Steven Yong's Weblog

After IE7 has been officially released, we can construct XMLHttpRequest natively,
i.e. var xmlHttp = new XMLHttpRequest()
for all the major web browsers including Mozilla, IE and Safari. Isn't that a good news for those who build AJAX web application?

Just got a sms saying that I got a picture message which I have to login to the website and view it. Here is the error page I get after trying to login in:


How the heck am I supposed to know what to do next? Does the error calm you down by telling you which layer is causing the error?

The first principle of error page design: Tell something meaningful and guide the user to the possible workarounds, or at least ask the user to try again later.

I bought a pair of new leather shoes yesterday. Guess what? it isn't for Chinese new year, it was merely because my previous pair have been (or has been? correct me please) stolen (it happened to my neighbor too). I'm suspecting it's done by the foreign worker because the place I stay is literally an industrial area which has huge number of foreign workers who came from Bangladesh, Myanmar and Indonesia. What is my point? I don't know, most of the times I don't have one. But the lesson of this story: Never leave your shoes outside.

Now what I make a good use of my Apache http server because I managed to get both bug tracking system and version control system running.
Unlike setting up Bugzilla, setting up and configuring Subversion with Apache http server is pretty straight forward which I think everyone can do by following the manual.
If you're using Apache httpd earlier than version 2.0.x, more works will have to be done including manually copy the Subversion WebDAV Shared Library (mod_dav_svn.so, which comes with the Subversion package) to the Apache installation directory. The reason we have to use Subversion with Apache httpd is making your repository available to others over a network. Of course Apache httpd isn't the only option, you can also use IIS or even the custom server provided by Subversion, svnserve (which you can access your repository via svn:// or svn+ssh://). Since I've configured my local Apache for Bugzilla, I might as well use it for Subversion also.

Ok, the solutions of the problems have been somewhat figured out and I find them quite feasible and practical.
The only issue now is Bugzilla doesn't really solve the problem well and I got to know that from a friend there is a PHP based system which I'd like to evaluate it thoroughly as well. Hopefully the solutions will get implemented right after chinese new year.

PS: If you don't happen to know it yet, Subversion is designed to overcome some CVS limitations while providing most of the CVS features.

I got an interesting comment from my previous post which the commentor asked: "You really anticipate so much bug that you will need a bug tracking system?"
Of course he didn't ask this question for himself as he is a great developer who always strongly supports any good software development practices where bug tracking system is absolutely not excluded. In fact I believe he asked that to get attention from those who have the authority to implement bug tracking system but somehow have not yet done so.
I really want to borrow a good quote from Joel:

One of the biggest incorrect facts that programmers consistently seem to believe is that they can remember all their bugs or keep them on post-it notes

I find it true wholeheartedly because I can't remember more than two or three bugs at a time, and the next morning, or in the rush of shipping, they are forgotten.
A very simple question to those who think that bug tracking system is not always needed: "Would you like your manager come to chase you for fixes every morning?"