Pankaj Kumar's Weblog

Setting up this weblog took less than I had anticipated. I started today at around 5pm and had it up and running at my website by 7pm. Given that I am quite a novice with cgi and perl, the credit goes to MovableTypes for their ease of installation.

Had been thinking of having a weblog for quite sometime but kept postponing it. It was Ovidiu's weblog that sparked my interest, funnelled by various media reports and the observation that anyone who had an opinion had a weblog.

Now that I have got this up and running, I will hopefully be able to keep it alive !

In my quest for a good enough CMS/CP system, I resisted the urge to abandon Java all these days, but my patience is now running out. Found JLCP of JavaLobby too new. CoFax is interesting but is not designed for what I need. OpenCMS appears to be a closed system, simulating its own filesystem. Jetspeed is promising but would still require a good amount of work to get what I want. After spending better part of the long weekend trying to figure out how to make Apache Forrest and Apache Cocoon work for me, and finding them hopelessly complex (and I am not new to Apache Cocoon. In fact, it was I who wrote the Understanding Cocoon two years ago when the Cocoon documentation was really scarce ), I was quite disenchanted with the current state of Java CMS/CP systems (or their building blocks).

In a moment of vulnerability, Gyan (Veena's brother), who himself has been dabbling with PHP, convinced me to try out PHP-Nuke. In the beginning, I was bit scared of learning yet another programming language (i.e., PHP). However, this fear turned out to be baseless. Downloading and installing PHP-Nuke on my Linux system was a breeze. Browsed through some of the source files and an excellent guide. PHP-Nuke's completeness for what I was looking for and its relative simplicity was amazing. I was impressed.

However, running PHP-Nuke successfully involved some work. It needed MySQL and I discovered that my standard Linux installation didn't have it installed. Luckily, it was there on the RedHat Linux 8 CD#3 and installation using KDE Windows Manager (System Settings --> Packages) was quite smooth. But even after installing MySQL, I was not able to access index.php or admin.php of PHP-Nuke. I was getting a blank browser and the View --> Source indicated a html document with empty body. With some experimentation, I tracked down the problem to the failure of PHP to access MySQL. This was resolved by installing php-mysql from CD#3 and restarting Apache. I had to resort to line mode commands for this (for, the KDE Package Manager would just not show me this package):

#mount /mnt/cdrom
#rpm -i /mnt/cdrom/RedHat/RPMS/php-mysql*.rpm

Well, PHP-Nuke now works great on my box and I think, is quite simple to customize and extend. Perhaps this is what I will end up using.

Since the initial press release (which, frankly, seemed like an April Fool joke) announcing Google's Gmail, a lot has been said about it: the usual privacy concerns, the massive and not easily replicable distributed computing infrastructure that Google might have built, Forbes review predicting demise of hotmail and others, and so on.

However, all these analyses are missing an important angle: With Google's tremendous storage capacity and computing power, it would be able to do the same with Email that it has been able to do with Web pages. I am referring to its pagerank based analysis and categorization of the web pages, which simply exploits the fact that hyper-linked web pages have an inherent structure and this structure conveys something important about each page.

Now think about individual email messages -- each email message also has an inherent structure: who sent it? to whom? when was it sent? was it a response to another message? how did the communication evolve? Analysis of a very large number of email messages can yield some very interesting facts about individuals, their relationships with others, opinons, tastes etc. Combine these with an engine that builds personalization profiles based on other factors (such as search phrases and click-behavior on returned results) and what you have is a very powerful and adaptive personalization scheme. This can certainly be used for a number of consumer services or enhanced options for existing services, besides very targetted and focussed ads.

The pattern of email communication could have other uses -- think of social networks being promoted by likes of LinkedIn and Google's own orkut. The combination could be much more powerful, intelligent and perhaps significantly more useful to the consumers of these services!

All this will certainly redefine the very meaning of Email. No wonder Google invented a new word to describe this service: Gmail.

Like most of us, I have been a IE only user for a very long time. But this is changing. It started with a Firefox download few months ago out of curiosity -- there was so much buzz around it that I just had to check it out. It was good but not compelling enough, at least for me, to force a switch from IE. I realize that a complete switch is impossible even now, but what I noticed that I have been firing up Firefox more and more often because there are things that I cannot do with IE.

1. Last week I was dabbling with some DHTML, CSS and JavaScript programming and was particularly annoyed with IE's lame error reporting. Tried Firefox and just fell in love with its JavaScript Console. It not only gave a list of meaningful error messages but also allowed the source file to be opened with a simple mouse click!


2. While trying to figure out the page layout created by an AJAX app., I felt the need for a tool that could show me the underlying DOM tree. And I found mention of Firefox DOM Inspector on the Web. Though I had to do a re-install of Firefox with custom install option, it was worth it. Don't know how I lived without it for so long!


3. I tried to submit a Book Review yesterday to Slashdot. failed with IE, but succeeded with Firefox. Looks like their new CSS doesn't play very nice with IE. In fact, today also I saw overlapped and at places, missing text in certain Slashdot stories. There was no problem with Firefox. Good to see the same trick that MS used to promote IE be used against it (though, perhaps unintentionally).

At first sight the last experience appears to be unrelated to the previous two. Though I think there is a deep link -- because Firefox has so much better support for Web developers, developers are more likely to use FireFox for developing rich interactive sites and come back to IE only later. This is what seems to be happening with Slashdot and is likely to happen many more sites as they move towards DHTML, CSS and AJAX.

My website hosting company lets me access the raw webserver log files and analog reports. The logfiles have all the information but are useless without furhter processing (well, they are useful during development debugging of webapps, but that is a different matter). The pre-configured analog reports are better but don't have much visual appeal. They also don't allow me to do further drill-down (For example, it lists top 30 search terms but won't let me see the complete list).

So, I installed and configured open source log analyzer awstats. The generated reports are much nicer to look at and allow me to drill-down for more information. A full report includes information on:

1. No. of unique visitors, visits, pages accessed, hits and data transferred in a given month or day. Also, hits and data transferred in a given hour.


2. Hits and data transferred by specific robots and spiders, in a given month.


3. Distribution of visit duration, in a given month.


4. Hits and data transferred by file type, in a given month.


5. URLs sorted by view count, in a given month.


6. Operating Systems and Browsers used to access the website, in a given month.


7. Referring search engines and pages, sorted by view count, in a given month.


8. Search keywords and phrases, sorted by usage count, in a given month.

Note that most of the information is organized by month and it is not easy (or even possible) to specify a different duration. What if I want to look at referrers for a particular link on a particular day or week? What if I want to see the distribution of view count for a particular page for last one year? What if I want to know the dates on which Googlebot (or MSN or whatever) indexed my site over last one year? Possibilities are endless.

Answers to these, and many more similar queries, can be easily obtained by doing a more thorough analysis of the logfiles. Perhaps I could read the log entries, import them in MySQL and issue SQL queries against the relevant tables. I could even write a web interface to it using something like Ruby On Rails. Should make for a fun hobby project.

What I found most amazing about the eBay auction of Web 2.0 startup kiko.com was the runup to the sale price of $258,100.00 during final hour of the auction, as seen in the bid history:

What do you see in this bid history? There were at least five serious buyers who pushed up the price by bidding and counter-bidding. In fact, someone with eBay account name of wswire was beaten by powerjoe1998 in the final minute.

May be this is not so amazing in eBay auctions, especially for unusual items!

Not very long ago, during the heydays of dotcom boom, people talked about everything from toasters to refrigerators being connected to the Internet. Ofcourse, in absence of real, value driven use cases, most such talks died along with the dot com bust.

But may be, there are valid use cases for ubiquitous internet connectivity. For example, I would love to have wrist-watch that not only tells time at any place but also lets me "see" the satellite view of the place, as demonstrated by this Google Desktop Gadget.

Though I have a feeling that we might see this kind of capability in a Google enabled video iPod as a feature much sooner than in a wrist-watch.

Ran into an interesting situation with a Drupal powered sub-site: could log into the site via http://pankaj-k.net/ but not through http://www.pankaj-k.net/ (bot domain names, pankaj-k.net and www.pankaj-k.net, point to the same site). Did a bit of digging at Drupal.org and found that I am not alone.

Though, I did observe a few things that might help in identifying the real issue.

I have a Drupal installation on a host which can be reached via browser by either http://domain.com or http://www.domain.com. I can login into the site through http://domain.com without any problem, but not through http://www.domain.com (the login block comes up on page refresh, without any notification, as reported by others). Though, there are a few things that puzzled me:

1. If I login through http://domain.com, close the browser without logging out, open the browser again and point it to http://www.domain.com then I continue to be logged-in and can access the site without any problem.
2. If I log out from http://domain.com, delete all the cookies and then visit http://www.domain.com and login then the login SUCCEEDS. Logging out and logging in to http://domain.com also SUCCEEDS (ie; login to http://domain.com always succeeds but login to http://www.domain.com succeeds only when there are no cookies for http://domain.com)
3. The above observations are with IE6. With FireFox, login through either of the URLS, http://domain.com or http://www.domain.com succeeds. I even tried a number of sequences of login and logouts in different permutations and all work fine.

Makes me feel that there is something weird going on with cookie handling!

Have posted these observations to Drupal issue log.