I used Google to search the board, and some of the results had ‘fartypants.straightdope.com’ as the domain name. Is there an interesting story behind that?
It’s the codename for the new server.
Oops. Sorry, that happens when you’re old.
I’m still sticking with the theory I hypothesized in this post.
Summary: If anyone ever posts even one such link of the form xxxxx dot straightdope dot com, in any post anywhere (which several people have done already in these threads on the subject), and if anyone else (including a search engine spider) then opens the page via that link, then all the vBulletin-generated links within that page will also be generated as xxxxx dot straightdope dot com instead of boards dot straightdope dot com. Thus, it was suggested (I think it was Lord Feldon who said it first, IIRC) that the entire board ends up getting indexed under the new bogus domain.
Some of the known “xxxxx” values for which this seems to have happened so far are:
board; ards; senegoid; and fartypants.
And no, it wasn’t me that first posted a link to senegoid dot straightdope dot com.
Maybe they can fix this when the front end is upgraded. Just have everything do a hard redirect* to boards.straightdope.com. The fancy links will still work, but Google won’t index them.
*A hard redirect means that the URL in the address bar would still be boards.straightdope.com, even if you originally went to imalittleteapot.straightdope.com
It already does that.
I just tested this. fartypants.straightdope.com gets redirected to Straight Dope Message Board - Your direct line to thousands of the smartest, hippest people on the planet, plus a few total dipsticks. as do random strings of characters preceding .straightdope.com.
Yeah it has nothing to do with a ‘hard’ redirect. The answer was provided here.
Or <My favourite obscenity starting with F>.straightdope.com. Should make search results interesting!
No, you and Canadjun and BigT didn’t read my post above (Post #5) carefully, and more to the point, you didn’t read my original analysis that I linked to very carefully, and you didn’t do the same experiment that I described there. And spinky, whose answer Crazyhorse linked to, didn’t explain what was going on, he just offered a suggestion how to make it not happen any more.
Yes, if you type this into your address bar: fartypants dot straightdope dot com
then it gets re-written as: boards dot straightdope dot com
BUT
(I wrote this same BUT in extra-large bold all-caps letters the first time BUT you didn’t listen)
If you make a post, and you include in your post a link to another specific post using the fartypants URL then the problem happens that I described.
Look very carefully at the two links that DCnDC gave above in Post #3 above. They both point to specific pages within SDMB. The first one goes to boards dot straightdope dot com but the second one goes to fartypants dot straightdope dot com.
Now CLICK on the second link and see where it takes you. You get to the right page, but the URL in your address bar doesn’t get changed to “boards” – It still says “fartypants”.
BUT WAIT! It gets worse!
Look through that page, and examine all the links you can find. NO, not the links that various people have put into several posts in that thread. I’m talking about the links that vBulletin itself wrote into that page. Look at the line:
Straight Dope Message Board > Main > About This Message Board
near the top of the page. Look at every user name at the top of each post. Look at the post number at the top-right corner of every post. Look at the little right-arrow symbol at the top of every quote box. Look at the “Quote” button and the “Post Reply” button. Every one will have a URL that points to fartypants! Look at every other link you can find on that page. Many (not all) of them, likewise.
If Google finds just ONE page anywhere at boards.straightdope.com that has a fartypants links in it (as in Post #3 above), it will follow that link just as I have told you to do. And it will find all those other fartypants links and follow them. And ultimately, it will end up re-indexing all (or maybe some substantial part) of the board anew, all under the fartypants URL.
You guys!!! Senegoid noticed something and nobody is giving him credit!! Fonts are changing sizes! Bolds and italics are happening! UNDERLINES!
[ol]
[li]Is this still happening?[/li][li]Could this be the reason our server is being overloaded?[/li][/ol]
[QUOTE=Senegoid;16397042
Yes, if you type this into your address bar: fartypants dot straightdope dot com
then it gets re-written as: boards dot straightdope dot com
BUT
(I wrote this same BUT in extra-large bold all-caps letters the first time BUT you didn’t listen)
If you make a post, and you include in your post a link to another specific post using the fartypants URL then the problem happens that I described.
[/QUOTE]
You don’t even need to link to a specific post at fartypants, you can type ‘fartypants.straightdope.com/sdmb/’ into the location bar and you will remain at fartypants without a redirect. But this has nothing to do with Google indexing ‘everything’.straightdope.com. Everything except www.straightdope.com and chicago.straightdope.com are wildcarded at the DNS and as you have noted once Google gets ahold of one link to something.straightdope.com it might go ahead and follow all the links it finds there which will lead to more and more something.straightdope.com indexes.
The way to solve this is to tell Google the canonical (preferred) domain that should be indexed as spinky has been quietly pointing out for like a year every time this comes up.
I’m pretty sure this can even be done via google webmaster tools without having to change any HTML on the site.
This isn’t a startling new revelation, nor is it unique to the SDMB. It comes up all the time. For example http://www.example.com/page.html and http://www.example.com/page.html?parameter=1 would be considered two entirely different websites by Google even though they return the same content. It doesn’t require any special DNS wildcarding or redirects to run into this most basic of website design problems. It is described at length in RFC-6596 and even has its own Wikipediaentry.
We could just test this out by suddenly making a lot of subdomains, actually.
Yes, that is where the wildcard DNS entry is relevant to this situation. Since [anything].straightdope.com will lead to actual indexable content, and search engines won’t know it’s duplicate content for lack of any indication from the website which version they would prefer be indexed, more and more of these will appear in search results over time to infinity. Eventually the search engines might force one version to become the canonical domain and we can only hope the one they choose is ‘boards’.
But as noted many times in many threads, it’s easy to fix this with a single line of HTML code in the head section of the pages. (and in the case of Google I’m pretty sure it can be set using their webmaster tools without even changing the pages, but that won’t have any effect on yahoo, bing, etc.)
I hope fartypants prevails!
Me too. Imagine having the coveted fartypants domain I bet advertising revenue here would quadruple.
I, for one, welcome our new gas-pantalooned overlords.
Which means there is no hard redirect for those pages. I just made the erroneous assumption that the only way to tell Google was to make the server redirect all such pages to their boards.straightdope.com counterpart. A minor change to the markup would be a much easier solution.
And, again, maybe it can be done, since we are also considering implementing some bugfix updates to the software. Adding an extra markup line would be nothing.
Oh, and Senegoid, I needed to do no test because I’ve seen the results of previous ones. I understood you fine.