Newsfiend
07-07-2009, 01:41 PM
The problem with backups with regards to changing the structure of the character file is this: when you change the character file, your backups are no longer valid.
Regardless of how it's implemented, adding inventory space and bank space rates high on several "danger" axes.
It's very high on the effort scale, as all manipulation of the character file will need to take the new data structure into account. Any functions that use pointer math (as most of the early functions did for speed purposes) will need to be rewritten. Our character files have years of backups archived, and once the new format is complete, those backups will have to be unarchived, updated to the new format, and re-archived so they can be used as backups once again.
It's very high on the invasiveness scale, as this requires manipulation of one of the most basic classes that is used by nearly every function in the game. While technically it shouldn't matter what size the class is... this is EverQuest, and something, somewhere, will break spectacularly. Odds are this won't show up until live loads are put on the code.
It's very high on the process scale. Not only do our developers have to create, test, and regress the code, our Operations and Platform teams have to actually apply our changes to the millions of character files that are spread across all of our servers. CS tools have to be updated to see and manipulate the new storage spaces. EQPlayers will need to be updated to read and maintain the new storage. Platform and Operations time is split between all SOE games, so something this time-consuming will have to be scheduled well in advance.
It's extremely high on the consequences scale. If player files are broken, no one can play the game. If player files are corrupted, our customers can lose everything they ever had. If the conversion process corrupts player files, restoring from backup can become impossible. In short, a mistake that isn't caught has the potential to kill the game entirely. This is NOT something that anyone takes lightly.
All of the above is true for any increase for personal storage, whether it is a second banker, a trophy bank, a tradeskill bank, or simply more bank slots. The presentation to the player is not what makes the process complicated.
That said, there are two important notes here:
We have done this before. It isn't impossible, it's extremely complicated. If we're at a point where the benefit outweighs the costs, it can happen.
I'm not a programmer on our team, so the above list isn't exhaustive. There are very likely more costs and risks that I'm not aware of. Those are just the issues I know of based on my knowledge of our game's structure.
Hope that helps!
More... (http://forums.station.sony.com/eq/posts/list.m?topic_id=152934&post_id=2233189#2233189)
Regardless of how it's implemented, adding inventory space and bank space rates high on several "danger" axes.
It's very high on the effort scale, as all manipulation of the character file will need to take the new data structure into account. Any functions that use pointer math (as most of the early functions did for speed purposes) will need to be rewritten. Our character files have years of backups archived, and once the new format is complete, those backups will have to be unarchived, updated to the new format, and re-archived so they can be used as backups once again.
It's very high on the invasiveness scale, as this requires manipulation of one of the most basic classes that is used by nearly every function in the game. While technically it shouldn't matter what size the class is... this is EverQuest, and something, somewhere, will break spectacularly. Odds are this won't show up until live loads are put on the code.
It's very high on the process scale. Not only do our developers have to create, test, and regress the code, our Operations and Platform teams have to actually apply our changes to the millions of character files that are spread across all of our servers. CS tools have to be updated to see and manipulate the new storage spaces. EQPlayers will need to be updated to read and maintain the new storage. Platform and Operations time is split between all SOE games, so something this time-consuming will have to be scheduled well in advance.
It's extremely high on the consequences scale. If player files are broken, no one can play the game. If player files are corrupted, our customers can lose everything they ever had. If the conversion process corrupts player files, restoring from backup can become impossible. In short, a mistake that isn't caught has the potential to kill the game entirely. This is NOT something that anyone takes lightly.
All of the above is true for any increase for personal storage, whether it is a second banker, a trophy bank, a tradeskill bank, or simply more bank slots. The presentation to the player is not what makes the process complicated.
That said, there are two important notes here:
We have done this before. It isn't impossible, it's extremely complicated. If we're at a point where the benefit outweighs the costs, it can happen.
I'm not a programmer on our team, so the above list isn't exhaustive. There are very likely more costs and risks that I'm not aware of. Those are just the issues I know of based on my knowledge of our game's structure.
Hope that helps!
More... (http://forums.station.sony.com/eq/posts/list.m?topic_id=152934&post_id=2233189#2233189)