Loading ...
Sorry, an error occurred while loading the content.
 

Re: [XP] Theologists and missionaries

Expand Messages
  • Paolo Bizzarri
    ... Yes. But most of formal methods research was trying to tackle exactly the same problems as XP and TDD. This was funny for me: there was lots of research
    Message 1 of 31 , May 1, 2007
      On 5/1/07, John Roth <JohnRoth1@...> wrote:
      >
      >
      >
      >
      >
      >
      >
      > ----- Original Message -----
      > From: "Paolo Bizzarri" <pibizza@...>
      > To: <extremeprogramming@yahoogroups.com>
      > Sent: Tuesday, May 01, 2007 8:13 AM
      > Subject: Re: [XP] Theologists and missionaries
      >
      > > On 5/1/07, John Roth <JohnRoth1@...> wrote:
      > >>
      > >>
      > >>
      > >>
      > >>
      > >>
      > >>
      > >> ----- Original Message -----
      > >> From: "Paolo Bizzarri" <pibizza@...>
      > >> To: <extremeprogramming@yahoogroups.com>
      > >>
      > >> Sent: Monday, April 30, 2007 11:41 PM
      > >> Subject: Re: [XP] Theologists and missionaries
      > >>
      > >> > On 4/30/07, John Roth <JohnRoth1@...> wrote:
      >
      > ...
      >
      > >>
      > >> It did, however, come out of the Smalltalk environment. Most of
      > >> what we think of as XP came out of the Smalltalk environment.
      > >>
      > >
      > > But my point was that
      > > there was no trace of something like TDD in the Computer Science
      > > theory, at least as it was teached in Italy among the end of the '80
      > > and the beginning of the '90.
      >
      > I don't consider that surprising. A lot of university courses look for
      > "academic respectability." Djikstra, Hoare, Wirth, etc. are academically
      > respectable; they had Ph.Ds (or the equivalent), they were professors
      > of various things, they published papers. People like Alan Kay, Kent
      > Beck and so forth are not acacemically respectable. They aren't on
      > university faculties, they don't have a long list of academic publications
      > that have high citation indices, etc.

      Yes. But most of formal methods research was trying to tackle exactly
      the same problems as XP and TDD.

      This was funny for me: there was lots of research on things like Lotos
      and the like.

      >
      > >
      > >> Now consider the environment that Wirth, Djikstra, Hoare and their
      > >> collegues worked in. They did not have personal workstations
      > >> at their disposal, neither did their students. They did not grow up
      > >> in that environment.
      > >
      > > Yes. But this was exactly my point: they could not come with something
      > > better than weakest preconditions, because they were not developing
      > > code. They were neither interested, nor ready to produce a better
      > > theory for the current environment.
      >
      > It would certainly be interesting to see what people of that ability
      > level would come up with today.
      >
      > However, that doesn't mean that I consider what they did irrelevant.

      I don't think it is irrilevant. I think that their research, at
      certain point, missed to close the loop and going back to software
      development practice. If they had done, they would have found (IMHO)
      TDD ten to twenty years ago.

      > Learning formal methods is very useful: it will change the way you
      > look at code, generally for the better. Actually applying them is much
      > less useful: they're in use today for life and safety critical applications
      > as well as high reliability applications. They're very expensive overkill
      > for the vast majority of business administration applications.

      I know formal methods, and I have found that they are strictly related
      to unit testing and assertion.

      Most formal methods try to model your code, and define properties that
      are invariants through your code.

      Things like xUnit and Junit try to do exactly the same. The critical
      difference is that formal methods were using another language
      (typicallly, some form of logic predicates or finite state automata or
      petri nets). xUnit framework used exactly the same programming
      language. I think this was the critical point.


      > >
      > > As soon as I have seen TDD, it was obvious that this was something I
      > > have learned in my courses in Computer Science.
      >
      > I presume you meant 'should have learned'. I agree. There are some
      > universities where XP and TDD are being taught, but it's very much an
      > uphill battle.

      No. I meant what I have learned. I have learned a lot about weakest
      preconditions and the like where I got my degree in Computer Science
      (in Pisa). But my professors were obviously a lot more confortable
      with formal systems than with code.

      But the idea is exactly the same.

      > >> Their entire background, and their day to day
      > >> working environment was the batch compile and test environment.
      > >> Is it any wonder that the problems they worked on, and the results
      > >> they got, fit the batch compile and test environment and consequently
      > >> have little relevance to people who have personal workstations that
      > >> can do an edit, compile and test run in a matter of a minute or two?
      > >
      > > No. But things have changed from then. And the theorists were quite
      > > late at seeing the changes.
      >
      > I don't think we have anyone of that stature today, except Knuth,
      > and he's engaged in finishing The Art of Computer Programming
      > before he dies (and I sincerely hope he makes it!)

      Me too.

      Paolo Bizzarri
    • Steve Freeman
      There are actually a few. Sheffield has been running a very interesting programme for several years now (that s why the XP conference was held there). I ve
      Message 31 of 31 , May 3, 2007
        There are actually a few. Sheffield has been running a very
        interesting programme for several years now (that's why the XP
        conference was held there). I've been teaching a few bits and pieces
        with Ivan Moore at University College. I think Huddersfield does
        something. It's not exactly a landslide, but it's catching on.

        S.

        On 1 May 2007, at 16:49, Elizabeth Keogh wrote:
        > IIRC Aberystwyth is the only university in the UK which teaches any
        > of the
        > Agile practices - there may be a second by now (?).
        >
      Your message has been successfully submitted and would be delivered to recipients shortly.