[raganwald]

And I can't tell you how many doomed Death March projects started with this assumption. "The requirements are done, the design is done, we know it can be done, simply copy what we see there."

The boil it down to a cliché, copying what someone else has built is exactly the same problem as rewriting an existing application from scratch, without the benefit of being able to read the source code.

http://www.joelonsoftware.com/articles/fog0000000069.html

[davidw]

Rewriting a cross platform browser in C++ is way more difficult than writing something that took a small team 6 months to do. I'm not trying to say that it's a weekend, but have a look at SO. Unless there's something I'm missing, it's not "deep hacking", but more "crank it out" (not to take anything away from that team - they certainly did nice work!). Scaling might be an issue sooner or later, and that would take some more thinking, but the basic thing just doesn't strike me as some Big Hairy Problem.

So, yeah, be wary of "hey, that looks easy!", but don't think things are impossible either.

[raganwald]

Rewriting a browser isn't impossible either, as Firefox proves. Perhaps SO is an order of magnitude easier than a browser, but I stand by my suggestion that copying an existing application you didn't actually write is often much more difficult than expected and has led many teams to their doom.

[davidw]

My comparison was "writing something from scratch" vs "copying". If a team has trouble with copying, wouldn't "writing it from scratch" be even more likely to end in failure?

Also, with something like SO - you can pretty much see the models, and how they interact. That's already giving you a lot.

[dinkumthinkum]

That's interesting. However, there are already several clones of SO and a couple that reproduce the vast majority of the functionality.

[wlievens]

You can't even check that, because a significant part of SO's functionality could be hidden from you.