[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: changing channels (Re: no split modes patch)



On Jan 07, Joern Westermann wrote:
| > i think history has proven time and time again, that such
| > requirement is a problem.
| 
| Sorry, I disagree. We have a finite number of servers.

true, but...

|                                                        If most of the
| server admins agree we could set a date for the global move to 2.11
| (e.g. 2000-06-01 00:00 GMT).

you do realize that attempts to reach such an agreement have
never succeeded before?

|                              It should be possible for every admin to
| compile+install the new version and even make a few tests.
| If I understand uid right then users can still be collided from a 2.10
| in a mixed 2.11-2.10 network. And it'll probably take a year for the
| last server admins to upgrade to 2.11 if you don't force them somehow.

It is not my job to force them.
it is the job of whoever/whatever administers a network.
(yes, i'm ducking but i think i am entitled to.  there are
also other issues which make such a move rather
difficult/problematic)

| Normal users are lazy. They don't want to learn anything new about
| channels if they don't have to. We had this discussion with users dozens
| of times after a channel takeover ("use #channel:*.de, use !-channel"),
| but after the channel was free again, they joined the old channel.
| At the current level of client support for !-channels I can really
| understand this!

i think i also forgot to mention that i can very well see
users not liking !channels because you can't ride splits.

| What we need is a user-proof client support, where a users gets a dialog
| box with all channel alternatives, if he tries to join a certain !- (or
| #-) channel. For example: User tries to join !germany. But due to
| several splits there are now 3 different !germany-channels. A dialog
| pops up with the number of users on each channels, user clicks on
| largest channel, voila!

if we so easily get to duplicate !channels, then i'll
consider !channels a failure.
however, a client should be smart enough to list all
duplicates when it happens.