StrongMad challenged readers of his journal to update their journal every day for a week. In trying to do so, I realized that I wanted to talk about nothing but object oriented programming. Being that almost none of you but Fractured would know what I was talking about, it was better that that entry was left unsaid, and thus, there was no Tuesday update.
Today, however, I'm going to talk about object oriented programming.
I've got a problem, my interviewing system is decrepid and aging. Last week I turned a statistics script over to a client, she said she had some problems with it, I checked it, and it was embarassing and broken, like a four-year-old had gone through it with an eraser and a keyboard. I put the script together by hand and send it over, then tried to fix the bug, but realized that it's an actual design problem rather than a bug and to fix it would take days. Like an old car with a busted cylinder, it wouldn't be worth it to tear everything apart and put it back together. At best, it would just extend the life of a semi-shitty system another couple of months. An upgrade is necessary.
Now, in computer programming, it's considered bad practice to throw everything away and start from scratch on a new version. The problem with doing that isn't that it takes longer (which it does), but that bugs you had sorted out the first time, get re-introduced:
A good example of this is Netscape 3.1, where Netscape had the most popular web browser back in the mid 90's. Netscape decided when it came time to release 4.0, that 3.x was mostly shit and they could do it better by starting from scratch. In doing so, they reintroduced nasty bugs that had been completely fixed in the 3.x version, along with a host of new bugs. It took them a year to get the browser stable (4.6, 4.8?) so they lost their fan base to Microsoft Internet Explorer and that was the end of Netscape.
I don't care.
My original interviewing system was written in two weekends at Starbucks in 2002. And it's not an exciting system, it has a lot of bugs from various upgrades and patches I've forced through it in the years between then and now. Most of the problems came from the changes being last minute things that I threw together to make a particular study work. And I've learned a lot since 2002, so I'm cool with starting over.
Except it's boring. Dreadfully, mind-numbingly boring. And since the new version is really going to be a wrapper around the perl programming language -- rather than a mini-language in it's own right like the old version, it's not easy to feel happy about design decisions. I've started and restarted this upgrade nearly a dozen times now and it's driving me nuts, like some ugly monkey on my back (who I'm sure is pooing, but I don't know where he's putting it.)
Today, however, I'm going to talk about object oriented programming.
I've got a problem, my interviewing system is decrepid and aging. Last week I turned a statistics script over to a client, she said she had some problems with it, I checked it, and it was embarassing and broken, like a four-year-old had gone through it with an eraser and a keyboard. I put the script together by hand and send it over, then tried to fix the bug, but realized that it's an actual design problem rather than a bug and to fix it would take days. Like an old car with a busted cylinder, it wouldn't be worth it to tear everything apart and put it back together. At best, it would just extend the life of a semi-shitty system another couple of months. An upgrade is necessary.
Now, in computer programming, it's considered bad practice to throw everything away and start from scratch on a new version. The problem with doing that isn't that it takes longer (which it does), but that bugs you had sorted out the first time, get re-introduced:
A good example of this is Netscape 3.1, where Netscape had the most popular web browser back in the mid 90's. Netscape decided when it came time to release 4.0, that 3.x was mostly shit and they could do it better by starting from scratch. In doing so, they reintroduced nasty bugs that had been completely fixed in the 3.x version, along with a host of new bugs. It took them a year to get the browser stable (4.6, 4.8?) so they lost their fan base to Microsoft Internet Explorer and that was the end of Netscape.
I don't care.
My original interviewing system was written in two weekends at Starbucks in 2002. And it's not an exciting system, it has a lot of bugs from various upgrades and patches I've forced through it in the years between then and now. Most of the problems came from the changes being last minute things that I threw together to make a particular study work. And I've learned a lot since 2002, so I'm cool with starting over.
Except it's boring. Dreadfully, mind-numbingly boring. And since the new version is really going to be a wrapper around the perl programming language -- rather than a mini-language in it's own right like the old version, it's not easy to feel happy about design decisions. I've started and restarted this upgrade nearly a dozen times now and it's driving me nuts, like some ugly monkey on my back (who I'm sure is pooing, but I don't know where he's putting it.)
VIEW 6 of 6 COMMENTS
Take care.