Tried to log into my Comcast account to check my account this weekend and discovered that I had forgotten my password. No problem...click the forgot password link and go through the email password reset drill, right?

Wrong! Comcast's site suggests re-creating my user profile instead. And when I do try to re-create the profile, it complains that the profile already exists. Well, DUH!

I cannot believe this stupidity. To make matters worse, I called the Customer Service number and described the problem. The first thing they asked me is why I didn't try to reset my password. Sheesh!

In late Sept. 2007, I ditched Outlook for good and switched to Zimbra. My experience with Zimbra has been great -- exceptionally stable, wonderful user experience and every feature I have ever wanted in a webmail client. That was until v5.0x was released. It had even cooler features and I decided to take the plunge.

Bad move...after spending almost a whole day in aggregate, my smooth running Zimbra server on a Linode VPS is toast. I mean the data is all there, but I cannot get the server to work right. I have spent lots of time on the Zimbra wiki and forums, scoured the documentation and tried lots of fixes -- bottom line, the new webserver Jetty that replaces Tomcat just does not work. It basically responds to requests, but does nothing...for infinity. I have done upgrade installs, clean installs, clean installs on a clean server...no dice...it just won't work.

Today, I finally admitted defeat...Zimbra has lost some of its lustre (but only on the server side). For the first time ever, I have my mail hosted on a server that I do not control. When I first started my Zimbra experience, I found 01.com to be the best Zimbra host out there, so today I setup all the family email accounts on 01.com.

So far the experience has been pretty decent. Everything works as advertised (although it's a bit sluggish...but maybe that's because I'm coming from a VPS). The only area I was underwhelmed by was support...it took several hours to get a couple of domain aliases added to my account...a pretty simple operation. But so far, so good.

Now, all I have to do is figure out how to get the message store from my server into my 01 Zimbra account.

[Update]

The VPS on Linode has a cool feature that allows you to purchase more memory for your server on-demand. On a whim, I did this and presto...my Zimbra server was back to normal operational status. I was running it on 1Gb, while the minimum requirements for Zimbra v5.0 appear to be 2Gb. After doing some research I found that this is only for the default settings. By making config. changes for how much the Java VM and MySQL consume, you can run it on 1.5Gb.

Now, I have all my email from my Linode transferred to my 01.com Zimbra mailbox (using IMAP). Even though I like the thought of having complete control on my mail, this experience has convinced me that for something as critical as email, delegating backups, upgrades and compatibility to a provider is a better option. I have been using 01.com's service for over a month now and am a happy customer. I even signed-up for the Blackberry OTA sync of emails, calendar, addressbook and tasks that they provide. It works great.

Glad to be enjoying Zimbra once again. The lustre is back.

Almost exactly a year ago, Sramana Mitra offered a simple equation for Web 3.0: 4C + P + VS. I think she is right on. It's a neat and concise way to express something quite complex. Of course, since it is only neat and concise if you know what the variables in the equation mean, here's the long form of the equation:

Context + Commerce + Community + Content + Personalization + Vertical Search

A few days ago, Sramana requested help in naming this equation. Since I love playing with words, I was immediately drawn to this challenge. Looking over some of the suggestions posted in the comments for inspiration, I came up blank. Sramana's suggestion of "Personal Concierge" is also not a good fit. What's needed is a simple word or phrase that represents the equation and can be used in multiple contexts with ease. I came up with just such a word --

HYPERWEX

Here's why I think this is the perfect word to describe the Web 3.0 equation:

1. Google results are zero (at least at the time of this writing), so it's a fresh new word.
   
2. Hyper = Beyond; the concepts of Web 3.0 are beyond anything we have ever experienced, online or offline
   
3. Hy = Highly and Per = Personalized; it's all about a high degree of personalization
   
4. We = Web; it's all about the world wide web
   
5. Ex = Experience; ultimately, context/commerce/community/content all come down to user experience
   
6. Per = Personal and We = Web; Web 3.0 is about the web coming to you...the personal web
   
7. We; Web 3.0 is also about social networking -- "we" conveys that quite well
   
8. Hype; Web 3.0 is already being hyped as the coming utopia

As you can see, Hyperwex is a great way to describe Web 3.0. Not only that, it is usable in so many different ways:

• Hyperwexable applications will dominate the web in the future.
• Once a person is hyperwexed, there is no going back to the sheer time waste that was Web 2.0.
• Hyperwexing saves me time and money and makes me a better person.
• Google today acquired two hyperwex startups to kick-start its own efforts in this space.
• That site has a truly hyperwexory user experience. I love it.

What do you say Sramana?

 

When you go to the Freebase Data page, there is a list of categories. Drilling down into these, you can see the top-level domains and their types. Looking at all the types in a domain is easy enough, but I could not figure out a way to get the categories and top-level domains. The documentation was no help.

I finally found the answer by looking at the mjt source code at this site:
http://mjtemplate.org/examples/schemas/index.html

Here's the query to retrieve the categories:

{
  "query" : [
    {
      "name" : null,
      "type" : "/freebase/domain_category"
    }
  ]
}

And here's the query to retrieve the top-level domains for a category:
(Caution: This query will take a long time to run as it returns the top-level domains and their properties

{
  "query" : [
    {
      "domains" : [
        {
          "id" : null,
          "name" : null,
          "type" : "/type/domain",
          "types" : [
            {
              "/freebase/type_profile/instance_count" : null,
              "/type/reflect/any_reverse" : {
                "limit" : 1,
                "link" : {
                  "master_property" : "/freebase/type_hints/hinted_type"
                },
                "mediator" : null,
                "optional" : true,
                "type" : "/freebase/type_hints"
              },
              "id" : null,
              "name" : null,
              "optional" : true,
              "properties" : [],
              "sort" : "name"
            }
          ]
        }
      ],
      "name" : null,
      "type" : "/freebase/domain_category"
    }
  ]
}

I wish some of the types such as "/freebase/domain_category" or "/freebase/type_hints" were documented somewhere.

Can you imagine creating biological machines just as simply as writing code?

I am guessing your answer is no. Until today, I couldn't either. Then I watched this amazing video of Drew Endy at the Chaos Communication Congress. It is totally cool stuff. Here's a synopsis of the talk: