09 Mar, 2009, Dean wrote in the 101st comment:
Votes: 0
I apologise, I ate the crawler thinking it was a rather tasty treat. It wasn't.
09 Mar, 2009, Rendelven wrote in the 102nd comment:
Votes: 0
I got an MSSP request from MudBytes. That the crawler working? :)
09 Mar, 2009, kiasyn wrote in the 103rd comment:
Votes: 0
Rendelven said:
I got an MSSP request from MudBytes. That the crawler working? :)


that was quick.

I'm toying with it now, its almost ready to go.
09 Mar, 2009, Rendelven wrote in the 104th comment:
Votes: 0
Yeah. I have it where it notifies me if someone requests it atm.

Is it going to be linked to existing mud listings? How is that going to work?
09 Mar, 2009, kiasyn wrote in the 105th comment:
Votes: 0
So far I'm just dumping MSSP into a database - Davion will be handling the front end.
09 Mar, 2009, Tyche wrote in the 106th comment:
Votes: 0
If your server or client crashes due to an improperly formed IAC DO/DONT/WILL/WONT/SB/SE, stray NUL or any other rare oddity, I guarantee the probability goes up to 100%.
And I'll be laughing when I do it…. ;-)

*humps Scandum's other leg*
09 Mar, 2009, kiasyn wrote in the 107th comment:
Votes: 0
current data collected from mudbytes crawler http://www.kiasyn.com/mssp/
09 Mar, 2009, Tyche wrote in the 108th comment:
Votes: 0
Mister said:
Most diku-derivatives read input not line-by-line but 'packet-by-packet'.


Packets and write/send/read/receive are different animals. At the level of the C TCP sockets interface all you can say is you are reading/writing bytes to a stream, and nothing with any certainty about how many packets that might be.
09 Mar, 2009, Scandum wrote in the 109th comment:
Votes: 0
Tyche said:
If your server or client crashes due to an improperly formed IAC DO/DONT/WILL/WONT/SB/SE, stray NUL or any other rare oddity, I guarantee the probability goes up to 100%.

I'll need to proof read the code with crashability in mind sometime.

kiasyn said:
current data collected from mudbytes crawler http://www.kiasyn.com/mssp/

That looks better than expected. The protocol states the crawler should provide the IP value if it's omitted or left blank. How would the crawler handle a MUD that sends a PORT variable and value twice?

For the MUDs:

AzerethMUD is incorrectly adding the port to the hostname field
Both muds report USA instead of the ISO 3166 location which would be "United States"

Also, for muds that support more than one intermud protocol, should they report it multiple times, most important protocol last, or all stuff it in one value?
09 Mar, 2009, quixadhal wrote in the 110th comment:
Votes: 0
Scandum said:
That looks better than expected. The protocol states the crawler should provide the IP value if it's omitted or left blank. How would the crawler handle a MUD that sends a PORT variable and value twice?

That might be a policy decision that the protocol should spell out. If a MUD lists multiple IP's and/or PORT's, it may actually have multiple homes and allow connections from any of them – or one may be a builder port – or they might have realms where good logs into one side and evil logs into the other.

In any case, the protocol might want to specify something like "if only one IP or PORT is to be kept/listed, take the first one given".

Scandum said:
AzerethMUD is incorrectly adding the port to the hostname field
Both muds report USA instead of the ISO 3166 location which would be "United States"


Data normalizing probably falls to the crawler. I would probably say that what ends up in the database should be sanitized, so the crawler may need to implement some extra code to sanity-check all the input values. The MUD admins aren't likely to do it consistently, and rejecting the data because they used US or USA instead of United States would be counter-productive.

Scandum said:
Also, for muds that support more than one intermud protocol, should they report it multiple times, most important protocol last, or all stuff it in one value?

I would treat that the same way you treat multiple IP's or PORT's. Most people will list the most important ones first (inconvenient as that may be).
09 Mar, 2009, Scandum wrote in the 111th comment:
Votes: 0
quixadhal said:
Scandum said:
Also, for muds that support more than one intermud protocol, should they report it multiple times, most important protocol last, or all stuff it in one value?

I would treat that the same way you treat multiple IP's or PORT's. Most people will list the most important ones first (inconvenient as that may be).

In the case of hostnames and ports the protocol specifically says to use the most important hostname and port last. The same would go for intermud, that way a lazy crawler can overwrite previously reported values without giving it a second thought.
09 Mar, 2009, elanthis wrote in the 112th comment:
Votes: 0
I like how the MUD banner is shown. It's really neat looking, and will be great once it's cleaned up (interpret the ANSI color commands to translate them to HTML, and I'd probably try stripping off the last line of input if it has no \r\n so the login lines don't show up).
09 Mar, 2009, quixadhal wrote in the 113th comment:
Votes: 0
elanthis said:
I like how the MUD banner is shown. It's really neat looking, and will be great once it's cleaned up (interpret the ANSI color commands to translate them to HTML, and I'd probably try stripping off the last line of input if it has no \r\n so the login lines don't show up).


Actually, the best way is to form a png graphic from them, that way it always looks the same, regardless of the weird fonts the viewer uses, or the IE vs. Firefox minefield. I suggest png's because they're not lossy like jpegs, and even when thumbnail'd down, they're still reasonably useful.

I actually have code to do this, but it's in Perl and makes heavy use of CPAN, thus it's of limited use for a ruby crawler, unless one wanted to break it off as a cgi script.

Here's my example. It's a bit slow, so be gentle.
09 Mar, 2009, David Haley wrote in the 114th comment:
Votes: 0
Why exactly are we using the term "PREROGATIVE" to denote the general MUD family?
09 Mar, 2009, mtfox wrote in the 115th comment:
Votes: 0
I like the way the banners are handled on the mud gallery
mud gallery
they load up nice and fast that way.
But then it depends on if they are displayed in a group like they are currently, or if it is only going to display whatever mud you clicked on to see the stats for.
09 Mar, 2009, Rendelven wrote in the 116th comment:
Votes: 0
I was using 'USA' because it is listed under the ISO 3166 standard. It is the 'alpha3' or whatever.

Are we supposed to use the Full Name instead of the abbreviated versions?
09 Mar, 2009, Scandum wrote in the 117th comment:
Votes: 0
David Haley said:
Why exactly are we using the term "PREROGATIVE" to denote the general MUD family?

Are you trying to tell us you've been discussing the protocol definition for over 2 weeks now without reading the actual protocol? That'd explain a few things.

Suggestions for a better variable name are welcome.
09 Mar, 2009, Scandum wrote in the 118th comment:
Votes: 0
Rendelven said:
I was using 'USA' because it is listed under the ISO 3166 standard. It is the 'alpha3' or whatever.

Are we supposed to use the Full Name instead of the abbreviated versions?

The official country name should be used (preferably not in all caps) so the crawler doesn't have to parse it to make it readable. I didn't want to use alpha2/3 because those codes are pretty obscure for countries other than the USA.
09 Mar, 2009, David Haley wrote in the 119th comment:
Votes: 0
Scandum said:
Are you trying to tell us you've been discussing the protocol definition for over 2 weeks now without reading the actual protocol? That'd explain a few things.

It would explain the fact that I think your protocol has enough fundamental issues that I haven't bothered to comb through all the details, yes. I think it's pretty clear that I have been discussing several of the field values. In fact, this is hardly the first time I've suggested different names, values, etc. For whatever reason, I just didn't notice the term 'prerogative' up until now. Unless you have more cheap repartee up your sleeve, shall we put this aside?

I would think that the term "Family" is a quite preferable.
09 Mar, 2009, Rendelven wrote in the 120th comment:
Votes: 0
Scandum said:
The official country name should be used (preferably not in all caps) so the crawler doesn't have to parse it to make it readable. I didn't want to use alpha2/3 because those codes are pretty obscure for countries other than the USA.


You may want to add that to the protocol so other people don't make the same mistake as I did.
100.0/154