Re: [EPIC]notify interval proposal
On Monday 03 February 2003 10:32 pm, Jeremy Nelson wrote: > It seems that the current compromise solution seems to be acceptable to a > large number of people: [snip compromise that seems to be acceptable to a large number of people] This is fine by me. I would support intervals of less granularity than 60 seconds, and the obvious extension of that is that whatever limits the client sets, the user is able to circumvent it trivially; thus config.h is a reasonable place to put it. I would prefer it there rather than not limiting the /set, for a couple of reasons; - Someone is going to /set notify_interval 1 and wonder why they lag / the server keeps disconnecting them. - It's preferable to warn users exactly what they're doing, for that reason and others, and a comment above the option in config.h would be a reasonable way to achieve that. Regards, Edward. ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list
Re: [EPIC]notify interval proposal
nsx said: >> Yes, I am campaigning for both better precision, and a change of the lower >> notify interval limit. I suppose they sort of go hand in hand, since to >> either abolish or lower the lower limit, you'd probably need to provide >> better precision anyway. Kev said: >Personally, I don't see the problem with the first part--increasing the >precision. As for a lower limit, I would strongly suggest that it should >not be less than 30 seconds. It seems that the current compromise solution seems to be acceptable to a large number of people: * Remove NOTIFY stuff from the "top_of_minute" timer, into a second timer. * Allow user to modify timeout interval of that second timer via /SET NOTIFY_INTERVAL * Add a #define in config.h which will set an absolute minimum value of /SET NOTIFY_INTERVAL, so that the user can be assured that a script will not set a value lower than they wish to have. * The default value of this #define will be 60 for backward compatability. Of course with every compromise, not everybody got everything they wanted, but this seems to be very much in the middle of what everybody hoped for. Jeremy ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list
Re: [EPIC]notify interval proposal
> Yes, I am campaigning for both better precision, and a change of the lower > notify interval limit. I suppose they sort of go hand in hand, since to > either abolish or lower the lower limit, you'd probably need to provide better > precision anyway. Personally, I don't see the problem with the first part--increasing the precision. As for a lower limit, I would strongly suggest that it should not be less than 30 seconds. -- Kevin L. Mitchell <[EMAIL PROTECTED]> ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list
Re: [EPIC]notify interval proposal
On Mon, Feb 03, 2003 at 12:48:30PM +, Edward Brocklesby wrote: > Correct me if I'm wrong, but the impression I got from nsx on IRC and in his > mail was that he wanted a notify interval of less than 60 seconds, which is > what I'm discussing here. > > I wouldn't have any problem with allow more than +-60 second precision. > Yes, I am campaigning for both better precision, and a change of the lower notify interval limit. I suppose they sort of go hand in hand, since to either abolish or lower the lower limit, you'd probably need to provide better precision anyway. -- John ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list
Re: [EPIC]notify interval proposal
On Sun, Feb 02, 2003 at 02:08:44PM -0500, Ben Winslow wrote: > Edward Brocklesby wrote: > [snip] > > and I'm not likely to talk to someone who's here for 30 seconds and then > > gone again > [snip] > > > > -larne. > > Try using dial-up service over the US's PSTN. ;) Nothing RFI chokes, modem strings, and harassment of the local telco can't fix ;) ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list
Re: [EPIC]notify interval proposal
On Mon Feb 03, 2003; 12:48PM + Edward Brocklesby propagated the following: > On Monday 03 February 2003 7:27 am, [EMAIL PROTECTED] wrote: > > having finer granularity doesn't necessarily mean that it would be used to > > have notify interval shorter than 60 seconds. it would also allow notify > > intervals of, for example, 90 seconds. there still might be a lower limit > > of 60 seconds. > > Correct me if I'm wrong, but the impression I got from nsx on IRC and in his > mail was that he wanted a notify interval of less than 60 seconds, which is > what I'm discussing here. > You are correct. Personally I'd like to see this tuneable down to 20-30 seconds (at lowest). It would be nice to have a more precise notify level. I also think that if servers find this load burdensome than they should either add code to disconnect users who do this or add WATCH support, both of which are fairly trivial things to do. > I wouldn't have any problem with allow more than +-60 second precision. > > Regards, > Edward. > ___ > List mailing list > [EMAIL PROTECTED] > http://epicsol.org/mailman/listinfo/list -wd -- chip norkus; unix geek and programmer; [EMAIL PROTECTED] "question = (to) ? be : !be;" --Shakespeare http://telekinesis.org/ ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list
Re: [EPIC]notify interval proposal
On Monday 03 February 2003 7:27 am, [EMAIL PROTECTED] wrote: > having finer granularity doesn't necessarily mean that it would be used to > have notify interval shorter than 60 seconds. it would also allow notify > intervals of, for example, 90 seconds. there still might be a lower limit > of 60 seconds. Correct me if I'm wrong, but the impression I got from nsx on IRC and in his mail was that he wanted a notify interval of less than 60 seconds, which is what I'm discussing here. I wouldn't have any problem with allow more than +-60 second precision. Regards, Edward. ___ List mailing list [EMAIL PROTECTED] http://epicsol.org/mailman/listinfo/list