- From: Sandro Hawke <sandro@w3.org>
- Date: Tue, 18 Sep 2007 07:01:56 -0400
- To: Adrian Giurca <giurca@tu-cottbus.de>
- Cc: Sandro Hawke <sandro@w3.org>, "Boley, Harold" <Harold.Boley@nrc-cnrc.gc.ca>, RIF WG <public-rif-wg@w3.org>
Adrian Giurca <giurca@tu-cottbus.de> writes: > I guess one good document is Scott Ambler's > http://www.ambysoft.com/downloads/javaCodingStandards.pdf Nice document. I take it as supporting my proposal. It doesn't generally use the word "noun", but all the relevant examples fit. (Sections 3.11 for properties/fields and section 6.1.2 for classes, also overview in section 11.1) > Some other which might be used for guidance are: > > Gosling, J., Joy, B., Steele, G. (1996). The Java Language=20 > Specification. Reading, MA: Addison Wesley Longman Inc. > Ambler, S.W. (1998a). Building Object Applications That Work: Your=20 > Step-By-Step Handbook for Developing Robust Systems with Object=20 > Technology. New York: Cambridge University Press. > > All of them argue in the favor of a notation as in [2]. > > Other possible notations are: > Hungarian notation (McConnell, S. (1993). Code Complete =96 A Practical=20 > Handbook of Software Construction. Redmond, WA: > Microsoft Press.) > > > - Adrian > > Sandro Hawke wrote: > > Harold, you made changes to the Naming Conventions page [1] which seeme= > d > > to me to combine editorial changes, changes to the conventions, and > > arguments motivating those changes. I've refactored the page to try to > > clean that up. I hope they all were true to the spirit of your edits. > > > > Meanwhile, I hear you arguing that the grammatical considerations are > > not appropriate in an international context -- that there's no reason > > class names or property names should be nouns/noun-phrases. Is that > > right? But the Java convention that class names should be nouns or nou= > n > > phrases is very well accepted, isn't it? [1] In fact, google's second > > hit (after [1]) on "class naming conventions") also supports the use of > > nouns for property names [2]. > > > > I think the counter-argument against the conventions I'm proposing is > > that "ExistentiallyQuantifiedFormula" is too darn long, so "Exists" is > > fine. But if that's the case, let's just have a convention that names > > be short, and justify the shorter names on those grounds, because there > > was no short noun-phrase name for this class. =20 > > =20 > > -- Sandro > > > > [1] http://java.sun.com/docs/codeconv/html/CodeConventions.doc8.html > > [2] http://www.objectmentor.com/resources/articles/naming.htm#Nouns_Ver= > bs > > > > =20 >
Received on Tuesday, 18 September 2007 11:02:12 UTC