[FRIAM] anthropological observations
Prof David West
profwest at fastmail.fm
Wed Apr 15 16:33:54 EDT 2020
True, it does advance an argument for a rational process (i.e. guidelines). But it also states that such a process can never be other than an idealization (as you note).
There could be value in after-the-fact faking of a rational process, if the fake was used to inform and "improve" the process such that it became a closer approximation to the ideal.
If you establish a continuum
1. hacking ---> 2. ad hoc heuristics / best practices ----> 3. informed heuristics / best practices ---> 4. rational process
then my interpretation of the paper seemed to suggest that faking the rational process could lead to iterative improvement and move the profession to stage three of the continuum.
But, everything I have read about Parnas, and especially when he was in charge of the SDI, seemed to be to be very pessimistic about the profession/practice ever getting beyond stage 2. His resignation letter from the SDI project stated that such a project mandated the use of something much closer to a rational process than was achievable in the foreseeable future.
I don't know if CPSR has any archives of the papers / communications from the early years when Parnas continued this argument and generalized it to most "mission critical" software. I read them at the time but do not have them in my files.
I don't think I am arguing against your point, but I am highly skeptical and deeply cynical about its being realized.
We have had very different careers in software development. Mine started in 1968 as a COBOL and Burroughs Assembler programmer and reached CIO before I returned to graduate school. All of that experience was in corporate or government software development organizations. Ninety-five percent of the students I taught at St. Thomas had professional experience in the same kind of organization and they shared their experiences and observations with me. There were a few students, and I had a few consulting encounters, with development processes that were reasonably formal — at Medtronic, IBM, CDC, and Cray for example.
>From that experience I concluded that formal, scientific, engineering approaches were nearly useless for developing the kind of software business/governmental organizations and teams were charged with. Clearly, there is some software that is amenable to and enhanced by engineering approaches. It is a matter of the nature of the underlying problem domain and the degree to which that domain is itself substantially a formal system.
The 2006 DoD funded Ultra-Large Scale Systems Study Report — Linda Northrop, lead, SEI at Carnegie-Mellon — confirmed my cynicism when they noted that software engineering was not and could not be the means for developing software for a complex domain like Ultra-large-scale-complex-adaptive-systems.
davew
On Wed, Apr 15, 2020, at 12:24 PM, uǝlƃ ☣ wrote:
> That paper:
>
> https://users.ece.utexas.edu/~perry/education/SE-Intro/fakeit.pdf
>
> argues *for* guidelines for software development. So, it validates my
> point in the most direct sense. It *also* argues against inferring from
> Nick's idea that there might be such a thing as Laws of Software
> Development Procedure, in that the ideal is never met. So, it validates
> my point about heuristics and best practices from that perspective, too.
>
> Did you intend to say that this paper is contrary to Nick's point? Or
> contrary to my point?
>
> On 4/15/20 8:02 AM, Prof David West wrote:
> > A contrarian position: David Parnas, "The Rational Design Process: How and Why to Fake It."
>
> > On Wed, Apr 15, 2020, at 8:43 AM, uǝlƃ ☣ wrote:
> >> No guidelines for how much to ship to any given hospital. No guidelines
> >> on dosage. No guidelines. We don't build bridges that way. We don't
> >> write software that way. We don't cook food that way. Etc. Why should
> >> we "treat" patients that way?
>
>
> --
> ☣ uǝlƃ
>
> .-. .- -. -.. --- -- -..-. -.. --- - ... -..-. .- -. -.. -..-. -.. .-
> ... .... . ...
> FRIAM Applied Complexity Group listserv
> Zoom Fridays 9:30a-12p Mtn GMT-6 bit.ly/virtualfriam
> unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> archives: http://friam.471366.n2.nabble.com/
> FRIAM-COMIC http://friam-comic.blogspot.com/
>
More information about the Friam
mailing list