To be clear I'm not saying that Sommerville invested software
architecture as a term, what I'm saying is that his work and that of
others detailed the importance of separating
data/components/connections/constraints and certainly in large scale
projects this was normal practice before Perry & Wolfe was written.

The application of the term "architecture" rather than "engineering"
isn't the important bit, its the actual "practice" that matters.
Definitions help us to then have a common reference point (which is
good) but they don't mean that previous work didn't fulfil the
requirements.

Steve


2008/11/11 Nick Gall <[EMAIL PROTECTED]>:
> On Tue, Nov 11, 2008 at 12:21 AM, Mark Baker <[EMAIL PROTECTED]> wrote:
>> As I say, I haven't read the book. If you say there's a model there,
>> then great. But none of his top several cited papers seem to have
>> anything to do with a model, so perhaps it wasn't very popular.
>> Editions of his book prior to the Foundations paper are not highly
>> cited (~25).
>
> Not to reopen this can of worms when you and Steve have just closed it, but
> I checked out the Sommerville slide deck that Steve referenced
> (http://sunset.usc.edu/~neno/cs477_2003/February11.ppt). Slide 10 is
> entitled "Definitions of Software Architecture" and it lists three
> definitions. The first is Perry and Wolf, the second is Shaw and Garland,
> and the third is Kruchen. As far as I can tell, Sommerville does not offer
> his own definition of Software Architecture in the deck. And he certainly
> doesn't date it.
> So on the surface it appears that Sommerville credits others with defining
> Software Architecture.
> -- Nick
> 

Reply via email to