What is a sitemap worth if it doesn´t contain the complete content of your site ?

Will it damage your site, your PR your ranking ?

Shouldn´t a sitemap always contain everything and all of your site links ?

Opinions ?

Well, I think the discussion should really be about two different kinds of sitemaps (maybe more?), or really "uses" for them: 1) Google/SE sitemaps, and 2) end-user sitemaps.

Unfortunately, without having done any research on the Google site about their own sitemap capability (where you submit your sitemap URL), I do not know what their practical limit is on the number of links. I also get concerned about the enormous size of the XML that could be produced and have to be processed. It would seem that all links would be desirable IMO.

For end-user sitemaps or even those that are intended to help other SE's to crawl, IMO, they should include paging. I still think more than 100 links on a page is not looked upon favorably by the SE's. What human being is really going to "process" that many links? Therefore, since the SE's are trying to assess the value of a page based upon how a human being would process the page, it just makes sense that too many links would not be good.

Always willing to give an opinion... lol.

The problem is e.g. my site contains a lot of html sites wich are not included in Nuke and several older modules wich I ´m not able to use modrewrite for and to include these in this sitemap isn´t possible.
Therefore a sitemap without a full overview isn´t perfect for me I think.

Also I don´t know how Google will handle this cause my site is several years online and the most urls are already known but the NukeSeo sitemap my sitemap.html I submitted doesn´t contain these links.
I´m not a sitemap freak and therefore I had in the past very good results just with a good structure but the site grows very quickly and I can´t do this manually like I did in the past.
With 100 links per page I would have between 38 - 40 sitemaps sites.
I don´t use xml.

I have the same "problem". I still use the sitemap tho... I'm on the lookout to see if the other pages disappear in google. I doubt it but I'm guessing that those pages might lose in rank.

Sitemaps should contain as many links as possible - it's a map of the site. What's the point of looking at a partial map? Humans can use it with Ctrl-F to search for specific content (this works better than the default Nuke search function, but not better than mSearch... ). Search engines are looking for all the links, too, and aren't interested in having to follow multiple links to the content.

That said, some humans may prefer the flexibility to switch back and forth between all content and specific types of content. My thought was to use a cool jQuery script that would let a human hide the content they do not wish to see by content type or subtype (category, topic, forum, etc.) in a tree / hierarchy format.

Thoughts?
_________________

Sounds interesting. I believe such a tree menu is the solution.

kguske wrote:

That said, some humans may prefer the flexibility to switch back and forth between all content and specific types of content. My thought was to use a cool jQuery script that would let a human hide the content they do not wish to see by content type or subtype (category, topic, forum, etc.) in a tree / hierarchy format.

My concern would be how the SE's would look at this. They have gotten pretty sophisticated at recognizing various attempts to "game them" into crawling links that the end-users may not even be seeing through the use of CSS and/or javascript.

Now, I am not saying that you are "gaming them". What I am really saying is that since so many games have been played over the years, the SE's can also render false positives... I mean, they can consider you trying to "game" even when you are not. Problem is, no-one knows their complete algorithms to know for sure what you propose would be looked at unfavorably.

My "gut" and reading still tells me that having a very large HTML page with tons of links is going to be a negative from an SE standpoint. Everything that I have read is that the SEs are trying to "think like a human" and there is no way that that kind of page would be considered human friendly. That is why I suggested paging.

Unfortunately, the content within *nuke is such that it is quite "flat", and therefore, doesn't lend itself very well to any kind of deep categorization breakout (maybe News and Forums are somewhat better).

oh well, I'm out of thoughts... lol.

Very interesting topic but I just want to add something here specifically for Susann before going off track as I tend to do
Susann, you have multiple sites hosted within the same domain?
No real problem there but if you were not aware, Googles XML sitemap reader will throw an error if you include a link not under the webroot - by that I mean if you are using sub-domains etc then the sitemap will fail at Google even though it might be acceptable for a human reader.

The purpose of a sitemap is to make crawling deeper level pages easier for the SE's so they do not have to dig so hard.

I agree with Montego's thoughts to a degree BUT I also think it has become somewhat acceptable to have a fairly hefty page of links (sitemap) provided all those links are internal. Google itself will allow a fairly hefty file size before zipping the file becomes another option (which Google will unzip and read on the fly) and also allows for you to link sitemaps together so you can have one master sitemap which in essence is the 'map' for related 'other' sitemaps (provided they are all internal to the same domain).

Yahoo will read a hefty file size too but prefers a HTML sitemap and then there is textual sitemaps and ROR sitemaps......

Google also has a seperate special sitemap for 'products' which ight work well for e-commerce and/or download sites but I gave up on that one after spending a day trying to figure that one out.

Well, I had to crawl my site and created a sitemap.xml to avoid different issues.
The difference between the nukeseo sitemap and the xml sitemap is beneed other factors the number of urls.
The xml sitemap contains 4 x urls.
To avoid a slow death of my site cause of duplicate content and other nonsens I didn´t list in the xml. sitemap:

-----------------------------

-ftopic-reply-
-new-
-reply-
account
account-pass_lost.html
addlink
article-friend-
days0
download-seldate-
faq
feedback
forum-userprofile
fsearch
group



login
mark
modifylink
next
orderasc
popular
previous
print
Private_Messages
profile
quote
random
redirect
register
site_id
SPGo&site_id
stats
submit
unanswered
userinfo
Your_Account


I hope thats correct.


I only need a sitemap for Google thats enough.
People expect sometimes too much from a sitemap I believe.

I agree, there are far too many links in nuke which go to identical or very similar content - News is a good example where you have one url for the topic, another for 'read more', another for news comments which goes to the same data through a different url and of course the 'vote for article' link which uses yet another url to go to simiar content.
Then there are tons of links which I believe do not belong in a sitemap like memberlist, profile, topic reply, topic quote..... the list goes on.
I think if nukeSEO (sitemap) had some regex facility to ignore some links from being included in the sitemap it would make things much easier but given that we are working with the very imperfect (from an SEO standpoint) nuke I think what we have at the moment is certainly far better than having nothing.
I'm really excited at what future versions will bring to the table especially if we can finally combine meaningful title/anchor text to links in the map but I fear the more we consider SEO the more likely solutions will have to become 'core' rather than add-on code.