watch problems

[Aaron Sethman]

On Tue, 15 Nov 2005, Piotr NiżyÅ~Dski wrote:

> Aaron Sethman napisał(a):
>> Then why bother implementing something when you are going to do it 
>> half-assed.  You'd be better off not shoving out WATCH in 005 and letting 
>> clients work correctly using ISON than breaking clients who expect WATCH to 
>> behave like WATCH and not something sort of like WATCH.
>
> It's just like you said we can't return RPL_LOAD2HI in reply for /LIST, 
> because it's not included in the RFC. Or any other kind of rate limiting, 
> because the "standard" doesn't let us to do so.

I never said any such thing regarding LIST, or not that I recall at least.

>
> Here, as far as I know, WATCH is not a matter of any absolute standard like 
> RFC, so we're more than welcome to improve its implementation.
>
> Please point me at least one widely used client "for plain people" which 
> needs to do multiple 'l' in WATCH. It's not a feature to be used for good 
> reasons, but rather a toy for kiddies and abusers.

Can you say for certain that a client doesn't do this?  And its something 
called a de facto standard.

> Finally, we just don't like inventing new 'standards' where it's possible to 
> evolve well known ones. Going your path, we'd have to give up on /LIST and 
> others and introduce /ELIST, /ELUSERS etc. with some limiting allowed. This 
> is not the way unless we want to be the only people up to current IRC 
> 'standards', with others seeing it as a complete mess changing with every 
> ircd revision.

So you don't invent new 'standards' you go and implement something in a 
different way from everyone else and then go and busticate a bunch of 
clients.

I'm not sure why I'm bothering continuing on with this thread anyways.  If 
you want to break clients go for it.  After all an ircd is supposed to be 
about how it strokes the coder's ego and not about the actual users 
experience anyways.