Dragon's Armory developer wishlist

In no order, below is a list of things I'd love to be able to access or see fixed in the data api for EQ2:

• Any missing spell properties added to constants.generalspellproperty_list. To be honest, I believe the features I found these in came into Census after the data in constants was created so it's not a major pain, just annoying.
  • Examples: mana (from reforging), heal (from set bonuses)
• A change to how the lists are done. Right now, the resulting data in the list is just an array which makes this unsearchable directly from the Census API. I propose, like the PS2 character.stats elements are done, that the id numbers are used as the key for that array.
  • Examples, equipmentslot_list[0].item.id be able to use equipmentslot_list.0.item.id, or achievements.achievement_list[0].id be able to use achievements.achievement_list.32100782.id.. One could do count queries to see how many folks have an item equiped, or an ordered list of who completed an achievement sorted by completed date, etc.
• An export of raw unmodified spell data, maybe name the collection as spells_raw, so we can run our own calculations on them. With the data points of min damage, variance, and duration for each effect on the spell, missing components from the spells collection (range, resistability, the type identifier to use against the crit multiplier data, etc), and any special limitations (some spells do not use ability mod, some have crit multiplier overrides, etc.)
• An export of the critical damage multipliers per class and damage type, maybe added onto the class component in the constants collection.
• An export of the reforging stat to stat multipliers.

I reserve the right to add more to this!

 

Feldon and I have actually been talking about that for quite a while now.

PS2 does some of their arrays in a way that would make them much more useful. The current way it's done in EQ2:

Code:

<achievement_list>  <achievement id="123" completed_timestamp="3454345543"/>  <achievement id="543" completed_timestamp="2345433453"/></achievement_list>

With this structure, you can't query on achievement 123 or 543 and do things like return the characters that have completed that achievement sorted by the completed timestamp.

The PS2 method would have been to do it like this:

Code:

<achievement_list>  <n_123 id="123" completed_timestamp="3454345543"/>  <n_543 id="543" completed_timestamp="2345433453"/></achievement_list>

The n_ is a prefix for numbers used as node names. This lets you query achievements by id since you can reference the specific one you want, n_123 for example. It still lets you iterate through them though since the id is also available as a property of the node.

Personally, I would prefer the more verbose hybrid of the two:

Code:

<achievement_list>  <achievement_123 id="123" completed_timestamp="3454345543"/>  <achievement_543 id="543" completed_timestamp="2345433453"/></achievement_list>

This lets you query achievements by id using achievement_123 for example, but doesn't make it look quite as confusing as I believe using n_ in front of every number used as a node name does.

But that's just my 2cp.

 

Yeah, the JSON view of the PS2 method looks much nicer (no "n_" prepended), that's seems to only be an xml thing.

 

Never even looked at the JSON view before since I haven't tried to use the data on a site anywhere. I'm guessing it's invalid to start a node name with a number in XML so they simply put n_ in front of the numbers for the xml view.