link rel="canonical"

[nend]

Here is the URL that the search engine will index since it does not have any extra parameters in it to manipulate the content.

Page 1
http://www.simplemachines.org/community/index.php?topic=481995.0

Page2
http://www.simplemachines.org/community/index.php?topic=481995.20

[alphacaveman]

That's NOT how Google intended the "no index" to be used. You're just supposed to add the rel="canonical" information and Google can figure out the rest. Putting a no index on a page expecting that to get Google to index the correct page as the index is just very bad practice when it comes to SEO. Can't believe it was put in there.

Actually, Google never specified how it should or should not be used (and I think you'll find it predates Google anyway). It's used to indicate that the current page should NOT be indexed, since canonical does not necessarily indicate that the other pages should or should not be indexed. You'll note even from the page you link that Google itself says it won't necessarily rely on what canonical says. Other search engines still do not listen to it anyway, and noindex is the safest way to make it work as expected (since there are enough links through other navigation that will be indexed)

I find it hilarious to note that the only people who complain about how SMF's 'SEO' does or does not work are the people with the tiny forums who think they know best, the guys with the bigger forums have absolutely no troubles with it working the way it is supposed to...

Yeah...LOL, cause I have a tiny forum. Right....

I've never had any trouble getting SMF to do what I want it to do SEOwise. I just recently upgraded to 2.0.2 and am noticing this.

Telling Google not to index the page is not the way to go. Using rel="canonical" allows you to get link PR, juice whatever-you-want-to-call it, from links that are going to one of the other pages. For example, I have tons of links to the foldername/  and to the foldername/index.php. Since I can't do a 301 redirect because of the way it's structured, the way I can tell Google that both of those pages are the same is to use the rel="canonical" attribute. Google is smart and knows how to take it from there but telling them not to index a page just means that none of the links that go into the page will be counted.

[Arantor]

And that's actually exactly what you want. You do not want index.php?topic=1234.msg123 being indexed at all. Indexed is not the same as followed, it will still follow everything.

All the links in the page navigation (Pages: 1 2 3 etc) will be indexed as expected because they're not topic=1234.msg123 links.

Seriously, leave it alone. It works very well and doesn't generate any problems as it stands.


And you're saying you know better than multi-million post forums about what works for SEO, eh?

[nend]

It should be ok the way it is, noindex is for compatibility with some other search engines. If you would like drop support for some other search engines then go head and remove it.

However with noindex links will still be followed, it just tells it not to include it in the search listings. You can also do noindex follow, but it is still the same as noindex.

[alphacaveman]

And you're saying you know better than multi-million post forums about what works for SEO, eh?

Well, my forum has a million posts, but I don't use that as a measuring device for someone's SEO knowledge. I'm just saying that I know SEO and any person who does should pause over seeing noindex, nofollow, etc. I'd rather ask because I've found that software developers don't always have SEO as a priority. I'll leave it in though, since you say it's just to stop Google from indexing the anchors. Though I wish that structurally SMF would've done something else so that the noindex or other maneuvers would not be necessary.

[Arantor]

*shrug* So you don't like how SMF does it. What would you suggest they do, exactly?

The only way around this is to do what something like XenForo does and introduce a redirect between pages under certain circumstances. The main reason for having the topic=x.msgy format is so that you can go to a thread and to a specific message in that thread without knowing what page it is on, and that's done without having to redirect (at the expense of using canonical and noindex the way it is). XenForo et al does something similar, except instead of doing the noindex/canonical like SMF, they have a URL of /goto/post?id=x#post-x which then performs a redirect.

Now, doing a redirect like that, every time you have something like a post quoting another (which is one of the reasons SMF does it too) is actually at least as bad as using noindex/canonical like that, except you also have to factor in another server round trip, and all the performance implications associated with it (and remember: speed is also a factor in SEO)

[alphacaveman]

Thanks Arantor. Take care.

[alphacaveman]

Arantor,

What about the session ID's? Does the noindex tell Google NOT to index them?

[Arantor]

No, it doesn't.

There is, in fact, a bug that the dev team has historically refused to understand with respect to the canonical link.

The deal is that, generally, the session id won't appear in the page if cookies were provided. If cookies were not available (like some search engine cases), the session id will be injected into the page in every single link - including the canonical URL, and this is done after the templates.

There is actually a mod for it to fix the situation so canonical URLs will be dealt with, though it's not as much of an issue as it could be, knowing that Google is quite happy to ignore the PHPSESSID (doubly so if you tell it to do so in Google Webmaster Tools)

[alphacaveman]

Thanks. I told Google in webmaster tools. Kind of scary how they say, "Use this feature only if you're sure how parameters work. Incorrectly excluding URLs could result in many pages disappearing from search."

[Arantor]

Well, it's not scary at all.

Consider the effect if you put in 'topic' as a parameter to exclude. It will exclude a very large number of pages then.