jerbil@cit-vax.Caltech.Edu (Stainless Steel Gerbil [Joe Beckenbach]) (02/22/89)
As I was following the thread of discussion about comments and 'good coding style' with regards to these, I thought of a neat little shift which I think embodies what this facet of software engineering should head towards: not "self-documenting code" but "self-coding documents". The Hyper-Programming idea thrown out by one participant gives the right slant: combining the documentation, graphics, and code into a more comprehensive source. I'm not saying that this type of environment is for everyone [heck, it'll have trouble fitting onto most of the current computers, which are smaller than the Vax I'm working on]; it's just that this concept brings the commentary, the documentation, and the human-readable context of the code in question [including design!] into a spotlight where the driving expression behind the code can be given at several levels and in several forms. I'm getting ready to go back and help work on some code I wrote last summer; at the time I thought I was writing enough comments and documentation; I'm pretty sure I'm going to have quite a surprise when I start trying to correct my misunderstandings of my own code. Just my two cents and a new twist on an 'old' viewpoint. -- Joe Beckenbach joe@csvax.caltech.edu Caltech 256-80, Pasadena CA 91125 Users I'd like to see: Postmaster@link.L5.edu gorby@party.kremlin.gov.cccp
brooke@ingr.com (Brooke King) (02/25/89)
In article <9674@cit-vax.Caltech.Edu> jerbil@cit-vax.caltech.edu (Stainless Steel Gerbil [Joe Beckenbach]) writes: | | As I was following the thread of discussion about comments and | 'good coding style' with regards to these, I thought of a neat little shift | which I think embodies what this facet of software engineering should head | towards: not "self-documenting code" but "self-coding documents". | | The Hyper-Programming idea thrown out by one participant gives the | right slant: combining the documentation, graphics, and code into a more | comprehensive source. | Just my two cents and a new twist on an 'old' viewpoint. Some old information which is somewhat related: Iris Vessey and Ron Weber, "Stuctured Tools and Conditional Logic: An Empirical Investigation," _Communications of the ACM_, January 1986, Vol 29, Nr. 1, pp. 48-57. The article empirically hints at what we all feel, that linear programming languages are inferior to well-done multi-dimensional (perhaps graphical) programming languages. The whole August 1985 issue (on visual programming) of _IEEE Computer_ is pretty interesting and some of it was referenced in Vessey and Weber. Anyone have any pointers to some more recent work (that is also not too heavy)? | Joe Beckenbach joe@csvax.caltech.edu Caltech 256-80, Pasadena CA 91125 | Users I'd like to see: | Postmaster@link.L5.edu gorby@party.kremlin.gov.cccp -- brooke@ingr.com uunet!ingr!brooke W+1 205 7727796 H+1 205 8950824