October 7 turned 40 years, Software Engineering, because 40 are those that have passed since the NATO conference (1) which he named this new professional discipline, born to solve the rudeness of the software on military projects, which was losing millions of dollars because it always wrong and never delivered late.
40 years ago that launched the first tracer bullets into the solutions, even those who might believe they were shot dead and accurate and definitive.
Even those still blindly in their wake, think that this is the goal of knowledge, not a point on the road (particularly the beginning v. synthesis)
The more we strive to follow the path, uncorrected the shot, the more enlarged the angle error, and the more we move away from the target.
The initial directions of Software Engineering, were: rigor and precision in the requirements and product design, planning and project control.
and all solutions were: from and to their context. A context of major military projects with millions in losses in the software sub-systems, and would be foolish to criticize them for not serving for different projects in a scenario 30 years more advanced technology. Criticize them for heavy and unsuitable for small start-ups "or groups that do not program the ballistic missile guidance, and do not need predictability but fast and continuous innovation.
As is also awkward question is not what we do.
Consider that the solution we have been taught is valid for all, and at all times. For all, convinced that our new ceremonial dress is appropriate for any occasion. And at all times, with an "attitude" Peter Pan "that does not evolve professionally.
If to err is human and to rectify wise, Tom DeMarco is the latter. His work is one of the main contributions to the development of software engineering.
is the author of a major work on metrics in the software project management concerning the agendas as PMBoK: Controlling Software Projects: Management, Measurement and Estimation and to mark the 40 anniversary of Engineering of Software, July / August SOFTWARE IEE published an article in which, rather than say anything, and do not know if using or abusing the right to quote, I prefer to stick his words literally:
The metrics initially described in my book Controlling Software Projects: Management, Measurement and Estimation, have defined the way in which many software engineers built and planned work. With a reflective mood state, now I wonder: Was the advice correct metrics? Is it still relevant? and what still believe that metrics are a necessity for any successful software development?. My answers are no, no and no.
Tom DeMarco is also the author of the statement that in recent decades has been an axiom for many processes and management models (all?) "You can not control what you can not measure."
And now, the author asserts that the control may not be important in many software projects:
"Many projects have proceeded without focus management on the control, but the creation of wonderful products as GoogleEarth or Wikipedia.
Understanding the true function of control is necessary to distinguish between two drastically different types of projects:
A project of type A, with an estimated cost of one million dollars and an estimate of return of approximately 1.1 million.
A Type B project with an estimated cost of one million dollars yields a value of more than $ 50 million.
What is immediately apparent that the control is important in the project A, and yet their importance is minimal in B. This brings us to the strange conclusion that control is important extrict minor projects, and vice versa.
Me comes to mind the principle of "subtle control" identified by Nonaka and Takeuchi in the fields of Scrum to continue reading in the article draws comparison with the type of control that a parent makes in educating its teen:
"In applying the principle 'can not control what you can not measure' to education in their teens, most of the really important things, honor, dignity, discipline, personality, values, ethics, intelligence, loyalty, humor, kindness ... are not measurable.
You have to train your child the best you can without feedback of metrics. It's hard, because being a parent is difficult. You have a certain metric on the type of notes from school, and intuit note that the math is more important than the English and the note says quzá behavior over the student teacher. "
For those who believe that knowledge is continuously evolution, and sometimes nausea and arrive in the developing lines of management metrics and inappropriate for many projects, read these statements by Tom DeMarco reassured and glad to see that there are people that question honestly, refined and improved continuously (2) . That ultimately claimed that years of experience make them challenge and improve.
course, this is what it seems to me. Surely those who prefer online metrics PSP, CMMI models and for all, would argue that Tom DeMarco is a shame. With good ideas I had, and has now spoiled. It has gone to the dark side. :-)
* Article Software Engineering: An Idea Whose Time Has Come and Gone?
(1) Conference on Software Engineering. 1968 Garmisch, Germany. Conference Report
(2) Since its initial ideas on the line "You can not control what you can not measure" to the current findings on metrics and project management (agile, why not say so) through the at the end of the 80 summarized in his statement "Most of our work problems are not technological but sociological" (Peopleware: Productive Projects and Teams 1987)
Source: http://www.navegapolis.net
