I'd really like to know why it's not being very random.

Is it possible that there's no lower limit to the randomization? What I mean by that is that I once set up a random number generator that generated values from 0 to MAX, then forgot and numbered my records from 10 to MAX, with the result that record 10 got selected a lot more than any of the others, because it would show up for values of 1 - 9, as well as 10.

I don't think Mickie looks suspicious. They're user number 390, for one thing.

Yeah, Mickie was one of the very first newbies that signed on when Joss first showed up, although she's only posted twice, and only in TTT.

edit: Whoa, speaking of record numbers...

Is it possible that there's no lower limit to the randomization?

Thing is, I think there's only been two different amounts of quotes. Of course, we can't pinpoint the date of stickiness, but I lean towards maybe a MySQL upgrade that changed an algorithm that's peeing in our ointment.

One approach that would help work around the problem is to keep a list of all the quote numbers, and when serving up a random one, remove it. When the list is empty, refill it from original list. That way, each quote only gets served up once per cycle.

A similar approach might be to create a table of the quote indexes repeated one thousand times and just march through it. That will look random to any single viewer.

Or, you know, live with it as it is.

Would it be possible to, when a post has been bookmarked, where it says MARKED, have a little link right there that would allow you to un-mark it without going to your "view bookmarks" page?

Someone I know is looking for the name of a good webhost. We use HostRocket? We like HostRocket?

~j

We use HostRocket. We've had far less downtime than WX, or TT, for that matter.

I've made another change to his ID in the database -- I don't have the code to hand, but it should make a difference and log him out.

So what happened earlier when we weren't able to stop Michael from posting? I'm really uncomfortable with the idea that only ita can "instantly" deactivate someone. 'Cause, what if she's not around? Can we put a high priority on fixing the problem?

And these questions are really directed at ita, because she coded it, so I probably shouldn't refer to her in the third person...

t edit Would it work to have a button on the edituser page that logged out a specific user?

We use HostRocket. We've had far less downtime than WX, or TT, for that matter.

Just to note, the "downtime" on WX and TT were failures of the software and/or database back end of the sites.

That has never happened here, except for, literally a second or two during which ita was uploading a new version of an essential file.

What we have had here, although it hasn't been for a while, touch wood, is communication outages, in which our server was not connected to the internet for some reason.

That, conversely, never happened on WX/TT to my knowledge.

I can't remember if we had those outages explained -- were they Hostrocket's fault or was there some backbone/pipe/major issue?

Jon, I can't instantly deactivate a poster. The tweak I tried didn't work.

That's going to have to be a modification.

John -- I know I've been unable to get to pre-Phoenix Buffistas at least once because of a connectivity issue on their end. Which I was sure of because it turned out another discussion board I frequented was hosted by the same people.

To be honest, I haven't paid much attention.

For personal use, I'd say Hostrocket is just fine. For an application like this which gets almost 24/7 use, you want to do what you consider reasonable in the way of increasing redundant hardware and connectivity, which takes you out of most standard hosting packages.

That's going to have to be a modification.

Does my idea of having a "logout" button on the edituser page make sense?