Perl's Prospects Are Brighter Than Ever

Aug 26, 1998 by Jon Udell

Jon Udell on the Perl Conference

I’m back from the second Perl conference. Over on my home page, listed below, you can find the paper on distributed HTTP that was the basis of the talk I gave there, and also the ZIP file containing the software described in the paper.

The Perl community is really rocking these days. The long-awaited merge of Win32 Perl and Unix Perl is finally a reality – this is a huge improvement because now Win32 Perl programmers can build and use the same extension modules that have long been available to Unix Perl programmers.

The conference was chock full of people whose use of Perl is a far cry from what many people commonly think of. These folks aren’t just writing little CGI scripts. They’re tying together huge far-flung organizations – investment firms, HMOs, universities, you name it – using Perl as the glue for large-scale component aggregation. For example, PerLDAP, a Perl module that manipulates the LDAP API, is now widely used to populate and query the directories that hold together these mega-networks.

A fellow named Gebran Krikor, who works for Digex, demonstrated the most comprehensive network management software I’ve ever seen – graphical displays, massive data reduction and correlation, the whole nine yards – all implemented using Perl (plus, of course, graphics and database components). This is as mission-critical as it gets – Digex relies on his app to keep their circuits flowing, warn of impending problems, and diagnose trouble. Why did he build it, rather than use something commercial, like what Cisco provides? Not because they couldn’t afford the Cisco stuff, but because it just wasn’t good enough.

Tim Bray, who was a founder of OpenText and is now an independent document- and knowledge-management consultant, gave a great presentation on XML, and the new Perl XML::Parser kit. To an audience of Linux faithful, he showed how lame it is that every piece of Linux – /etc/password, httpd.conf, smb.conf, inetd.conf, lilo.conf, and on and on – invents its own weird and different structured-text format. Quite clearly XML can and should unify all this junk so nobody has to waste time inventing syntax, then inventing parsing logic to parse that syntax.

There have been times when I wondered whether Perl would remain relevant. But in the last 12 months it has progressed farther and faster than I thought it would, and now I would say its prospects are brighter than ever. With object-orientation, threads, a common Unix/Win32 foundation, Win32-specific extensions (OLE, COM, registry, event log, etc), and now Unicode/XML, plus a huge base of knowledgeable users – there were over 1000 people at the conference – Perl is on target to retain its title as the Swiss army knife of the Internet.

There was an open-source theme to the conference as well. Tim Howes, from Netscape, gave a talk in which one point struck me particularly hard. Someone asked how Netscape folds in the check-ins from the mozilla.org source tree into the “real” version of Communicator 5. Howes pointed out that there is no such “fold-in” process. Developers inside Netscape are using the same source-control and build tools, and follow the same check-out, check-in procedures, as do developers outside Netscape. This absolutely floored me, for some reason. What an amazingly bold experiment!

Jon Udell http://udell.roninhouse.com/