The process in the large doesn't have to be rational

Edward G Nilges — Thu, 14 Sep 2006 07:51:45 -0700

In ECLIPSE OF REASON, Max Horkheimer speaks indirectly to the plight of MIS developers who'd rather write tools.

In the Renaissance of the West, the merchant class learned to adapt to short-term needs. This was the "material basis", in part, to skeptical philosophy and modern science.

The contemporary pretense is that the West is the simple avatar and progressive evolution of this willingness to at all time to take short-term advantage based on "reason", and that there is no alternative.

But as experienced by real people in real life, "reason" becomes indeed the wreckage of THEIR "reason" (their software architecture) by a trumping "reason" which uses authoritarian structures to insist on its way, and which in the case of corporate and government secrecy need not divulge why the user now "wants" a different architecture, and get back to work.

The fact is, as Horkheimer shows, it becomes impossible for individuals to be "individual" ("in DUH vidual" in the cruel mockery of Dilbert) in a society which proclaims its valuation of individuality!

Horkheimer shows how in the 17th century, Leibniz' "Monad" became the model for individuality. A "monad" exists independently of all other monads, and can step over homeless men in the street.

But what we find is that in a process which strangely resembles crystal formation, the monads, in optimizing their in DUH viduality, become, if they know what's good for them, individual crystals in a lattice of ... identical crystals.

In MIS, in fact, they all know the same required laundry list of technologies even if they harbor a voiceless resentment of Crystal Reports, knowing that a Crystal palace is not Utopia.

[The origin of Crystal Reports was the simple fact that circa 1978, if you had coded 50+ report programs, you realized if you didn't totally suck that they all were implementing the same basic algorithm, and that you only needed to generalize the ontology from "print line" to "block of information meaningful to the user containing small n lines", and then scale up, to get a report GENERATOR.]

[This is the type of thinking that denies the constructed firewall between MIS and "tools" and it shuddha made me rich but my inner Buddha had other plans for me.]

In MIS the proud, individual Masters of the programming Universe wear all the same khakis and troop on time to the Bangalore Hooters for lunch.

[I do hope you bad boys have a Hooters in Bangalore by now, with all due respect for local custom goes without saying. My point is that in places like Shenzen, the culture I saw in Schaumburg, Illinois, was replicated like a self-duplicating crystalline cancer on the shores of Shekou bay.]

Not a mathematician, Horkheimer had a mathematical insight. Rule driven monads, or cells in a cellular automaton, generate not stochastic patterns but patterns in the large, for the same reason I discovered that John Conway's LIFE would generate spaceship builders in 1971.

Horkheimer took this evolution seriously. To step back and creatively craft a tool to anticipate requests goes backward to a time when producers controlled the market.

British coal miners asked Thatcher for the right to keep their jobs even when there was no demand for coal, and to stack the coal they dug in neat piles awaiting an energy shock.

In the typical software company, to develop a general tool is to stockpile and it goes directly against global evolution/devolution. Even as the merchants of the 17th century learned to treat silk from India and China interchangeably, and not describe themselves exclusively as dealing in the one or the other, the evolution is in fact towards even more reactive, even more opaque coding of proprietary solutions.

To be a toolcrafter in 1978, you had to leave renumerative and paid employment even as Wozniak left Hewlett-Packard. To "write a compiler" in 2003 I had to come close to starving in a San Francisco hotel room.

No company is going to pay you for disruptive technology. And "being an entrepreneur" isn't the answer, because entrepreneurs entrep, they don't write code.

Enterprise Software Considered Harmful (2) - comments on Ravi Mohan's blog

Edward G Nilges — Tue, 12 Sep 2006 22:38:32 -0700

Further comments on Ravi Mohan's blog post about enterprise software development vs. tool development, picking up where Part 1 leaves off.

developer.* Blogs - Enterprise Software Considered Harmful (2) - comments on Ravi Mohan's blog - Comments

The process in the large doesn't have to be rational

Enterprise Software Considered Harmful (2) - comments on Ravi Mohan's blog