Kinda funny that people are still admiringly quoting Joel's criticism of Netscape for going with the big rewrite, now that that rewrite has turned into Firefox.
It's been a while, but iirc the rewrite of Netscape's browser started soon after they opensourced it, when they were still a going concern. Later Firefox came along and repackaged the core code.
How long did it take to get to the halfway decent (and I do mean halfway) browser that it is from the rewrite? How long would it have taken to get there /without/ the rewrite?
EDIT: If you're going to downvote at least explain why...
You know how you think you're 90% done with a project, and there's a ton of little details, so that in reality, you're actually a long way out? That's where Cassandra is at.
I think there problem is that they are already riding on the edge in terms of service stability and availability, trying to conduct a large scale migration now would probably just add to this. I don't think they can really afford extended downtime now just when they're beginning to implement their monetisation strategy.
Unless if you have to go back and keep on fixing it again and again and again and again... until one day, the cost of fixing (please also include the cost of re-testing) can't be justified any longer.
Anyhow, that's why software should be developed in modules/components.
