are certain design choices (e.g. finite number of download fields per prod) that are tuned for running the site efficiently rather than strictly 'proper' design. Right now, I think I'd prefer to create a schema from scratch (and steal all the concepts from Pouet that work particularly well) rather than being constrained by the existing structure. I do agree that extending Pouet would also be a good thing for me to work on, though (and it's good to know that the Pouet team would be happy for that to happen) - if any sort of data migration is going to happen, having the sites meet in the middle is kinda helpful.

psonice's suggestion of two front-end sites sharing one database is pretty much spot on what I had in mind - scenedb would be a database first and a website (or websites) second. If Pouet were to become one of the front-ends, it could either use the scenedb database directly or keep its current one, with some kind of two-way import going on.

Somewhere along the line zxdemo.org would probably become another specialised front-end of scenedb, so all of this will be a route towards an up-to-date zxdemo.org, albeit a rather long one :-/

forums for serious stuff: It's certainly too early to say where scenedb will fit into the scene socially / 'politically' - it's the sort of thing that just happens by itself rather than being engineered. It sounds like it would fit in with the general tone of the site though ("Pouet's trainspotter-ish big brother"?) so I guess it's something for the wishlist.

hosting would require huge amount of resources: I'd say the scene has no shortage of the sort of resources we require (techy skills, enthusiasm, generous people with servers, people to buy you beer). It would only be a problem if the site grow to cover stuff outside the demo scene, and I don't propose to do that. Fair enough, it's harder to regulate that sort of thing when it comes to standalone music and every random netlabel wants a slice of the pie, so we'll need some arbitrary cutoff point to determine 'what is scene'. I have some ideas on that, but we can cross that bridge when we come to it.

catering for non-sceners: One of the biggest criticisms of zxdemo.org (and one I agree with) is that there's nothing to welcome newbies and point them towards the best demos (unless they find their way to the party results section, which is still only slightly helpful). I didn't want to just pick out some personal favourites, so right from the start I had vague intentions of adding a rating system. It never happened, and it just set me off on this endless quest for new features instead...

"show me demos with voxels and Boris Valejo": What about "show me all demos where I'm greeted"? :-) I dare say it can wait until we run short of more critical data to enter...

ergh. tired. Hope some of the above makes sense.
added on the 2006-01-13 03:02:36 by gasman gasman
I haven't read your posting yet, but I've seen once again the hosting question. Write some facts/asumptions down, go to untergrund.net and get in contact with scamp.
gasman joining the pouet forces, that's a better idea. If you have so much enthousiasm, then perhaps it's better to help things going on here, since Pouet is teh shit!
added on the 2006-01-13 03:18:14 by Optimus Optimus
gasman: some of your ideas ive been trying to expand on pouet for exactly same reasons:
- the demoblog.php for letting newbies surf the veterans tastes in a "what do people who liked this prod also like" kind of way
- the lists.php to do cross aggregation of information that wouldnt normally be listed (listing demologies and country sceners is just the start, the plan goes to extend it to smaller details, like all demos containing BASS or boris vellejo, and then supporting that info on prod.php so that people checking the prod can see some sort of description by lists its part of)
- the aggregation from several scene sources.. matti has been working on sceneID, and i've been allowing the sceners to add themselfs the corelations to the other sites, so that technically it would already be possible to have global sceneID's and partyID's and prodID's and facilitate the extraction of all the info. we've also been trying to facilitate it's export (via pouet's xnfo rss feeds, and the sceneID project itself)

so, i think we have the same vision, just that we are not enough people on scene.org to implement it, and pouet.net devteam is next to dead, it's me and matti who keep pushing the development despite it not beeing our site..

and about the download links beeing limited, that is already on the todo list of things to be fixed as suggested by nosfe before, i just havent had time for it yet.. we hired seffren to scene.org to lower the php upkeeping load on scene.org stuff so we could focus on our (me+matti) plans to unify the databases more, but seffren has been MIA and we're still too few people to be able to pull it off anytime soon, with the org of the scene.org awards also stealing some time, and trying to make material for new demos and our music projects.. we're people with too little free time for the dimension of the project. so i think i can speak also for melwyn when i say that your willpower to push this further is most welcome and we'll facilitate in whatever way we can (sceneID, access to pouet's prod db, slengpung, etc) drop by ircnet #scene.org if you wanna discuss things realtime :)
added on the 2006-01-13 06:33:04 by psenough psenough

login

Go to top