10 Mar, 2009, Scandum wrote in the 141st comment:
Votes: 0
FAMILY provides a list that should work for 95% of the MUDs out there. CODEBASE is more difficult because it continuously evolves, so I think Crawlers will have to provide guidance here.

This should be easier once Mud Wiki launches officially which I've organize using a similar hierarchy, it's open for editing if anyone is interested.
10 Mar, 2009, quixadhal wrote in the 142nd comment:
Votes: 0
Scandum said:
quixadhal said:
The most distant ancestor would almost always be AberMUD, since LP, Diku, and Tiny all derived from it.

They were inspired by AberMUD, not derived from it. The only AberMUD derivative I know of is Dirt.

I'll leave it to the mud crawlers to translate the FAMILY and CODEBASE values into a hierarchical structure of derivatives.


Perhaps, but that doesn't invalidate my point. Knowing the most recent ancestor will automatically tell you the most distant ancestor. Knowing the most distant throws away information. If I start a new mud using AFK as my CODEBASE, by your list I have to put DikuMUD as my FAMILY. Thus, people think the mud will work just like someone who starts a new mud using TBA as their CODEBASE, which also uses DikuMUD as their FAMILY… even though a Circle based MUD and a Smaug based MUD will feel quite different.

You can expect the crawlers to know that Circle and Smaug both derive from Diku, but that Smaug also derives from Merc. You can't expect them to know that FlubberBASE and FooBASE both derive from Smaug, but BooBASE derives from ROM. As you say, CODEBASE will be constantly changing, and will often be invalid as folks want to claim their own "FooBASE 1.0". If FAMILY is the most distant ancestory, then you lose data and just have a big pile of DikuMUD's, a small pile of "Custom", and a small sprinkling of all the others. Might as well not bother with it.
10 Mar, 2009, Scandum wrote in the 143rd comment:
Votes: 0
It seems to work on TMC where most muds properly report their codebase as something TMC can work with.
10 Mar, 2009, quixadhal wrote in the 144th comment:
Votes: 0
If you're looking for a mud to play, and you see it has a codebase of ROT, is it more useful to know it's a descendant of ROM, or that its distant ancestor is DikuMUD? Is the data meant to be useful, or just complete for the sake of completeness?
10 Mar, 2009, Rendelven wrote in the 145th comment:
Votes: 0
I agree with Quixadhal in regard to the 'FAMILY'. Would not the nearest ancestor reveal the most information ( including every 'ancestor' before it ) ?
10 Mar, 2009, David Haley wrote in the 146th comment:
Votes: 0
Thirded. (I think I said so before, but not very clearly. I think the closest ancestor gives the most information, and public codebases don't change so often that this will be a problem for crawlers to maintain.)
10 Mar, 2009, Scandum wrote in the 147th comment:
Votes: 0
quixadhal said:
If you're looking for a mud to play, and you see it has a codebase of ROT, is it more useful to know it's a descendant of ROM, or that its distant ancestor is DikuMUD? Is the data meant to be useful, or just complete for the sake of completeness?

FAMILY is supposed to be all inclusive, CODEBASE is meant to be informative. I'm sure that MudBytes can maintain a proper hierarchy that includes ROT.
10 Mar, 2009, David Haley wrote in the 148th comment:
Votes: 0
You're putting the onus on the crawlers to have a full hierarchy so that they can compute some interesting midpoint between family and custom. It would be so much simpler to just give the answer that people care about anyhow in the MUD output…
10 Mar, 2009, Scandum wrote in the 149th comment:
Votes: 0
It's impossible to compute a complete branch from the middle though.
11 Mar, 2009, David Haley wrote in the 150th comment:
Votes: 0
What? I'm not sure I follow. But if something is impossible to compute, isn't that just another reason to give the answer straight away?
11 Mar, 2009, kiasyn wrote in the 151st comment:
Votes: 0
Honestly, 90% of these variables are pointless anyway. How often is your ancestor going to change? The MUDBytes crawler only crawls MUDs in the MUD list… when you're changing your /codebase/ I expect you'd update your MUD Listings anyway.
11 Mar, 2009, David Haley wrote in the 152nd comment:
Votes: 0
I thought the idea was that the MUD admin wouldn't have to enter a lot of these by hand, so why not just use one mechanism to provide information? Sure, it won't change often if at all, but who cares?
11 Mar, 2009, kiasyn wrote in the 153rd comment:
Votes: 0
Just seems rather pointless.
11 Mar, 2009, David Haley wrote in the 154th comment:
Votes: 0
If the whole idea is to give information automatically, why are you requiring things that just happen to not change often be given manually? That would force MUD authors to do more work than just submit their contact information. It also means that I can't start a crawler on my own using connection information I got elsewhere, because I'll have incomplete information.
140.0/154