Boffins: Dark times for application development
Ex-Microsoft man holds flashlight
You know it's bad when two top programmers at different conferences in a two-week period say we're in the dark ages of software development.
Speaking at MIT's Emerging Technologies conference in Cambridge last week, Charles Simonyi - creator of Microsoft Office, friend of Bill Gates (and Martha Stewart - thanks ZDNet's Dan Farber), cosmonaut and all-round achiever - said current paradigms and tools for developing applications are primitive and must evolve beyond languages and methodologies.
Evolve to what? To include code written by business analysts.
"They are still in the dark ages," Simonyi said of current programming techniques, including model-driven development paradigms such as Unified Modeling Language (UML) and Domain Specific Languages (DSLs) - lately favored by Microsoft. They invite more input from domain experts and business analysts, but that still fails to encapsulate the business problem within the code.
"We are on the verge of the renaissance but we are not there in the practical world," the billionaire told hundreds gathered at MIT's Krege Auditorium.
That's the intent of Intentional Technology, a company Simonyi co-founded in 2002 after a 25-year career with Microsoft. The 20-employee start-up in Microsoft's Bellevue, Washington, has developed a platform and workbench technology that allows programmers to create generic "generators" that ordinary business analysts use to project a representation of the business problem in terms they understand.
"It's like a super-duper PowerPoint because it projects something that is machine processible," said Simonyi, who invented the first What-You-See-Is-What-You-Get (WYSIWYG) word processor, dubbed Bravo - the precursor to Microsoft's evergreen Word.
This approach, which is platform-independent and not wedded to Windows, focuses on the recipe - and how to mix and cook the ingredients - rather than the meal itself. According to Simonyi programmers today spend too much time rearranging or adding ingredients rather than on finding better ways to mix them.
Software is just a tool, not a solution
I am primarily an engineer so software to me is as intuitive as any other method of implementing a logical machine. Its easy. Except that if you take me out of context, into, say, an applications environment my productivity flounders. What's the problem?
The answer is simple. Programming is just an expression of problem solving. The real money is knowing how to solve the problem, not how to crank out code. Non-programmers and even some less enlightened programmers don't understand this, they like their applications templates because they can make things work with them, but all they're doing is driving the car that someone else designed for them. If the need to step outside the boundaries of their tool they're, not too put too fine a point on it, screwed.
Business people like to reduce creative processes to rote because it allows interchangeability ("flexibility") of staff. Large companies spend a lot of time and effort on processes, they're supposedly to increase quality but it all boils down to trying to de-skill skilled workers. (Skilled workers have power, they don't like that because it makes them uncomfortable.)
I avoid the bloatware. I have my own problems. Its amazing how easy it is to run off the edge of your knowledge or capabilities once you start messing around inside a computer -- sometimes really simple tasks are impossible, intuitive algorithms don't scale, really Mickey Mouse stuff is just off the scale difficult. Its what makes life worth living.
Purple Prose... HyperRadioActive Pros
"Simonyi was quick to point out his friend is leaving behind his day-to-day duties at the company to focus on the Gates Foundation, solving bigger problems than software development"
How very bizarre that he doesn't realise that they are both inextricably linked.
A most definite case of "Not seeing the wood for the trees".
""Software programming for multicore is hot, and some of you have the opportunity to be the next Microsoft if you can solve these multicore issues," Agarwal said."..... and aint that the gospel truth.
Who would bet against it being developed Virtually Online, Registering ITs Prior Art Patent credentials in Simple Text Shared?
And I wonder when El Reg will realise what is going on and what is be floated, intentionally, out of the Bletchley Archives, Turing Module?
Simonyi's race has already been run and won. IT is now just a case of collecting the Danegeld reward which reveals the code which will automatically, by selfish Third Party Proxy Default, completely destroy Systems in Markets..... and just so as not to be ambiguous, that does mean any System in any Market.
AIMagic Spell from Blithe Blighty Boffins? Or a Rampant, Rutting Stag of AIdDriver Program? And is there any difference?
Another layer to add some bloat - just what is needed...
Except by the sellers of bigger and better hardware who can sell us boxes that will just about let the code limp along.
I actually wanted a smart car so I bought a BMW and configured it so that it looks and runs like a smart car.