[Ohrrpgce] some thoughts on reload

[Ralph Versteegen]

2008/11/7 James Paige <Bob at hamsterrepublic.com>:
> Having caught up on the IRC log and the on-list discussion from last
> night, I have a few thoughts:
>
> * I like binsize.bin. I think the system is nice and simple, and it
> feels less hacky the more I work with it. The only downsize is the
> somewhat slow upgrade because of lump resizing

I agree, I don't see it as a problem either. The only strange thing
about it is that the record size is stored in a separate lump instead
of in a header.

>
> * If a goal is to decrease upgrade times, a format change will not
> accomplish that goal, even if the new format can be upgraded more
> efficently. Remember, the "upgrade" sub CAN'T go away, because of the
> vast library of old games that still float around, most of which will
> only ever be upgraded by players at load-time rather than by authors.
> The upgrade time for these games will be
> oldupgradetime+conversiontime+newupgradetime
>
> * Flat binary files with binsize.bin are prefectly adequate for many
> lumps in the OHR. Please don't upgrade a format without a much better
> than "because I can" reason.
>
> * All that being said, the plan for RELOAD does seem pretty cool, and
> there are already some lumps that could benefit from conversion.
> menus.bin+menuitem.bin are top candidates. Also the slicetree_0.txt
> format which I haven't even bothered to document on the wiki, since I
> was so confident it would be replaced by something else later.

Again, agreed entirely. We shouldn't go out and replace existing
formats without a compelling reason. I hope the whole RPG conversion
idea is dead.

>
> * RELOAD could be awesome for the new save format.
>
> * RELOAD could be awesome for the EditEdit lump definitions.

Ah, I hadn't thought of those two. Any more details for the save format then? :)

>
> ---
> James