perlmeditation petdance On Saturday, June 23rd, [http://www.conway.org|Damian Conway] had a little free for all free-for-all workshop that he gave at [http://www.cod.edu|College of DuPage] in Wheaton, IL. He talked about a ton of different Perl topics, all of them enlightening. He's a great speaker, and I wish I could have gone to the 3-day class that he gave earlier in the week. Here are some random notes of interestingness that I scribbled down. Perhaps [frag] will want to add some of his own comments. <UL> <LI><A NAME="item_%22%24%2F_tells_when_to_stop_reading%22">``<CODE>$/</CODE> tells when to stop reading''</A> We all know about the <CODE>$/</CODE> variable, but a new light dawned when Damian said ``<CODE>$/</CODE> tells when to stop reading''. That's why <CODE>$/ = undef</CODE> makes Perl read an entire file: There's no definition of when to stop reading. How clear it is now. <LI><A NAME="item_Perl_6_may_have_the_%2E%2E%2E_operator">Perl 6 may have the <CODE>...</CODE> operator</A> Perl 6 may have the <CODE>...</CODE>, which Damian pronounced as the ``yada yada yada operator.'' You'd use it as a placeholder for undefined code like so: <PRE> sub mangle { ... return 1; } </PRE> Yes, that's a literal <CODE>...</CODE>. Executing <CODE>mangle()</CODE> would have Perl kick out a message like ``sub mangle is pending future code''. <LI><A NAME="item_Typeglobs">Typeglobs</A> ``Typeglobs are like a box of chocolates'', he said, and then explained why in a way that let me finally understand the method behind the madness of typeglobs. Of course, typeglobs are going away in Perl 6. <LI><A NAME="item_Perl%27s_phases">Perl's phases</A> Big discussion of Perl's 5 phases: BEGIN, CHECK, INIT, execution and END, and when you'd want to use which one. <LI><A NAME="item_The_benefits_of_golf">The benefits of golf</A> He didn't specifically explain that golf in the extreme way we golf here is good, but rather that ``The fewer characters you type, the fewer characters you can get wrong.'' He claims that there's a linear correlation between the number of characters in a system's source code and how many bugs it has. <LI><A NAME="item_Qualifiers%2C_not_control_structures">Qualifiers, not control structures</A> In this code <PRE> foreach ( @foo ) { whack($_); }</PRE> the <CODE>foreach</CODE> is a control structure. In this code: <PRE> whack($_) foreach @foo;</PRE> the <CODE>foreach</CODE> is a ``qualifier'', which is why you can't apply it to a block in that syntax. <LI><A NAME="item_Perl_optimizes_a_lot">Perl optimizes a lot</A> The Perl pseudocode has no <CODE>if</CODE>s in it. Anywhere. ``How can that be?'' I asked. He then showed the following reduction: <PRE> if ( $x > 50 ) { do_something(); }</PRE> is the same as <PRE> ($x > 50) && do_something();</PRE> It's all just a lot of booleans in there. <LI><A NAME="item_The_use_of_the_v_notation">The use of the <CODE>v</CODE> notation</A> Version strings are tracked with the <CODE>v</CODE> notation, like <CODE>v5.6.0</CODE>, which gives a string equivalent to <CODE>chr(5).chr(6).chr(0)</CODE>. This is also handy for IP addresses, as in <CODE>v127.0.0.1</CODE>. <LI><A NAME="item_Attributes">Attributes</A> Big discussion of attributes and the Attributes::* modules, most of which I didn't note because I don't see any direct application for them in my work. However, Perl 6 is going to make a LOT of use of attributes. <LI><A NAME="item_Object%2Doriented_design">Object-oriented design</A> He presented his Spinal Tap-style list of [91080|Ten criteria for object-oriented design]. This list will eventually be part of the standard Perl distro. <LI><A NAME="item_What%27s_coming_in_Perl_6">What's coming in Perl 6</A> Since Damian is sort of Larry's right hand man for language design issues, he's on top of all the changes that are going to be coming. Some of my favorites: <UL> <LI><A NAME="item_Here_document_improvements">Here document improvements</A> Here documents won't have to be flush left. <PRE> while ($foo) { for ( @list ) { print << "END"; Here's the text of the message I want to print. Note that I'm not flush left like I would have to be in Perl 5. The amount of whitespace before the terminating "END" is stripped from the front of each line. END } # for } # while</PRE> Many of us literally applauded and cheered. <LI><A NAME="item_Typeglobs_go_away">Typeglobs go away</A> Everything will still be in symbol tables, but not as they are now. The fully qualified name of <CODE>$foo</CODE> will be the key in the symbol table, not like it is now, where <CODE>$foo</CODE> really means ``the scalar in the <CODE>foo</CODE> typeglob''. Also, lexical variables go into their own special symbol table <CODE>MY</CODE>. <LI><A NAME="item_Globals_go_away">Globals go away</A> All the global punctuation variables will go away or become properties of something else. For instance, instead of setting <CODE>$/</CODE>, you'll modify the <CODE>insep()</CODE> attribute of the file handle in question. </UL> </UL> And here is where my notes end. I wish I had a videotape of it. He's just great. xoxo, Andy -- I was dreaming when I wrote this, so sue me if I go too fast.