Argo Filch
BetaMax. CRONOS.
15
|
Posted - 2013.05.04 21:40:00 -
[1] - Quote
I know i will sound like an ungrateful bastard down the line, that's why i'm saying this first: I really really appreciate the effort you guys put into making and releasing those data dumps for us. 3rd party devs will really be grateful for this stuff and hopefully will make great tools from them.
That said... well... the first try is nice and all... but there's a bunch of issues with those two files you released.
First, the Skill_Tree.xlsx . Yes... its nice and pretty... but in the end pretty useless, because you can't really extract any meaningful data from it. Sure i can copy the data in an excel sheet and tabularize it but if i have to do it myself well then why have an xslx file to begin with. You could just release that thing as a pretty image to show people.
The most needed information this document doesn't even provide. Like what these skills actually do (bonuses) and what rank you need to get the next ones in the tree. A tabular presentation for 3rd party devs with a parent column that shows the relation between the skills and a rank column would have been really nice.
I even looked into that xslx file with an editor to check if there's any salvageable data one could programmatically extract, but there's really nothing there.
Second, the more usable dust_market_items.csv file. This file has a bunch of interesting and usable information. But this file is really missing a few crucial things and has some issues.
1. Its Category column needs to be populated for all items, not just for the skillbooks. Would be really nice to have the different categories show up in that column like Assault Rifles, Shotguns, Shield Extenders, Shield Regulators, et cetera.
2. It needs a Group column that holds a more general information on what kind of Item it is. For example an entry for Skills that groups all the skill book items, or an entry for Dropsuits, Vehicles, Weapons (Infantry), Weapons (Vehicle), Modules (Dropsuits), Modules (Vehicles), Equipment, Low Powered Modules, and so forth. So all infantry weapons would be grouped under Weapons (Infantry). This would make it easier for people to sort and present the data. As it is now, if i want to make real use of the data i have to add this manually. EVE already has those two column types, but there it's reversed. Group is the more specific and Category the more general classification.
3. There needs to be two new columns that reference the skills (could use the typeID of the skillsbooks in that list) needed for specific items and what rank that skills needs for it to be usable. Right now this information isn't present in this file atm. E.g. the 'Viziam Laser Rifle' needs typeID 353465 with a rank of 4 (don't actually know if that's true) to be usable. This information is needed so developers can present for example their users a way to only see equipment they can actually use. Would also need to put that data in it manually to make this kind of functionality. Those two columns would also relativize the need for a Skills_tree file. And right now i think there's not really a item or skills that needs two prerequisites so a simple parent<->child relation would suffice. If that's not the case we'd need a cross-referencing table to present that data.
4. Putting unit labels after numeric values completely kills any kind chance to import that data into database columns of other types then string. Or have any kind of use for it in a spreadsheet application. You just can't add '80 HP' from the Armor Bonus column of some Plates to the '150 HP' from the Armor column of some Dropsuit. You have to remove that HP unit label to be able to use the data. Why not rename the column to Armor (HP), Armor Bonus (HP) so you know what kind of unit it is but the data in that column would still be usable. What i really want to do with this csv file is to import it into a database... but atm i'd have to manually search and replace every unit label in every column with an empty string to get the actual numeric value to import into a database.
This brings me to point 5. You have Data in columns that really belong split into two columns. For example the column Scan Precision. For Dropsuits it has values with dB units and for Precision Enhancers you have a percentage value. This data really really needs to be split into two columns. You can't even save that stuff in a meaningful way in a database the way it is atm. Shield Recharge Delay is another example of this, as is Shield Depleted Recharge Delay, Scan Profile, Shield Bonus, PG Bonus, et cetera.
Those are the things that caught my eye in the first few minutes of looking into those files.
I hope those issues get looked at when there's a bit spare time on hand for the guy making the files in the first place. Who was that actually?
Thanks for reading this stuff :) |