Of late, I had a number of discussions revolving around the topic of "whether to start all over again." All softwware projects go through these questions. Every good developer has advocated just "throwing it all out and doing it right the second time" at least once. Many of us multiple times. Experience tempers our eagerness a little. So I've had these discussions with a number of people at various sites, including projects across 3 employers over the last three years. There's really nothing new here.
The "new" came as I thought about it earlier today, and realized I have one man to thank for some of this confusion: Alan Kay. He's quoted lots and the hero of many.
I've heard Alan speak at a number of conferences and usually gobbled up just about everything he wrote or said. One of Alan's big things is the cellular biology metaphors found in Smalltalk. He's been noted as saying it's the stuff in between, the messages that makes the magic happen. He likes to mention his involvement in the DARPA project that created what would become the internet. He points out that the "internet program" has never been rebooted as a whole even once. That it's ability to evolve and morph and become what it is is a testament to the power of loosly coupled messaging.
Now wait. Rewind one paragraph. Try it again.
I've heard Alan speak at a number of conferences and usually gobbled up just about everything he wrote or said. One of Alan's big things is the "Burn the Disk Pack" creed that was prevalent in the Xerox Parc days. Doing so allowed the group to not get "stuck in a rut" but repeatedly try new tacts of creativity. Smalltalk underwent 4 major changes (70, 72, 76, 80) in 10 years time. One might argue that there are more differences than similarities between them. ParcPlace, the company that commercialized Smalltalk -80, was berated for letting Smalltalk stagnate. Original OOPSLA attendees were admonished to get out there and do something original instead of attending a conference. I heard the same challenge at Smalltalk Solutions 2001.
So which is it? Should developers of the internet have burned the disk pack? I don't think there's a right answer. Maybe the contradiction's not as real as it seems to me. I'm definitely not trying to belittle Alan Kay. If anything, noting the juxtaposition occurs even for one of the computing greats, makes me feel a little better as I work my way through it.