|
|
| Author |
Message |
olene
Newbie
Joined: 31 Jul 2004
Posts: 61
Location: olene on DALnet
|
 Posted: May 17, 2005 11:51pm Post subject: Top ten really minute things i dont like about IRCds/RFC1459 |
 |
|
1. 481 numeric... "Permission denied -- You are not an IRC operator". God. I know thats how it is in RFC1459, but can we at least shorten it to "You are not an IRC operator". You don't have "permission denied -- You are not a channel owner" or "permission denied -- You have not registered". Shorten it!. IIRC I noticed that recently one of the major networks (i think ircnet) had shortened it to just "Permission denied". In any case.. in all 13 of my crappy ircd's, i've always used ":You are not an IRC operator"
2. Cannot send to channel. Why the hell not? I'm not sure if unreal does this or not, but i somehow got the idea early on that the "(+m)" in "Cannot send to channel (+m)" was standard, but on recent inspection.. it's not! Think about.. now adays, you have +m +M +b +n +c +C and a whole other host of limits on what you can say. At least include a little caption (like "Cannot join channel (+i)") so you know why.
3. People using NOTICES where they should be using NUMERICS. I cringe when i connect to a server and i get "*** Total global connections: 49489 (39393 max)" I'm pretty sure thats what RPL_GLOBALUSERS is for, and though its not in the RFC, it's pretty damn standard.
4. "-mesa.az.us.undernet.org- on 6 ca 1(4) ft 10(10) tr".
What the !@#$ is that?!? It's not so much that I can't understand it, but its the fact that it's noticed to you, everytime you connect.
The tao of irc states "The client should always respond the luser with messages that will not astonish him too much. The server likewise. If the server does not, then it is the clients job to explain what the server says."
I have not yet seen a client that interprets this for me.
5. Quit message prefix fux0red up syndrome
Ok. Quit message prefixes are ok (e.g. "quit:"). I especially like dancer's solution of simply enclosing it in quotes. My annoyance is with unrealircd and other ircd's that will send the entire quit message as "Quit: " with the space and all, if you never specified a message. I mean... "*** blah (blah@blah) quit (Quit: )" looks really shitty, IMO.
Which brings us into #6...
6. The default message being your nickname.
Who the !@#$ thought this shit up? What's wrong with an empty string? i wanna know what the !@#$ is wrong with "*** $#$@!@# was kicked by olene ()". This isn't webtv.. the kickee is perfectly capable of seeing who kicked them. Clients are perfectly capable of parsing an empty string (RFC1459 even notes that the trailing parameter may be an empty string). So WTF? Some servers, thankfully, do allow empty strings. Some only allow empty strings if I supply the ":" with nothing after it. And some simply will always fill in my nickname.
7. Ritz crackers that are too salty.
8. Servers without WATCH. Bandwidth saved alone from elimination of periodic ISON queries is worth it.
9. I ran out of things that annoy me. But i'll think of more. |
|
| Back to top |
|
 |
angelic
Lurker
Joined: 01 Aug 2003
Posts: 148
|
| Posted: May 18, 2005 2:20am Post subject: |
|
|
| Now i know why men think women are whiny. |
|
| Back to top |
|
|
Dr-Voodo
Eleet
Joined: 07 Nov 2003
Posts: 535
Location: IRC
|
| Posted: May 18, 2005 3:17am Post subject: |
|
|
Lol! Well it sure explains alot  |
|
| Back to top |
|
|
Roku
Newbie
Joined: 18 Apr 2005
Posts: 92
|
| Posted: May 18, 2005 8:37am Post subject: |
|
|
olene,
The way I see it you have 4 choices:
1. live with it
2. /quit
3. create your own protocol
4. ignore RFC lke so many others
Complaining here will only get you flamed or get you ignored but it won't change anything.
So many people spend so much time and energy complaining about how something is/isn't and not enough time or energy doing something about it.
Roku |
|
| Back to top |
|
|
Ashen
Idler
Joined: 05 Jan 2004
Posts: 285
|
| Posted: May 18, 2005 11:09am Post subject: |
|
|
I don't think olene was trying to get things changed. I just think they were venting their psychological frustration. A little bit of bitching is sometimes a good thing, becuase it clears the air of the things that were irritating you inside, and keeps you sane :)
The point isn't to change things, or to convince people, just to air your frustrations so they annoy you less.
Having said that though, If I was olene I'd go around submitting patches for the current major IRCDs that fixed these problems and hope they got incorperated :) Complain once, thereafter, fix.
-Ashen |
|
| Back to top |
|
|
Roku
Newbie
Joined: 18 Apr 2005
Posts: 92
|
| Posted: May 18, 2005 12:04pm Post subject: |
|
|
Ashen, perhaps your right. However, "bitching" in a public forum tends to leave a negative impression on others. Which means she maybe venting harmlessly but someone else may just view her as annoying. If that was a form of venting, why not use it in a positive manner? such as put it in an email and send it to IRCd dev teams, maybe they will agree with her and team up and finally revise the RFC.
Impressions one leaves are just as important as any product that person yields. For example, I stopped using IRCServices because IMO Andrew Church is a snide, condescending person. I know people that have switched from mIRC to other clients because they feel that way about khaled. It sucks .. but it's true. |
|
| Back to top |
|
|
[DiMENSiON]
Eleet
Joined: 06 Sep 2003
Posts: 621
|
| Posted: May 18, 2005 12:41pm Post subject: |
|
|
| Life sucks, deal with it.. And we where all n00bz once, well most of us. |
|
| Back to top |
|
|
rachelllorenna
none
Joined: 14 Jan 2005
Posts: 3
|
| Posted: May 18, 2005 3:45pm Post subject: |
|
|
| Quote: | | 1. 481 numeric... "Permission denied -- You are not an IRC operator". God. I know thats how it is in RFC1459, but can we at least shorten it to "You are not an IRC operator". You don't have "permission denied -- You are not a channel owner" or "permission denied -- You have not registered". Shorten it!. IIRC I noticed that recently one of the major networks (i think ircnet) had shortened it to just "Permission denied". In any case.. in all 13 of my crappy ircd's, i've always used ":You are not an IRC operator" |
Worst-case scenario, a "Permission Denied - You're not an IRC operator" message takes:
| Code: | :some.server.net 481 frequency :Permission Denied - You're not an IRC operator
^ ^ ^ ^ ^
1 63 5 30 48 |
Which is a grand total of 147 bytes. If you don't like it, change it in your client. It's not even an issue. As well, I have no doubt that your ircd installations are indeed 'crappy', and I haven't even gotten to all your other points.
Your colorful language really makes your point seem valid. *rolls her eyes*
| Quote: | | 2. Cannot send to channel. Why the hell not? I'm not sure if unreal does this or not, but i somehow got the idea early on that the "(+m)" in "Cannot send to channel (+m)" was standard, but on recent inspection.. it's not! Think about.. now adays, you have +m +M +b +n +c +C and a whole other host of limits on what you can say. At least include a little caption (like "Cannot join channel (+i)") so you know why. |
This is a good point, but most ircd's will list reasons or send different numerics depending on the reason. The problems occur when ircd's decide to implement a plethora of non-standard modes, such as +M, +c, +C and whatever else.
Write something in your script to check channel modes if this is an issue, or just /msg an op asking nicely. If this really bothers you, speak to ircd developers personally about implementing such a feature. But based on how you wrote your first point, I doubt your relations with other people are good enough for this to be feasible.
You get more flies with honey than you do with vinegar. Unless you learn how to deal with people (read: be less anti-social), IRC (and indeed life in general) is probably not the place for you. Find a hobby that doesn't involve interaction with humans.
| Quote: | | 3. People using NOTICES where they should be using NUMERICS. I cringe when i connect to a server and i get "*** Total global connections: 49489 (39393 max)" I'm pretty sure thats what RPL_GLOBALUSERS is for, and though its not in the RFC, it's pretty damn standard. |
Agreed. Granted, server notices are useful for some things; and numerics have in the past been created to replace notices. Sometimes servers are old or their developers don't believe that using such a numeric is preferable. A philosophical conflict, in many cases.
| Quote: | 4. "-mesa.az.us.undernet.org- on 6 ca 1(4) ft 10(10) tr".
What the !@#$ is that?!? It's not so much that I can't understand it, but its the fact that it's noticed to you, everytime you connect.
Does one line of text really bother you that much? Re-write your IRC client to hide such notices if they really bother you.
The tao of irc states "The client should always respond the luser with messages that will not astonish him too much. The server likewise. If the server does not, then it is the clients job to explain what the server says." |
I don't think you realize that that document is a normative, not a standard. It was written by people with a lot of free time on their hands for their own personal enjoyment. The document makes a good point, but it is by no means a standard. It was written when 'irc' by jto was the only ircd software available. That is no longer the case; with many teams developing many different servers.
Besides, many networks end up using their own numerics without consulting anyone else, which results in a lot of conflicting numerics. It's really not a good idea to make it worse. Confusing clients with the same numeric meaning different things in different contexts (based on network/server software) would cause a lot of problems - using on-connect notices solves this issue nicely.
| Quote: | | I have not yet seen a client that interprets this for me. |
Also, it's better to get an unexpected NOTICE than an unexpected numeric.
| Quote: | 5. Quit message prefix fux0red up syndrome
Ok. Quit message prefixes are ok (e.g. "quit:"). I especially like dancer's solution of simply enclosing it in quotes. My annoyance is with unrealircd and other ircd's that will send the entire quit message as "Quit: " with the space and all, if you never specified a message. I mean... "*** blah (blah@blah) quit (Quit: )" looks really shitty, IMO.
Which brings us into #6... |
Hide such messages if they really bother you. The fact is, such a small thing isn't worth changing ircd for. The current system works perfectly. This stuff is better done on the client side than the server side - it's a bad idea to put a lot of load on the server for nothing. Using quotes (as on freenode/dancer-ircd) works, but it's more work than simply prefixing "Quit:". You would need to do comment[MAXLEN-1] = '"'; comment[MAXLEN] = '\0';
The prefix is just to disambiguate between a forced KILL/KLINE/etc and a user-generated QUIT. For what it was intended, it seems to be working nicely.
| Quote: | 6. The default message being your nickname.
Who the !@#$ thought this shit up? What's wrong with an empty string? i wanna know what the !@#$ is wrong with "*** $#$@!@# was kicked by olene ()". This isn't webtv.. the kickee is perfectly capable of seeing who kicked them. Clients are perfectly capable of parsing an empty string (RFC1459 even notes that the trailing parameter may be an empty string). So WTF? Some servers, thankfully, do allow empty strings. Some only allow empty strings if I supply the ":" with nothing after it. And some simply will always fill in my nickname. |
If this really bothers you, specify a reason. Maybe a '.' or something. It doesn't matter who kicked them anyway, most of the time kicks are followed by bans to get rid of abusive/annoying users such as yourself. I'm sure you see yourself being kicked a lot, and don't want to see your kicker's nickname twice. Do yourself a favor and /quit, remembering to specify a message so that your quit doesn't end up being 'Quit: '! Removing your IRC client would probably be the best thing you can do for the IRC community at large.
| Quote: | | 7. Ritz crackers that are too salty. |
I'm not even going to say anything about that.
| Quote: | | 8. Servers without WATCH. Bandwidth saved alone from elimination of periodic ISON queries is worth it. |
Again, a philosophical reason. WATCH is a broken implementation (refer to ircd-ratbox-2.1.x documentation for discussion regarding this matter.) There are, however, alternative commands that one can use for notification systems (ie, ircd-ratbox-2.1.x offers MONITOR.) ISON is pretty hackish, and certain servers are removing it, slowly. It was created because a) it was easy to do and b) it didn't use memory on the server side. IRC is an old protocol with many holdovers from older times: live with it.
With that too, if you're THAT concerned about bandwidth consumption, you probably shouldn't be running a service that a) consumes a lot of bandwidth in and of itself and b) attracts denial of service attacks. If a few extra megabytes are really that big a deal to you, stop running ircd now.
| Quote: | | 9. I ran out of things that annoy me. But i'll think of more. |
I can certainly see why many people think of women on IRC as idiots. You certainly don't make us look very good. I try and try to prove that women are every bit as capable as men on male-dominated environments such as IRC. However, women such as yourself continually make me (and other women on IRC) look that much less credible. I gained a reputation on Linknet and elsewhere about how much I complain about things, but your complaints are both illogical and a waste of everyone's time. If you cannot substantiate your complaints with more than "I don't like it," then nobody will listen to you. Moreover, I don't know why I'm devoting so much time to pointing out your fallacies; perhaps I just pity you.
'angelic' definitely makes a good point: | Quote: | | Now i know why men think women are whiny. |
|
|
| Back to top |
|
|
olene
Newbie
Joined: 31 Jul 2004
Posts: 61
Location: olene on DALnet
|
| Posted: May 18, 2005 5:01pm Post subject: |
|
|
| The original reply I had here was out of line. Apologies. Please lock this thread. |
|
| Back to top |
|
|
trelane
none
Joined: 18 May 2005
Posts: 1
|
| Posted: May 18, 2005 7:24pm Post subject: |
|
|
| Quote: | | The fact is, such a small thing isn't worth changing ircd for. The current system works perfectly. This stuff is better done on the client side than the server side - it's a bad idea to put a lot of load on the server for nothing. Using quotes (as on freenode/dancer-ircd) works, but it's more work than simply prefixing "Quit:". You would need to do comment[MAXLEN-1] = '"'; comment[MAXLEN] = '\0'; |
I've always noted an incredible amount of paranoia about adding simple functions to improve human readability of ircd output. . This takes little effort, a small calculation to deal with odd or improper output from the quit message is similarly little output. IRCd's ran fine on a 486 (or as well as could be expected from the code 15 years ago). IRCd's, even those on large networks are not about to run out of cpu power. An arguement that we must leave malformatted output because "it eats cpu cycles" is asanine.
| Quote: | The prefix is just to disambiguate between a forced KILL/KLINE/etc and a user-generated QUIT. For what it was intended, it seems to be working nicely
|
Odd since most IRCD's attempt to obfuscuate "unnatural" parting of clients via kill kline dline etc to prevent DDoS attacks on the server. The "remote host closed connection" code in Hybrid 5,6, and 7 being an excellent example of this. Dancer can even kick users from channel and send a PART to the other clients in the channel. Dont trust an IRCD to differentiate output, this seems foolhardy at best. |
|
| Back to top |
|
|
codemastr
Idler
Joined: 05 Feb 2004
Posts: 353
|
| Posted: May 20, 2005 11:55am Post subject: |
|
|
| Quote: | | Again, a philosophical reason. WATCH is a broken implementation (refer to ircd-ratbox-2.1.x documentation for discussion regarding this matter.) |
You're the first person I've ever heard refer to WATCH as "broken." The author of ratbox's MONITOR refers to it as "flawed" (which is very different), though he/she never actually provides any reason for this opinion. So where exactly is the discussion of this issue? |
|
| Back to top |
|
|
Guest
|
| Posted: May 20, 2005 9:23pm Post subject: |
|
|
4. "-mesa.az.us.undernet.org- on 6 ca 1(4) ft 10(10) tr".
They are IPCheck statsistics
on <number of connected users on your ip> ca <recent connection attempts>(<max allowed>) ft <freetarget>(<max>) |
|
| Back to top |
|
|
Guest
|
| Posted: May 20, 2005 9:25pm Post subject: |
|
|
just to extend a little bit further
freetargets: Initial targets when you connect, ever seen those notices when you cant join a channel or message a user for X amount of seconds on undernet? you have run out of targets |
|
| Back to top |
|
|
null
none
Joined: 21 Jun 2004
Posts: 26
|
| Posted: May 20, 2005 11:16pm Post subject: |
|
|
| Rightio, Guest.. I'm sure everyone in the IRC world understands that. I sure don't. |
|
| Back to top |
|
|
w00t
Eleet
Joined: 09 Jun 2004
Posts: 699
Location: Nowra, Australia
|
| Posted: May 21, 2005 6:53am Post subject: |
|
|
| Then why not say "connection attempts" instead of ca? Why send it to every user? (ie: who the hell cares?). |
|
| Back to top |
|
|
|
|
| |
Register
Log in
