The New NASA Calendar

Perl Development Essentials

A recent question on the SF.pm mailing list about IDEs prompted me to think about what I depend on to code Perl.

I've been spending some time recently in GUI IDEs for other languages, especially MonoDevelop and Visual Studio.  MonoDevelop is snappy and helpful.  VS is of course a bloated mess.  (Though if I were developing for Windows I probably couldn't live without it.  Fortunately this is not my lot.)  GUI IDEs are like C++ socket wrappers: Making one is easy, but making one that actually helps you get your work done is harder than it looks.

However, I am accustomed to Emacs and I simply cannot do development without shell-fu.  Frankly I'm surprised anyone does.  IDEs are simply not well suited for power coding.  Too much clicking and so little progress for all that.  However, some real, earnest, full-power IDE support for Perl inside Emacxs would be excellent.  For example, automatic runs of "perl -c" and highlighting of error lines would be beautiful, and I would give real money for automated refactoring to e.g. rename things globally and safely.  Perhaps given the libperl integration of vim it would be better to build the smart bits with vim and PPI rather than trying to shoehorn it into Emacs?

I've had recommended Perlnow for Emacs and perl-support for vim.  I'll give them a look.

Meanwhile, today, I count nine tools that are vital for my Perl development.  In roughly descending order of importance, they are:

• CPAN and search.cpan.org, though that's a given
• emacs with cperl-mode (make sure it's up to date!) and some easy customizations -- e.g. a function key for grep which defaults to ack at project root (well my own script that predates ack)
• zsh.  Like I said, there is absoutely no substitute for shell-fu
• perl -d.  I simply cannot live without this, and neither can you.  If you think you're doing OK without it, you're kidding yourself
• Thinkpad.  These keyboards simply have no equal.  I don't understand people who code on Macs ... not because of the software, though that is a minor issue, but because of the perverse industrial design of the keyboards.  Beautiful pains
• git, or hg in a pinch.
• Debian.  (Sorry, Mac OS, if you ran on a machine with a decent keyboard I'd think about it.)
• cpanm
• vi.  Emacs is too heavy for quick hacks

Let me know if you know of any good Perl IDE hacks for emacs/vim, or want to suggest additions to the above list.

 

The Relative Dangers of Nationalism and Religion

Nationalism and religious partisanship both use the fundamental us vs. them dynamic; they're modern tribalism. Religion, though, has an additional necessary ingredient: faith -- belief without evidence. While nationalism can be based on unreason, religion actually requires it, and unreason is harder to dislodge than entire armies.

"Those who can make you believe absurdities can make you commit atrocities." - Voltaire

(Corporate) Culture is Destiny

"It's a question of which skills are most highly regarded. Everyone has different strengths, and the people who are going to build the next Twitter aren't necessarily going to build the next BigTable as well. A culture that has been very successful with one strength can have difficultly recognizing others."
   — former Googler Paul Buchheit

Indeed, this explains why Google and Apple have almost entirely complementary product lines.  They're all in it to make money, of course.  But Google's culture is built of people who love tools and hacks.  Apple's culture is built of people who follow Steve Jobs; and Jobs is by nature a dictatorial art director, whose art forms are luxury consumer gadgets and corporate imagery.

iOS and Android are fully complementary in philosophy, though it's easy to miss how different they are since their facades are so similar.  Jobs has a vision of the art piece that is the iPhone + iOS, and he deeply and personally resents any compromise of that vision.  Did you see the press conference in which he gave users free bumpers for their iPhone 4s?  He was not contrite, trying to win back the unhappy customers and repair Apple's image.  He was *pissed* -- not at his antenna engineers, mind you, but at the *users* and the *press* who forced him to put a rubber bumper around his magic sculpture.  And he makes no allowance for alternative implementations of either his hardware or his software.  They are what he sells you, and they do what he sells you, and if that's not good enough for you, he'll happily tell you to fuck off.  Sure you can have apps in your phone, but their icons will be THIS big, they won't touch each others' data, and if he thinks his (his!) store is worse off with your app in it, well, again, he will happily tell you to fuck off.  The only magnets on his refrigerator are the ones HE likes.  (One camera app got booted for using the side buttons to take pictures.  How dare they?!  The side buttons are for volume!)

In contrast, every aspect of Android is as open as the carriers and a decent security model allow... Not because Google is altruistic, but because Google is populated with people who like that kind of thing and believe in the collective mind, the unplanned and unscheduled group accidents that make magic.  And fart apps.  But when you're managing a Google-like zeitgeist and try to squeeze a consumer appliance out of it, things can go horribly, horribly wrong.  Case in point: Google TV and its remote from hell, the worst example of The Inmates Running The Asylum I've seen in a decade.   "More buttons?  Sure!  We can just crowd in closer to the primary wheel control ...."  "Works for me!"  So sad.  But what else can one expect from an organization that implements customer support in Python?

Granted, Mac OS doesn't really fit the Apple paradigm as well as iOS -- yet.  Sure, Steve couldn't pass up the price tag of $0 on FreeBSD, so Mac OS X is a good OS even by engineering standards - solidly designed (if not well-patched), and fairly open.  But it's way too open for Steve to be happy.  Mac OS Lion will have an app store.  The writing is on the wall for Mac OS X -- and recall that in the Biblical story that gives rise to that phrase, the writer was God, and his VPs knew better than to argue.

Google is now suffering from the early stages of Microsoft Schozophrenia - the tendency of groups and divisions without strong connections to central leadership to go off and do things on their own which end up being unsuitable for corporate adoption.  They end up either poorly integrated and dying of neglect, or killed outright.  The examples from Microsoft are legion, of course.  As for Google, Wave is just the most obvious example, but I understand that many cool things inside Google just don't make it.  For example, there was a real estate listing plugin for Maps.  It worked, it was very effective; but for whatever reason it didn't attract users, and it was simply allowed to die.  No plan; no vision; just stuff that happens.  It's more like an open source community, for good or ill, than anything to be found in Cupertino.

The Devolution of Robert Wright

I found Nonzero - The Logic of Human Destiny a revelation. I was entirely persuaded of its basic premise: that zones of more non-zero-sum interaction will consistently outcompete their smaller and closer-to-zero-sum rivals. In the book's final section, though, Wright sadly devolved into wooly-headed anthropocentrism; for the most part I look at that last section as his Highlander II (which, incidentally, was never made).

The Moral Animal was good, enlightening in parts, but not tightly edited. The structure grated as well: though I enjoyed learning about Darwin, you can't really use one man as an example of evolutionary principles for an entire book without sounding like Johnny One-Note.

Now we have The Evolution of God -- a timely topic, to be sure. But Wright, interviewed on the Colbert show, has not given up his anthropocentrism. He seems to see evidence that the universe has a plan for us. It's as if he took the only truly bad section of his first book and said to himself, "That really needs more attention. I'll write a whole book about it." (As bad as Clone Wars was, imagine an entire movie about Jar-Jar.)

I'll look at the book ... but I'll wait for paperback, read a chapter, and be ready to walk away.

The Hard-Wiring of Respect ... and Theft

Tim O'Reilly recently observed the spontaneous arisal in Twitter of the idioms of "retweet" or "RT", i.e. tagged repost, and "via", indicating from whom the poster got his information.  Human social wiring provides a strong motivation not to be seen as a status thief, so it actually would have astounded me if somehow using the medium of short text messages was able to defeat the normal human urge to give attribution.

However, there is also another relevant human urge: To steal when one can get away with it.

I do wonder how may retweets and vias are not tagged.  And I also wonder what would happen to those rates once a search service to detect plaigarism were made public.  A determined thief would then, presumably, start rewording their plagiarism so that it was harder to detect.  In the words of the great Tom Lehrer:

Plagiarize!  Remember why the good Lord made your eyes!
Let  no one else's work evade your eyes, so don't shade your eyes, but
Plagiarize! Plagiarize! Plagiarize!
... but be sure always to call it, please, "research."

And yes, I'm sure I'd be caught if I tried to take credit for that.