Proposal: Schema for shipyards about eddn HOT 13 CLOSED

See: https://github.com/jamesremuscat/EDDN/blob/master/schemas/shipyard-v1.0-draft.json

from eddn.

Some comments:

I suggest removing the "minItems" restriction on ships , so that "ships": [] can indicate the absence of a shipyard.

Since only the name is mandatory all subscribers will need to handle it. This makes the fdId field redundant. Which field takes priority if the name and fdId conflict?
If you keep the fdId field I would suggest making it an "integer" for consistency with the commodity schema.

Do you intend to publish a list of names?
i.e. should that be "Cobra Mk III", "Cobra Mk. III" or "Cobra Mk.III" and "Type9", "Type-9" or "Type-9 Heavy" etc?

from eddn.

Well since the fdID is only proposed by a few softwares and you don't know how long it will last,
I tends not to rely on it too much. I think it should be to the subscribers to take care of that.
But it is thinking about future, as no one except those API software will use it for now ^^

fdID isn't an integer because on JSON, it can be unsigned and 0. By making it a number you can assure that it will not. See: http://spacetelescope.github.io/understanding-json-schema/reference/numeric.html

Agreed on removing the "minItems" on ships.

As for the list, we can add an enum field, but it would need us to be quick on updates when new ships are released.

I'll update in consequence.

from eddn.

I'm not necessarily suggesting that "name" be an enum - as you point out this would break when a new ship is introduced. I suggest maintaining a master list of expected ship names/spellings, but outside of the schema.

from eddn.

Or we can make the definition in the schema but let the field be a string

from eddn.

Some further comments on your updated version:

You're missing "Federal Dropship", "Imperial Courier" and "Imperial Clipper".
That should probably be "Sidewinder", not "SideWinder".

The last item (currently "Diamondback Explorer") in the enum should not be followed by a comma - it stops the schema passing jsonschemalint.

FYI here is a sample message in accordance with this schema.

I agree that it would be better for "name" to be a string.

from eddn.

Yep it was a quick update based on your gist
I have them like ingame in my database.
perhaps will move them on wiki.

from eddn.

I've updated that gist with the IDs of all of the currently flyable ships.

from eddn.

Looks OK to me.
I still think fdId is redundant and that it's not clear what subscribers are supposed to do with it, so I won't generate it.

from eddn.

@Marginal If it was for me, I woudn't have it either, some it's removed.
I think the proposal is good.

I leave for the week, but as soon as I return, I'll give it a shot in EDCE and ET-TD.

@maddavo What do you think of it ?
Does it suits your needs, your one of I think of when it comes to shipyard info ?

from eddn.

This schema could be simplified with no loss of information - please see my pull request #23.
Here is a sample message in the simplified schema that I'm suggesting in #23.

from eddn.

Thanks. This draft schema is implemented in EDCD/EDMarketConnector@f332f7f.

from eddn.

If anyone is looking for data, and schema examples everything is available here:

https://github.com/cmmcleod/coriolis/tree/master/data

Just an FYI :)

from eddn.