Billions and billions of...lines of proprietary code to go open source

Eric Raymond made the point years ago that most software is written for use, not for sale. Eric put the number at 95%; that is, 95% of all software is written for in-house use, rather than for sale.

If he's right, and I believe he's not far off, then banks, manufacturers, retail chains, etc. are sitting on a massive gold mine of software.

Ken Krugler, founder of Krugle, agrees (as noted in this article in InformationWeek), and has put some back-of-the-napkin calculations together that suggest that billions of lines of code have been written for use, with trillions of dollars in associated development costs involved...and most of it with a good reason to go open source:

When you factor in salaries and other expenses, Krugler says it costs $50 to $100 per line of code developed. That would put the investment in proprietary, internally developed software, measured in today's dollars, in the trillions. Companies are looking to capitalize on that massive investment. Krugler expects a significant percentage of corporate code to become open source as CIOs try to benefit from the dynamics of community development and software sharing.

Is there enough "community" to manage that much open-source code? I don't know. But I suspect that where there's exceptional code there will be exceptional entrepreneurs to make use of it.

Before any of this can happen, of course, armies of lawyers will need to formally bless all this code. Perhaps for this reason it will never happen. Perhaps we'll only get a trickle where we should get a torrent.

Either way, I can't wait. Whether trillions or billions or even millions of dollars worth of developed code, it will be welcome in the open-source world.

[JYReynaud]

How much duplicate code

I wonder how much duplicate code within this 95%. I guess a lot. Another good reason to open source code. Spend your energy improving rather than re-creating.

[kkrugler]

@JYReynaud - you're right, there's going to be lots and lots of duplicate (or rather, near duplicate, which means re-invented) code out there. If it's not visible and findable then there's no way to reuse it - and that's the real problem, which is why there have been all those failed code reuse initiatives and products and methodologies that we've seen over the years.