When the electric guitar was first invented, it was played just like an accoustic guitar but with amplification. Later, artists like Jimmy Hendrix came along and played it like it was a fundamentally different instrument. I think that a similar cycle is likely going on with the web, as the original article says. Like the electric guitar, the web has ways of "playing it" that are fundamentally different from the non-web counterparts. The best innovations have come and will come not from porting non-web faculties, but inventing new ones that could not exist without this medium.
Creating a POS system is easy. I've worked on many. Most POS's can be created with no planning and require little to no programming skill whatsoever. Or did you mean a Point Of Sale system?
There's also a common scenario in which "fixing" a broken module breaks the code. This can happen because others have coded to depend on the incorrect behavior or it was never determined to be incorrect behavior and now users are familiar with the program working that way. Either way, your fix will be perceived to be the problem.
So, I'd say rewrite/refactor only as an approved task that provides sufficient time to regression test and catch any unforeseen consequences of the change.
Slashdot Mirror
When the electric guitar was first invented, it was played just like an accoustic guitar but with amplification. Later, artists like Jimmy Hendrix came along and played it like it was a fundamentally different instrument. I think that a similar cycle is likely going on with the web, as the original article says. Like the electric guitar, the web has ways of "playing it" that are fundamentally different from the non-web counterparts. The best innovations have come and will come not from porting non-web faculties, but inventing new ones that could not exist without this medium.
Creating a POS system is easy. I've worked on many. Most POS's can be created with no planning and require little to no programming skill whatsoever. Or did you mean a Point Of Sale system?
There's also a common scenario in which "fixing" a broken module breaks the code. This can happen because others have coded to depend on the incorrect behavior or it was never determined to be incorrect behavior and now users are familiar with the program working that way. Either way, your fix will be perceived to be the problem.
So, I'd say rewrite/refactor only as an approved task that provides sufficient time to regression test and catch any unforeseen consequences of the change.