> as I recall, it still lacks things > like sessions or authentication.
Not sure about that. Randal evolved it for a shopping-cart type app, so I would think it has something like that. > More importantly, does anyone "get" Randal's CGI::Prototype? See http://www.stonehenge.com/merlyn/LinuxMag/col70.html and 71 and 72 for 3-part worked example(s) [Linux Magazine May-July, 2005] There's also http://sourceforge.net/mailarchive/forum.php?forum=cgi-prototype-users > Usually Randal's recommendations or modules have a readily apparent > elegance to them, Indeed. The elegance here comes from the use of Class::Prototyped, which is the inspiration. http://search.cpan.org/~teverett/Class-Prototyped-1.10/lib/Class/Prototy ped.pm > brings relative to the added complexity it introduces. CGI::Prototype offers a _different_ way of factoring out the "you always had to write this glue code" code. Catalyst uses the Perl Attributes annotations to factor out glue-code, which is classy demonstration that attributes are a good idea. CGI::Prototype uses prototypical (instance-based, or nonce-class) inheritance. Prototypical inheritance is not your classical C++/Java rigid Object Isa Class inheritance structure, or even SmallTalk 1st-Class 2nd-Order Class objects, so is somewhat foreign on first impression. Every object has it's own class, which may be same or similar to other objects'; the key relation is not "Isa" but "Has" (and "Can"). This model goes back to antiquity, via the Self language (1980's) to the earliest days of A.I. in Lisp (late 1960's or early 1970's?) and the Frame/Slot/Attribute/Facet theory of knowledge representation (later in Prolog and other engines as well). Some of us have done similar things in C++ and SmallTalk; there was a blossoming of this model in the Expert System rage of the 1980's. Prototype-based nonce-classes are highly flexible, very cool, but hard to get your head around the first time, since it's so heterodox (outside the Lisp / A.I. community, that is, and much of Lisp moved on with Scheme and CLOS standards). For background on Frame-based knowledge bases, see for instance http://www.cse.unsw.edu.au/~claude/teaching/AI/notes/prolog/Frames/Frame s.html http://deed.megan.ryerson.ca/~fil/I/Papers/kicIII/node10.html http://www.amzi.com/ExpertSystemsInProlog/06frames.htm (Aside -- Frame-structures are inherently XML friendly, http://www.ai.sri.com/pkarp/xol/xol.html but antedate XML per se by 20+ years -- there was likely cross-fertilization between Frame folks and SGML folks in the 80's.) -- Bill _______________________________________________ Boston-pm mailing list [email protected] http://mail.pm.org/mailman/listinfo/boston-pm
