More Mozilla RC1 Reviews
Thursday April 25th, 2002
Anonman writes: "c|net's article Don't Miss the Thrilla from Mozilla is quite positive. Their only complaint seemed to be a problem with Mozilla reading a Netscape profile. The reviewers liked Mozilla's quicklaunch, tabbed browsing and price, going so far as to suggest that 1.0 'may actually best its two most powerful competitors.'
#20 Re: Re: Re: Interesting article...
Saturday April 27th, 2002 10:56 AM
You are replying to this message
"The article was not about Netscape's market share. The article was about refactoring vs. rewriting. The plummeting market share over a three year period DURING THE REWRITE was mentioned simply to show that rewriting an application rather than doing stepwise refactoring leaves a company vulnerable since it's impossible to ship new features during the rewrite."
Netscape decided to release the code in January of 1998, with the code becoming public on March 31.1998. It was roughly another 6 months before the decession to do a rewrite was made in the early fall of 1998, around two years not three years before Netscape 6.0 was released.
IF, they had decided to refactor rather than rewrite at least another 6 months would have been required to get the code into reasonable shape and remove bugs introduced during the refactoring process. THEN large portions of it would have had to be rewritten anyway because the code that was released in 1998 wasn't a complete software package. Large parts of Netscape 4.0 were third party code that was licensed by Netscape, not owned by Netscape. These sections had to be rewritten anyway to have a fully open product.
IF the resources existed to maintain the old code AND work on the new code, we could have gotten an inferior version of Mozilla 1.0 sooner, and been looking at this release as Mozilla 2.0, but since the resources did NOT exist (Thanks to Netscape having their "air supply cut off") trying to work on two code bases with limited resources would have delayed the Gecko version of Mozilla even more.
Refactoring might have some advantages when you have to ship a product to maintain market share, but it dosen't apply here. Netscape was going to continue to lose share even if Netscape 5.0 had shipped the same day the decession to do a rewrite had been reached.
Beyond time constraints imposed by a need to maintain market share there is NO advantage to refactoring any many disadvantages. Look at all the problems legacy code has caused in Windows. Look at the Y2K mess we went through a couple of years ago. These problems were a direct result of having to refactor old code rather than rewritting the codebase because of market considerations.
IF you are so certain it was a mistake then you are more than welcome to pick up the old code base and refactor it as a starting point for your own project.