XML Sitemap Questions For Big Site

[keywordwizzard]

Hey Guys,

I have a few question about XML Sitemaps.

• For a social site that is going to have presonal accounts created, what is the best way to get them indexed?  When it comes to profiles I found out that twitter (https://twitter.com/i/directory/profiles) and facebook (https://www.facebook.com/find-friends?ref=pf) have directory pages, but Google plus has xml index pages (http://www.gstatic.com/s2/sitemaps/profiles-sitemap.xml).

• If we go the XML route, how would we automatically add new profiles to the sitemap? Or is the only option to keep updating your xml profiles using a third party software (sitemapwriter)?

• If a user chooses to not have their profile indexed (by default it will be index-able), how do we go about deindexing that profile?  Is their an automatic way of doing this?

• Lastly, has anyone dappled with google sitemap generator (https://code.google.com/p/googlesitemapgenerator/) if so do you recommend it?

Thank you!

[LesleyPaone]

If it were me and someone were asking me to design a system like that, I would design it in a few parts.

First I would create an application that handled the sitemap minus profiles, just for your tos, sign up pages, terms, and what ever pages like that.

Then I would design a system that handled the actual profiles. It would be pretty complex and resource intensive as the site grew. But the main idea flows like this

Start generation, grab the user record with id 1 in the database, check to see if indexable (move to next if not), see what pages are connected, write to xml file, loop back and start with record #2.

There are a few concessions you have to make, you need to keep up with the number of records in a file before you start another file. You can only have 50k records in one file.

The way I would handle the process in total for a large site would be this, sync the required tables via a weekly or daily cron to another instance (server). Call the php script (because that is what I use) that creates the first sitemap for the normal site wide pages. At the end of that site map, put a location for the user profile sitemap, then at the end of the scrip, execute the user profile site map generating script. At the end of each site map, put the location of the next site map file, because as you grow it might take 2-10000 site map files.

One thing that I would ensure to do is get a list of crawler ip addresses and in your .htaccess have an allow / deny rule. That way you can make the site maps only visible to the search engines.

[spencerhjustice]

I'm not a web developer, so this might may be wrong, but I feel like it might be easier to just add every user to the xml sitemap and then add a noindex robots meta tag ons users pages that don't want to their profiles to be indexed.

[LesleyPaone]

I guess the way I was explaining it was for scalabilty on a large site. You have to think a site like fb or twitter with hundreds of millions of users still has the limitation of only having 50k records in a site map. So if they are running site maps, they have hundreds.

[keywordwizzard]

Thanks for the input guys!

I believe Twitter and Facebook don't run sitemaps for their profiles, what they have is a directory for all their profiles (twitter: https://twitter.com/i/directory/profiles Facebook: https://www.facebook.com/find-friends?ref=pf) and use that to get their profiles crawled, however I feel the best approach is through xml sitemaps and Google plus actually does this with their profiles (http://www.gstatic.com/s2/sitemaps/profiles-sitemap.xml) and quite frankly I would rather follow Google then FB or Twitter...  I'm just now wondering how the hell they upkeep that monster!  Does it create a new sitemap everything one hits 50k?  When do they update their sitemap? daily, weekly, or monthly and how?

One other question I have is if their is any penalties to getting a lot of pages crawled at once? Meaning one day we have 10 pages and the next we have 10,000 pages or 50,000 pages...

Thanks again guys!