I started two startups and this is my journey.

bootload · on May 4, 2007

'... going for the Holy Grail of Software Engineering and developing an algorithm engine (deemed impossible) that will generate 1M + algorithms per second on a supercomputer or cluster of PS3s in (any) programming language. The other half of the problem is creating a visual 3D and manageable structure for software. Solving this problem has involved all of my brainpower ...'

Hey blake what's the problem your solving? Ah found it. Better still I read this article ~ http://myblog.rsynnott.com/2007/05/perils-of-outsourcing.html

'... My name is Blake Southwood and I'm the founder of Brontosaurus Software. I have recruited smart MBAs and 2 lawyers to work on the business plan and right now I'm in the midst of trying in vein to recruit lisp programmers for a software project to create high level machine thinking. ... The Catch-22 is that we can't pay programmers till we get funding ...'

Can you code? Ahh then I found this article ~ http://www.alu.org/pipermail/uk-lispers/2006q4/000169.html

'... now has hit a seemingly insurmountable roadblock which is the lack of demonstrateable software ...'

A demo? ...

'... In the end all startups morph their business plan ( I know that I did) and they go through lots of engineers and deal with management. But with a common goal and a collective brainpower it would be advantageous for foes (potential fledgling competitors) to become comrads and cooperate together for a common good. ... They would also immediately have more engineers on board (which is always good) ...'

But this ones my favourite ~ http://startupsmeetandmerge.blogspot.com .. Blake your a bloomin comedian. Your talent is wasted ... couldn't dream this stuff up even if I tried.

Hmm, reads like a troll, smells like a troll ...

blakesouthwood · on May 5, 2007

None of the lisp programmers that I recruited was brave enough to try; thus I'm learning lisp and coding it myself. Yes I can code. Even Franz chickened out. At least I have the courage to try.

bootload · on May 5, 2007

'... At least I have the courage to try ...'

it's not the courage bit that I'm skeptical about. Its the scope. It's fine to have a BHAG but try a smaller problem and solve it. This particular problem is someones academic career. Is it a a product someone would buy?

blakesouthwood · on May 5, 2007

Unless I figured out how it would work. I realize that I need to work on it in stages and get each puzzle piece working. At the same time with funding what would take me 20 years working alone could be accomplished in 2 years with say at least 20 programmers and some office space. In regards to whether it would be a product someone would buy it would be for companies that have programmers. The whole point is speeding up software development. It would make it possible to develop software in months rather than in years. I think that the empitus for the idea first occurred to me when I had to make a symbolic link at Schwab and no one at Schwab knew how Schwab wrote symoblic links except for one guy and he was in vacation in Lousiana. I decided that there needed to be a better way of 'talking to the machine" with a smarter programming language. It would definitely make money. You could do a month's worth or programming in a day. What's not to like?

blakesouthwood · on May 4, 2007

My first company (Big Bear) found a way of programming VCRs and DVDRs from a computer through USB, Ethernet or Wireless. It's possible to manage and program any number of video receivers in a house. It's sort of like Tivo with no subscription and it works with any brand.

My second company (Brontosaurus Software) is solving is streamlining software engineering with the algorithm engine so that programmers can write code by describing what the algorithm does and not how it works. So let's say on a good day you can write say 50 algorithms, what if you could write 1,000 instead and then the code that's generated is readable.

So the first problem is being able to write code based on human thinking as opposed to machine thinking. The second problem is generating readable code. The third problem is that the code that's generated is composed of reusable code. The fourth problem is creating structure for software so that managing the complexity is easier.

I eventually decided that only Lisp could so what I want to make it work.

The point is that the algorithm engine will make it possible for programmers to write code once and not spend time optimizing their code and fixing bugs. Fixing bugs and looking for them consumes too much time.

I was reading about the software bugs that cost billions and years of snags in the F-22 jet so I decided that I'd "discover" a way to solve the problem of writing code faster, making code more manageable, and making (most of all) source code readable and reusable.