A voluntary filter?

[Lunasa]

Well, there ya go, the title gave it away. XD

Well, anyways, a couple of others and I think a voluntary filter would be a great idea for these boards. Advantages of a voluntary filter?

Well, for one, this can get particularly annoying:

For Wren just fumbled the ball yet again. Kallenbach s******ed trollishly. As a result, Sirius took a bite of Slytherin’s cake and tossed his bat at the back of Kallenbach’s head. A whistle blew as Kallenbach, no longer s******ing was darting at Sirius, who had a superior sort of smirk. Madam Hooch flew between them.

Well, just a couple notable words, anyway. I said "sni[g]gered" above.

That and people should have the ability to toggle that sort of thing. You see, with the toggling, people who are offended by the words can still filter them. And vica versa for those who don't care about the curse words. If this requires ALOT of extra coding, sorry, but I think it's a pretty good ideaa nd it should be done.

[LocoColt04]

I can change it so that the filter blocks only the word by itself and its common variations, rather than when it is a portion of other words. I can do this for others as well (some, it's already set that way for). Also, it only takes me about 30 seconds to change, so no worries there.

As for writing up a toggle switch... it shouldn't be too difficult to add an option for users to turn the filter on and off. I'd have to look at the code. I'm a little more accustomed to coding for freeware, so there are some key differences between vBulletin and the software I usually code for.



On a tangent,

I know what some other people are going to follow up with though... "we're all grown up now, can't you just turn off the filter?"

And what happens when some 13 year old's parent strolls by and sees f-bombs on the screen? Look, we all know they say it anyway but the filter's been there for a decade. I don't see why we'd remove it now.





Snigger. I typed this before making any changes.
...seems to work fine. I double-checked the filter and the word "nigger" is already set up the way I was going to set it up, so I didn't change anything after all...

[Merlin]

The filter is done right when you submit a post. It physically alters the text before it even goes in the database. There is no simple way to have it dynamically filter the text only for certain people as it prints each post in a thread. The server lag would be ridiculous.

FYI: You could always type "snicker" instead. It's the same word.

[Lunasa]

Ahhh. Well, I knew that snicker can be rewritten, that was just an example, and actually, I caught that in the archives. Apparently the system was updated since then.

I'm wondering who determines this 'best answer' thing, though. It's strange that one says it's possible, one says it isn't. How can any of them possibly be the best answer, since they simply both address possible outcomes? And I thought it was the threadmaker who determined the best answer, not the Mods/Admins. I certainly don't approve of the instant "no".

It's too early to lay this to rest, anyway, unless we get more pros and cons. In the toggle thing, wouldn't that actually be less on the server, merely because everyone that browses does not require the filter to be on? Isn't that sort of things based on logic programming with a yes or no variable?

[SOLDIER #819]

The thread maker, admins, and possibly super mods can all choose the best answer, but you probably already knew that. I wonder if one can change the best answer after it's been selected and if one usergroup has priority over the other in terms of selection. Admins can obviously do anything, but that's not my point.

What Merlin is trying to say regarding filters is that the censor is applied before the post is logged into the database. I guess an example would be if I typed **** right now and then attempted to edit my post. **** will still be censored in the edit window, because the post is regulated before submission to the database. The database entry only ever sees asterisks. It's the same as having too many caps. A "post check" modifies the caps to lowercase before submission, not after.

If this was not the case, then every single time a user viewed a post with curse words (or too many caps), the board would have to change the post on the fly to suit the user's preferences. It is possible to have a user-opted censor, you see it in MMOs all the time, but it is not practical by any means for a server like this. Programming can have it go either way, but two ways aren't necessarily equal in pros and cons.

If you are programming for vB, would you go the route that could decimate a potential customer's server, or leave it to the customer to modify it afterwards if s/he so desires? It's a matter of efficiency. Even if this vB could do it, there would most likely be consequences in one way (performance) or another ($$$) for the person hosting the board.

Something like that. Correct me if I'm wrong.

Edit: Here are the default censorship options for vB, per the manual.

http://www.vbulletin.com/docs/html/v...s_group_censor

[Merlin]

Cesar hasn't been working with vBulletin for 6+ years, so you can't fault him for not knowing how the censor function works under the hood off the top of his head. On the other hand, I do, and can tell you without a shadow of a doubt that what you ask would require modifying a number of files every time vB gets an update, it would also cause a server hit that is not worth the end result.

It's too early to lay this to rest, anyway, unless we get more pros and cons. In the toggle thing, wouldn't that actually be less on the server, merely because everyone that browses does not require the filter to be on? Isn't that sort of things based on logic programming with a yes or no variable?

It isn't logic programming. It's programming based off the magnitude of a function identified with the term that increases fastest relative to the size of the problem. A simple true/false conditional in itself is not that time consuming. What is is the function contained within. Obviously, decreasing the number of times a function runs is a good thing. What you are suggesting doesn't decrease the number of times it is ran, it increases the times it is ran by quite a bit. Even if only half the people ran the filter, the difference is ridiculous.

The censor function uses Regular Expressions to do its work. Regexp's time is denoted by O(m*n), where m is length of the regexp and n is the length of the string.

The filter function runs a regexp to separate out all the censor words into an array. Then it loops through each censor word and runs another regexp on the post itself. This leaves you with roughly

O(m*n) + (O(p*q) + O(r)) == O(m*n + p*q + r) time to filter a single post, where:
m: length of split for censor (#[ \r\n\t]+#)
n: length of censor string
p: length of a specific censor
q: length of the post
r: number of censor words

Now, the way it works currently, this filter only has to run on a post right when it is written (or if it is edited). so, the penalty of O(m*n + p*q + r) is applied to the time it takes to make a post and will never affect the time it takes to display a thread. Therefore, the amount of time to display a thread will be simplified to O(s), where s is the amount of posts in a thread display page.

So, if we change it to filter posts as you view a thread, the time penalty will become:
O(s) * O(m*n + p*q + r) or O((m*n + p*q + r)*s). Even without doing an example, this is quickly proving to be an inefficient method of filtering, since this would be done every time someone views a thread.

[LocoColt04]

In the long run, what matters is that the words in the filter are far fewer than they were four years ago. The list is surprisingly small, and almost entirely made up of single-word instances (so that, in most cases, it won't censor the little word inside of a big one). The only complaint I ever really get is from those 20+ who just want the thing done away with altogether.

On the programming side of the field, I just learned something today. ^^

[Merlin]

I always found it hilarious when cocktail got hit with a censor. It's like you're talking about some new sleezy drink name. And hell, so many people at TFF are gifted at taking things out of context, figuring out censored words should be a cakewalk. [/grin] But in all seriousness, I think this issue is pretty much resolved. You've made it as streamlined as possible.