[game_preservation] Game Database question

[Jan Baart]

On 15.03.2010 02:11, Jim Leonard wrote:

> And before those individual releases, there was a single release that 

> contained both.  See the package variants here (take note of the 

> requirements sticker in the lower left corner): 

> http://www.mobygames.com/game/kings-quest-v-absence-makes-the-heart-go-yonder/cover-art/gameCoverId,43069/ 

>

> It's hard to see, with the watermark, but both sets of disks are 

> enclosed as well.


Yeah, I know. Was just pointing out that there actually also were EGA / 
VGA only boxed releases.

>

> Sierra was constantly reworking things, even their program releases.  

> I have seen three or more original disks of the same AGI or SCI game, 

> all with different version numbers and slightly different interpreters 

> (no discernible difference in the game itself though).


Yep, own a couple of updated releases myself.

> Okay, so MobyGames treated the two different Jones releases as two 

> different platforms because one of them contained a native Windows 

> version.  If you go to either version's main page, a link to the other 

> is shown via a grouping "Jones in the Fast Lane versions"

>

> Jones is a very old entry on MobyGames that might not be 

> representative of the entire database.  Heart of China is a little 

> better, showing all versions via covers 

> (http://www.mobygames.com/game/dos/heart-of-china/cover-art) as well 

> as some (but not all) releases here: 

> http://www.mobygames.com/game/dos/heart-of-china/release-info

>

> MobyGames is 100% volunteer, so not all entries may be completely 

> fleshed out with every single variant and detail.

>

>> So maybe not the best example. Let us instead go with Fate

>> of Atlantis which has a CD-ROM Version that's greatly enhanced.

>>

>> http://www.mobygames.com/game/indiana-jones-and-the-fate-of-atlantis/release-info 

>>

>>

>> Okay, it says "CD-ROM version has added speech" in the comments of one

>> release. And again, a trivia entry. It does not feature the two

>> different versions in any obvious way though. I'd somehow expect the

>> summary page to point out these exist. And I'm sure you can't run any

>> kind of database query to find all talkie cd-rom versions of all games

>> or anything like that because it is just some text hidden in generic

>> fields. That's all I'm saying.

>

> I don't understand your Atlantis example because the CDROM version is 

> clearly listed in the release info.  Maybe the problem is with our 

> presentation, and not necessarily our schema?


No, the problem is something else which I didn't notice before. It's the 
fact that the enhanced CD version of Jones in the Fast Lane gets a 
separate database entry (which I didn't notice before, hence the 
confusion above) but Fate of Atlantis CD doesn't. Other than the fact 
that JITFL has different platforms for the non-cd and cd versions and 
FOA hasn't I don't see why this is. And as you just witnessed during 
this conversation, this did a good job at confusing me, which is exactly 
what your database should prevent.

Now, about the actual FOA entry. Yes, it lists all releases in that 
list. But unfortunately, it treats the CD version just the same as the 
german version. I don't feel this is how it should be. Wouldn't it make 
more sense to have some structure like this?

Game: FOA
-> Version: Floppy
-----> Local Variant: US
-----> Local Variant: Germany
-> Version: CD
-----> Local Variant: US
.....

Also, you're duplicating this information in the Technical Specs where 
Media Type lists CD-ROM too but does not tell us anything about that. 
What I'm trying to achieve is that there isn't a catch-it-all entry but 
actually a separate list of variants that does tell you, okay, this had 
a different media type than the "main entry", namely CD-ROM. It differs 
in this and that way, has this UPC, etc. I know you have a list of 
variants too (called releases on your site) but it somehow seems to be 
not fully living up to its potential.


>

>> I don't know how on earth you got the impression I'm a pirate, part of

>

> I specifically wrote "I'm not calling you a pirate" so I wasn't doing 

> that.  I said that MobyGames was specifically created to catalog games 

> from a publisher perspective primarily and a collector perspective 

> secondarily.  I didn't feel at the time that a collector-centric 

> database was the best way to catalog publisher releases, so that's why 

> MobyGames is organized the way it is.

>

Yes, you said that but then again you did somehow feel the need to 
response that way just in case I was. So somehow, my initial message 
must have given you the impression I could just be some abandonware guy. 
Just curious how you got that impression, that's all.


>> Yes, your and other sites are great to

>> look things up but they just lack when it comes to simple requests like

>> "please give me a list of all NES games released in the UK".

>

> Our philosophy from day one was quality over quantity.  We require a 

> large set of basic information about a game before it is approved into 

> the database.  This wards off the less dedicated, plus it's also why 

> there's no import functionality, nor will there ever be.  I created 

> MobyGames specifically as a response to "game lists" of the 

> mid-to-late 1990s that contained title/publisher/year and little else 

> (which really bothered me, as I didn't feel that three pieces of info 

> was useful information from a historical or research perspective).


I'm sorry, don't take this the wrong way but actually imho your first 
sentence and the explanation after that contradict each other. You say 
your aim is quality but then you complain the old lists did not contain 
enough information which means not enough quantity of data. While I 
agree with that it is curious to me why you're still lacking behind some 
of those lists in terms of quality. What I mean is this. It's been years 
(6,7? not sure) since I did the first complete list of european snes 
releases for Digital Press. Despite this, your site is still missing 
some of those. This is something I just don't understand. If quality is 
your mantra (which I don't doubt a second) then why isn't the first 
thing to do make sure that you catch all releases for these closed 
systems. I'm pretty sure this is rooted in your initial focus on 
computer games rather than consoles and the contribution system in place 
but please correct me if I'm wrong here. I appreciated your efforts to 
collect as much relevant data as possible for the entries that are there 
but from an archival point of view, shouldn't the main priority be 
having all games in there? Now you end up "wasting" approver ressources 
on people contributing cheats for some US SNES game when the same guy 
could add a missing PAL SNES game in a couple of minutes.

(note: it's been a while since I checked snes, if it was completed in 
the meantime I'm sorry. Take it as an example only then)

Mind you, I'm not advocating having simple title+publisher lists only. 
Not at all. But in the end, having a simple entry for a game is better 
than not having it at all, at least for the closed systems. For the open 
ones like the PC it doesn't make sense to add what can only be called 
shell entries. No one has any benefit of having a pc game listed with 
only its name and publisher.


>

>> And if you want to be able to run a query like "all win 3.1"

>> games you probably have to do it that way depending on your db design.

>

> (http://www.mobygames.com/attribute/sheet/attributeId,59/p,5/ for the 

> record :-)  Anything in a tech spec can be clicked on.)

>

I'm sorry, I never noticed that it was clickable although I knew you 
could search for all Win 3.1 games. The above comment was a result of 
the fact that I did not notice the CD-ROM version of JITFL having a 
separate entry. Thus, Win 3.1 was only noted in the trivia section which 
I'm sure the above attribute search does not check ;)


>> I'd probably be ending up adding the CD-ROM version to my db twice, once

>> as Win 3.1 release and once as DOS release. And add a special link to

>> the database which lets you know they are sold in the same package.

>

> You need to clarify what you mean by "release".  If they came in the 

> same package, how can they be different releases?  Terminology is very 

> important if you will interact with other systems (not just mine).


About the terminology. The fact that I'm using somewhat vague terms here 
has two reasons. For one, every database and site is a closed system. 
Terms do not always and can not always reflect the real world 100%. As 
you know people use the term version for all kinds of stuff, same with 
release or any other term. For the database we have to decide on one 
specific term for one specific type of object. Just as you did. As your 
site and mine use different or differing terms it wouldn't help at all 
to strictly use the ones I've decided on in this conversation. And, 
secondly, I specifically asked the inital question because one of these 
version problems hasn't been solved for our database yet. Thus, I do not 
have a final term to use yet.

Well, how can they can be two different releases. The question is 
whether you want to track available game versions or available game 
packages.We obviously disagree on some points there as contrary to 
what's in your guidelines we will allow game entries for games only 
released in compilations too. Which is somewhat related. But, having 
said that, yes, I'd like to avoid having to add a game like this (JITFL 
CD) twice. I also know how to design the database to do so, not a 
problem. I'm just not entirely sure yet which route I want to go on that 
matter.

>

> I was very lucky in that my friend Brian Hirt who co-created MobyGames 

> with me used to write database-driven applications for a living, and 

> he pointed out the importance of proper design and normalization.  

> What you just described sounded like you were going to add two 

> different entries to your system for a single package -- I'm not sure 

> that's the best way, but that's why I have my system and you have 

> yours :-)  I'm just bringing it up as something to think about.  Once 

> you get pretty deep, it can be very difficult to reorganize.


Mind asking him if he's up for the job again? :p Just kidding of course. 
I'm an IT guy myself and normalization certainly is an important factor. 
Also, I've been working on game lists and databases for almost a decade 
so it is not like I'm new to this field. The design has been given lots 
of thought because I am well aware how crucial it is that some things 
never ever have to change once the database goes live. Having said that, 
I'm sure there are still flaws in it and I'll do my best to sort them out ;)

>

>> Yeah, names alone tell you nothing. Actually, this would probably have

>> been advertised as a conversion or port anyways. Some 8-bit conversions

>

> My point was that the core gameplay is nothing at all like Tomb 

> Raider, making it neither a port nor a conversion.  It's a completely 

> different game, only with the Tomb Raider title, so it got its own 

> entry in our system.


I know. What I meant was that they probably still advertised it as a 
port or conversion. There is no question here that it is a different game.

> Our database is nearly 100% normalized; there are over 300 tables. 

> Adding a platform is a lot more than adding a "BBC Micro" element to 

> the platform table -- now we need BBC Micro-specific tech specs, etc.


I see. Do you have system-specific tech spec tables then? My current 
approach is closer to 50 tables. Interesting to see how different 
solutions can be.  When you say normalized are you referring to 3NF?

>

> The bigger issue in creating a platform is identifying someone who 

> will be an approver for it.  Most of our approvers have firsthand 

> experience with the platforms they approve entries for, for obvious 

> reasons.  I've never touched a BBC Micro, so I'd have no business 

> approving BBC Micro game entries because I have no context in which to 

> judge if the entry was correct or a pile of BS.


That goes without saying, especially with computer systems. Which 
reminds me, I always wondered, does your team only approve user 
contributions or do they actively research and add information to, 
probably even in a systematic way (like, lets say, check the copyright 
year of all lynx games).

>> Well, thanks again for taking the time to reply to what is basically a

>> question by a "competitor". I appreciate and respect that a lot. If you

>> have any questions about my project feel free to ask although I doubt

>> you can have any seeing as nothing's public yet and there's no reason

>> yet to assume I/we can build anything worthwile. I know that and can

>> fully understand!

>

> No "competitors", there's always room for more historical data :-)  I 

> encourage you to make your ideas reality.  You can ask me anything, 

> privately or publicly.


Great :)

>

> If I were to be the slightest way snarky or malicious, it would be 

> because I know firsthand how incredibly difficult it is to get 

> something like this off the ground ;-)  For every hour you spend 

> implementing your system, you should spent 20 hours thinking about how 

> to implement it so that you avoid issues down the road.

>

> I'm not kidding about that 20:1 ratio.


I know you're not. It is not like I started designing this yesterday ;) 
There were months of thought before the first table was added to the 
database model!