Release Switch, Release State, and Unrelease State

[Tracy Poff]

In the current model, each ReleaseGroup has three properties: Release Switch, Release State (Data List 3), and Un-release State (Data List 4).

I propose that these two data lists be merged into a single data list, Release State, and that the Release Switch be removed from ReleaseGroups. The entries in the combined list will all be mutually exclusive, and, ideally, collectively exhaustive of the possible states. Since data lists are being implemented as database entities, we can add to each value (e.g. "Unofficially released", "In development") a flag for whether that state means the game has been released or not. This way, we won't be duplicating information (any release group with "Officially released" must have its Release Switch set to true, after all, or it's in an inconsistent state) and we won't have on each release group two properties only one of which can possibly be filled.

[gene]

Interesting!

You mean [url?http://wiki.oregami.org/pages/viewpage. ... Id=4490333]like this[/url] ?

[Tracy Poff]

Interesting!

You mean like this ?

Not exactly--I was suggesting a simpler model than this. I mean that each value like "Officially released" or "Cancelled" will have a fixed boolean value 'released' that determines whether it represents a released or unreleased state. I don't think there's much benefit in allowing released and unreleased to be chosen independently of the release states, and it would mean that (absent program logic preventing it) we could have incosistent states like Cancelled and Officially Released.

I think that if a game has never been released, then it must be either Cancelled, currently In Development, or in some Unknown state. No Release Group in any of these three states can ever have been released. The only one of these that might be questionable is In Development, but I think that a game that is released but in continuing development is in a sufficiently different state than one that is in development and has never been released that we oughtn't to label them with the same release state.

I propose the following release states:

Code: Select all

State Name      Released        Notes
Released        True            Official, not 'beta'. Demos and shareware count.
Open Beta       True            Marked as 'beta', but available (either for sale or free) to anyone.
Closed Beta     True            Available to some selected group of 'the public' as a beta.
Ongoing         True            Minecraft, etc.
Leaked          True            Complete or not, a release not officially sanctioned by the publisher.
In Development  False           As far as we know, on track to be released some day.
Cancelled       False           Officially cancelled. Includes completed (but unreleased) games.
Unknown         False           No word lately. Might be cancelled. Duke Nukem Forever.

Is there any situation where one of these states might reasonably have a different value for "Released"? Is there any state that a game could be in that doesn't fit one of these?

Problems I can see: 'Closed Beta' could be argued to be an 'Unreleased' state. The three 'Unreleased' states that I list really depend not on the state of the game so much as our knowledge about its state.

I'm not sure there's any important distinction between 'open beta' and just 'released' or between 'Closed Beta' and 'In Development', but I've kept them, since they were part of the original data lists. A simpler model could eliminate them.

[Ultyzarus]

The only other state that I could add to these is when a game has been released, but that it can't be played anymore due to cancelled service ie: Tetra Master, Final Fantasy XIV (version 1).

[Tracy Poff]

The only other state that I could add to these is when a game has been released, but that it can't be played anymore due to cancelled service ie: Tetra Master, Final Fantasy XIV (version 1).

That's definitely something worth noting, but I'm not sure that Release State is the right place for it. It's a complicated situation, actually--consider games that have private servers, but for which official support has ended, or games that have limited or single-player-only offline modes, but can no longer be played online. It seems like we might want to consider this as an entirely separate property.

[Rola]

consider this as an entirely separate property.

"official servers closed" tag of some sorts.

[MZ per X]

Good to re-visit this.

I like your approach, but have some suggestions.

1) Thinking again, I would define the state "released" as: Fully playable version was available to the public. This would mean that "closed beta" would count as unreleased, which seems logical to me, as there's no real difference to an in-house beta test.

2) Following 1), should we consider games as "unreleased" that only have a demo/promo version available? I'd say yes, but that would mean that even games which are "unreleased" would need releases attached to them later on.

3) I would add a new state for Ultyzarus' case: "Released, official servers closed". This state would only be allowed for MMO games without single player content. The other cases you mentioned could be mapped using tags, as Rola said.

[Ultyzarus]

3) I would add a new state for Ultyzarus' case: "Released, official servers closed". This state would only be allowed for MMO games without single player content. The other cases you mentioned could be mapped using tags, as Rola said.

Tetra Master is in the same state, with "servers closed". Can an online card game be considered a MMO?

[Tracy Poff]

1) Thinking again, I would define the state "released" as: Fully playable version was available to the public. This would mean that "closed beta" would count as unreleased, which seems logical to me, as there's no real difference to an in-house beta test.

As I noted above, I agree that 'closed beta' is probably more like an unreleased state. I think I would suggest just merging 'open beta' into 'released' and 'closed beta' into 'in development'. But I do think there's some difference between a closed beta and in-house testing: Blizzard held a closed beta of Diablo 3, as I recall, with thousands of players invited. It wasn't an open beta, but it wasn't much like in-house testing, either.

2) Following 1), should we consider games as "unreleased" that only have a demo/promo version available? I'd say yes, but that would mean that even games which are "unreleased" would need releases attached to them later on.

Well, the release state attaches to a ReleaseGroup, and demos and stuff get their own release groups (don't they?), so I'd say that it's totally appropriate to consider the demo as 'released'. The full game wouldn't be, but I suppose that we can easily check if a game has any RG that is released and isn't a demo, if we need to check whether the game has been released.

3) I would add a new state for Ultyzarus' case: "Released, official servers closed". This state would only be allowed for MMO games without single player content. The other cases you mentioned could be mapped using tags, as Rola said.

I don't like this idea. It mixes information about the status of the servers, the status of the release, and the type of game we're dealing with. Further, consider if we want to know whether a game's servers are still up. Then we'd need to check in (at least) two different places, depending on whether the game had single-player content available. This means duplication of data, which is obviously bad, and the possibility for inconsistent data, which is even worse. Ideally, we should track each property of a game in exactly one place in the database.

I'd suggest, instead, that we have some flag for 'online only' games, and then the combination of that flag with a 'servers down' flag would mean that the game was no longer playable at all (without resorting to private servers, at least). Very probably both of these flags should be attached to Release Groups. I can't verify an example right this moment, but I'm sure that some game (Phantasy Star Online, maybe?) has been released for multiple platforms and had the servers for a single platform shut down before the servers for other platforms.

[MZ per X]

Can an online card game be considered a MMO?

Oh yes. MMO is everything that's played over the Internet, and where you can interact with other players, even this.

I think I would suggest just merging 'open beta' into 'released' and 'closed beta' into 'in development'. But I do think there's some difference between a closed beta and in-house testing: Blizzard held a closed beta of Diablo 3, as I recall, with thousands of players invited. It wasn't an open beta, but it wasn't much like in-house testing, either.

Yes, let's merge these. We can have the distinction at release level, with a date attached to it, if we introduce a new property "release type" there. We need this property for other things anyway, I just didn't add it, yet.

2) Following 1), should we consider games as "unreleased" that only have a demo/promo version available? I'd say yes, but that would mean that even games which are "unreleased" would need releases attached to them later on.

Well, the release state attaches to a ReleaseGroup, and demos and stuff get their own release groups (don't they?), so I'd say that it's totally appropriate to consider the demo as 'released'.

Haha, you understand my data model better than I do. I've waited for a long time to write this sentence.

BTW, there's this PlantUML thing in the Wiki, with which we could work collaboratively on the data model. Maybe we could try finalizing smaller parts of the model using this, as unlike Gliffy it allows change tracking / reversing.

3) I would add a new state for Ultyzarus' case: "Released, official servers closed". This state would only be allowed for MMO games without single player content. The other cases you mentioned could be mapped using tags, as Rola said.

I don't like this idea. It mixes information about the status of the servers, the status of the release, and the type of game we're dealing with.

Agreed.

I'd suggest, instead, that we have some flag for 'online only' games, and then the combination of that flag with a 'servers down' flag would mean that the game was no longer playable at all (without resorting to private servers, at least). Very probably both of these flags should be attached to Release Groups.

Definitely, they should.