Posting URLs in Messages

what does WW prefer

[atadams]

I notice that when some members put a URL in their post they "hide" part of it (eg, h**p//www.foo.com) so that the BestBBS app doesn't parse the URL and create a link. Is this what the powers that be at WW prefer or do y'all even care. I've always figured that if WW didn't want URLs to link, they wouldn't link them.

[mcavill]

i guess some links are OK and some, spammy or promotional, aren't. So as a default they parse them, but then the mods kill them if they're against the TOS [webmasterworld.com...] #13

It's fair enough especially when you consider the PR a lot of the threads here carry, it's bit a bit of a target for message board spammers.

[Mohamed_E]

After about 18 months here I am still occasionally unsure. In those cases a sticky to the mods, either asking for advice or inviting them to edit if they think it is desirable, should do the trick.

My general impression is that:

• Links to your site (or to a competitor you are accusing of spam) are clearly not allowed.
  
• Links to relevant news stories are just as clearly allowed.
  
• The delicate balance is with tools, software or anything similar. We are all anonymous, and there is always the danger that a "Just found this great piece of software!" post comes from the owner of said software.
  

Brett's post on professional forum spammers [webmasterworld.com] will give you a good insight into his thinking on these issues.

[Brett_Tabke]

I'd prefer they were linked. Sometimes we will delink them if they are to a commercial product. We might also delink it if it is to a particularly troubling company or product.

[skipfactor]

I think I know the answer but is anchor text preferred over a raw URL for both internal thread linking & outgoing news links?

[Yidaki]

skipfactor, i'd say a good anchor text makes it more readable and allows you to focus what's important on the linked page. Even if the link wouldn't get followed by any bot and if the anchor wouldn't have any se value, i prefer using a goot anchor text over raw urls. It's just good webmastering - that's why we are here. ;)

[Brett_Tabke]

The raw url is always preferred. The only time it isn't, is when the url is too long.

If I could take out the option to do it any other way, I would.

[Yidaki]

>If I could take out the option to do it any other way, I would.

i don't get it ... why?

[trillianjedi]

i don't get it ... why?

You can see instantly just what you're clicking on.

That's why I prefer it anyway......

TJ

[skipfactor]

All this time I've been fumbling with the blasted url=...copy-&-paste it is. Thanks Brett.

[Yidaki]

>You can see instantly just what you're clicking on.

Browser status bar ...?

However, i gonna stick by the rules then and post raw urls instead if possible.

[Brett_Tabke]

That is right - to see the url...

[sun818]

Brett, why not take out the URL= function altogether then? And then raw links only display if you put http:// or https:// in a forum post?

[Brett_Tabke]

[google.com...]

[Yidaki]

Don't want to ride on this issue but why not add a space after xy characters and split the url - what google does? I know the url would get broken but if it outweights the trouble ....

[Brett_Tabke]

> why not add a space after xy characters

Trying to reduce complexity and increase speed (not vice versa)

There are 8 regexs that control [URL]'ization now.
Ever notice how fast WebmasterWorld is? (eg: the fastest board on the web on a moderate level server under the most intense and constant load of any forum on the web).

To "add a space" would be at a minimum another regex that could use up to a meg of system ram to process. That isn't an issue at lower levels, but at 2000pvs a minute, it is the difference between life and a dead server.

[Yidaki]

Yah, that makes sense. Wow one meg ram for one additional regex!

Thinking: are the posts stored as is and then processed on every load? Afaik that's how most bbs software works, no? Wouldn't it be another thing if the software would rewrite url's and such at posting time? I mean do all processing (all regex' like badword filtering, url rewriting, styles etc.) right at the moment before the post gets saved the first time? So that if the posts are loaded for viewing they are allready formatted? If a post gets edited, just do it reverse - reformat into bb code. However, i'm just guessing. No reall idea how this all works. You are the specialist. ;)

[Salsa]

Yadaki, I'll wait for the answer to that one, too. In my own content managers, I format the data before it is stored in the database for the reasons you said. It's a bit more trouble writing the management scripts to reverse the regexs for the editing process, but it saves a lot on server load because you don't have to run so many regexs for each page view. If there's a down side to that, I'd like to know it.

[killroy]

I'm a fan of caching... have caching middle-end that caches the formatted posts. It's just text... even a million posts aren't much to store in a db formatted and un-formatted.

SN

[Brett_Tabke]

> Wow one meg ram for one additional regex!

Perl always prefers to spit more ram at any given situation. (a meg might be over stating it though - even on a worst case)

>processed on every load

Yes.

> Afaik that's how most bbs software works, no?

Most yes. Some _can_ generate static content.

> at posting time

Yes, but since you have to generate or fill out the board template, then the added work of filling in the msg is easy and gives you the flexibility.

> I format the data before it is stored in the database

You bet - it is always better to do so where you can. While here, we are a slave to a flexible and even user modifiable template that needs to be built on-the-fly (fonts, colors, etc).

> caching

If I were to run nph scripts, I would allow caching of those that had not changed by using "if modified" headers appropriately.

[Yidaki]

Thanks for discussing with us about this interna, Brett. I again learned someting new.

Btw: awesome job - this board is the fastest i've ever seen! Nuff said. ;)