Forum OpenACS Q&A: March Madness on a AOLserver architecture

For those of you interested in March Madness or just curious as to check out a site built using aolserver architecture ... check out Users can build their own blogs, create their own groups, and build contests/pools of multiple types. The majority of the site never hits the db and uses nsv caches extensivly. Check out the process times and technical details at the bottom of every page.

I've incorporated Amazon web services, Blogger API, and XML-RPC into the site. Be sure to check out the site technical details in the company info section and for those brave enough to actually enter March Madness or any other pools ... good luck :)

Let me know if your interested in how I did any of the functionality.
Posted by Eric Wolfram on
It's going to be a difficult one to pick this year. Last year I picked Maryland to go all the way and I was lucky. It's obvious that Maryland ain't gonna save me this year. 😊 They might make it to the sweet 16. Arizona and Duke will go far, on the other hand, but that's what everyone is going to say...

I'm also interesting in allowing users to make new groups on my site. How was that accomplished on your site?

Posted by carl garland on
Groups are very similar to portlets for OpenACS so if you are using OpenACS look into that. For my approach creating new groups consisted mainly of:
  • Authentication
  • Group Profiles for Users
  • Creating Columns for group ids to calendar/bboard etc tables
  • Creating Group Activity Logs
I'm sure more stuff as well, if you have any more specific question I could send you the corresponding code.
Posted by Wes Reing on
I would be interested in how you connected to Amazon’s Web services.  Did you use a TCL RCP package?
Posted by carl garland on
You can use Amazon Web Services with the Rest method of just using HTTP Post requests and you will receive a XML response. Then use ns_xml to parse the xml with the ns_xml page and your pretty much there. I cache the results in the filesystem in both XML and my HTML Display of item versions. You need to make sure to update the cache as needed though because Amazon doesnt want some data to become outdated.

If you or anyone else wants I can send the code that deals with the Amazon stuff, you would have to tweak for your own use though. I offload some of the requests/processing to a separate server to reduce load on the main user/frontend servers. Don't want to tie up your connection threads with lots of users potentially waiting for their requests to Amazon to be completed.