Archive for the 'openid' Category

Next Entries »

This blog is OpenID enabled!

Monday, January 15th, 2007

OpenID logo

One thing led to another. After reading my previous entry on “mod_python OpenID Access Control”, Nate Olson contacted me to ask if I tried Alan Castonguay’s Wordpress OpenID plug-in. To which I replied, “no, but I will try it out!”

After upgrading my wordpress to the latest and greatest, and then installing the plug-in, this blog is now OpenID enabled!

You’d need to first get yourself an OpenID URL (better yet, an i-name). There are many OpenID providers out there. Once you have that, you can then comment on this blog (and many other sites) by logging in with your OpenID. If you stay logged in at your OpenID provider, there is no need to type in your password in pesty registration forms as you move across various OpenID enabled sites (Zooomr is one such early adopter.)

Posted in identity, openid | 5 Comments »

mod_python OpenID Access Control

Monday, January 15th, 2007

Since XRI is pretty much in bed with OpenID and NeuStar is an XRI shop, I get to play around with it quite a bit.

Here’s a little success report about using JanRain’s mpopenid module to protect certain restricted resources served by Apache. It works somewhat like your basic HTTP authentication, but instead of returning a 401 Authorization Required response causing the browser to prompt for username and password, it redirects you to the OpenID login page. Pretty neat eh?

So, here's how we go about doing it.

My ingredients:

Following the recipes in this README file, I was able to set it up with one caveat: Python ElementTree module is a dependency of the Python Yadis library and you need to first install that.

Did I mention that you could login using an I-name?

UPDATE:

The following patch for mpopenid.py is needed to make i-name login work:

--- mpopenid-1.2.0-pre5/mpopenid.py     Tue Nov 21 20:24:53 2006
+++ mpopenid-1.2.0-pre5-wil/mpopenid.py Sun Jan 14 16:12:16 2007
@@ -157,10 +157,17 @@
         url = s.strip()
         if not url:
             continue
-        parsed = urlparse.urlparse(url)
-        if not (parsed[0] and parsed[1]):
-            url = 'http://' + url
-        urls.append(urinorm(url))
+
+        if (url[0:6].lower() == "xri://"):
+            url = url[6:] # strip "xri://"
+
+        if (url[0] not in "=@!$+"): # doesn't look like an XRI
+            parsed = urlparse.urlparse(url)
+            if not (parsed[0] and parsed[1]):
+                url = 'http://' + url
+            url = urinorm(url)
+
+        urls.append(url)

     return urls

Posted in identity, iname, neustar, openid, xri | 11 Comments »

sxore commenting

Saturday, August 12th, 2006

sxore logo

sxore is an identity and reputation system for blog authors, readers and commenters developed by sxip. As with any respectable web 2.0 services, it is in public (perpetual?) beta. I’ve heard of them before via TechCrunch but have not tried it. With so many new services being reviewed daily, where would you find time for that unless that’s your full-time job?

Our work on XRI somehow introduced me to Keith Grennan at Sxip who mentioned that they’ve successfully integrated i-name authentication in their sxore service – so I decided it to try it.

Everything works as advertised. The sign-up process was a breeze; just entered my i-name =wil, it brought me to my i-broker which authenticated me and kicked me back to sxore.com where I filled in the URL, an alias and email address. There I am, instantly logged in, without having to stir brain juices to construct yet another password! Such is the power of XRI and OpenID.

Now I’m logged in, it tells me to download and install the sxore WordPress plug-in. Ok, wasn’t expecting that (in fact, I had no idea what the service did before I started!) I followed the instructions to install the plug-in and activated it with the code that was given to me and “claimed my blog”. There were a few manual steps involved there, but you’ve got to appreciate how smooth and pain-free it was. A lot of things actually happened behind the scenes and I’m sure I can’t do it justice. For example, sxore detected my blog publishing platform and provided instructions tailored just for me.

And that’s it! I can assure you that all that took less than 10 minutes.

When you click on the “Leave a comment” link below, you’ll be brought to sxore where you can leave your comment. You can also provide an email address to track replies to your comment, not only on this blog but others that use the sxore commenting system.

UPDATE: Was told that the sxore plug-in will eventually put the comment handling responsibility back into WordPress to take advantage of Akismet for spam filtering. The sxore plug-in will concentrate on the identity and trust instead. Having just gone through the pain of migrating posts and comments from the old blog, I have decided to turn off the sxore plug-in for a while, simply because I want my comments to stay in my database. The plug-in is great though, especially the feature for a commenter to track replies and the reputation system may really help the comment spam problem too but let’s wait for a new version. See the sxore plug-in roadmap for more information.

Posted in blogging, identity, openid, xri | No Comments »

Next Entries »