| |||||||||||||
|
| |||||||||||||
Self Hosting Guaranteed Long DevelopmentSelf Hosting Guaranteed Long DevelopmentPosted Jan 1, 2008 21:26 UTC (Tue) by chromatic (guest, #26207)In reply to: Self Hosting Guaranteed Long Development by roblucid Parent article: A Perl 6 status update
Perl 6 is a cleanup of perceived (and keenly felt) mistakes in Perl 5 and most other mainstream programming languages. I'm not sure you understand how deeply those mistakes run. If I may channel Larry perhaps badly for a moment, consider if you had the opportunity to reshape the practical world of programming languages for at least the next 20 years. Would you limit that to minor syntax cleanups, a reorganization of the standard library, and perhaps requiring parentheses around the arguments to
(Log in to post comments)
Self Hosting Guaranteed Long Development Posted Jan 1, 2008 22:44 UTC (Tue) by sbergman27 (subscriber, #10767) [Link] """ Perl 6 is a cleanup of perceived (and keenly felt) mistakes in Perl 5 and most other mainstream programming languages. I'm not sure you understand how deeply those mistakes run. """ Perl 6 is a re-purposing of the language. Once upon a time, Perl was a better Unix shell script and was damned good at that. Then the WWW and CGI came along and the choice was between C, shell script, and Perl. C and shell script were simply too hideous to contemplate in that context, and so Perl won out. Based upon its popularity in cgi, people started trying to write real programs in Perl. The rest is history. There is already a Perl 6. It's called Ruby, and is a very nice language. Not my language of choice. But a very nice language, none the less.
Self Hosting Guaranteed Long Development Posted Jan 2, 2008 0:19 UTC (Wed) by jordanb (subscriber, #45668) [Link] By "reshaping the practical world of programming languages," I assume you mean taking your incredibly popular language, that had become the de-facto scripting language for Unix, and announce to everyone working on it that their code is going to be obsolete when the next version comes out, zapping all the energy out of most of the projects using your language, then taking over EIGHT YEARS to produce that next version, ensuring that all the little languages like Python and PHP that used to tremble in your shadow got all the forward momentum. I admit Wall has completely reshaped the practical world of programming languages, he's made one in which Perl isn't relevant in anymore.
Self Hosting Guaranteed Long Development Posted Jan 2, 2008 2:27 UTC (Wed) by dskoll (subscriber, #1630) [Link] If I may channel Larry perhaps badly for a moment, consider if you had the opportunity to reshape the practical world of programming languages for at least the next 20 years. I assume this is the "hubris" part of "laziness, impatience, hubris." It assumes that people will still care about Perl when (if) Perl 6 is finally released. Major earth-shattering rewrites of software are rarely successful. The only one I can think of that was was the Netscape 4 to Mozilla transition and even that was hugely painful. By the time Perl 6 is released, many Perl 5 programmers may have moved on to non-Perl. At my company, we've always found it hard to find good Perl talent, and the Perl 6 development process is not attracting more people to Perl. Once you start rethinking fundamental assumptions, you might find that the problem is a lot larger and more interesting than you initially thought. That's endemic to software engineering. Unless you say "stop" at some point, you never ship a product. And when you're positioning the new product to be the successor to a wildly popular and useful product, it's very dangerous to let your pie-in-the-sky thinking run wild.
Self Hosting Guaranteed Long Development Posted Jan 2, 2008 2:28 UTC (Wed) by roblucid (subscriber, #48964) [Link] "Perl 6 is a cleanup of perceived (and keenly felt) mistakes in Perl 5 and most other mainstream programming languages. I'm not sure you understand how deeply those mistakes run." May be not, but what was clear to me, as soon as I saw the main proposed ideas for perl6, was that it would not be done any time soon... perl was a practical pragmatic choice, it gave access to the C library which was an advantage over diddling with shell, and with perl5 a large and growing collection of CPAN modules. The language perl4/5 was not exactly small, but tended towards having everything but the kitchen sink included... TMTOWTDI became well known for a reason. Somehow seeing the responses here, leans me towards thinking, that the "rethinking fundamental assumptions" ought to be done with a tiny, research toy language, that could then be fleshed out, for production using the learned lessons. Launching perl6 tainted perl5 with FUD, about the future. This mattered as the size of perl5 script systems was an order of magnitude larger than those attempted with perl4. Depressingly, distro's seem to have moved back to shell scripting, for jobs that I *know* from experience on old hardware would run much faster and be quicker to code with the relatively self-contained perl4. Shell has re-asserted itself through ubiquity, and the limitations become accepted through familiarity. Ny comment, was not meant to be critical of the perl porters, nor Larry Wall; I intended to point out how ambitious the project was, therefore it is not suprising that progress is slow. Part of me, wanted not to rule out "hacking miracles" but my head suspected 2nd System Syndrome, OS-370 (read Brooks) style problems, and a very, very long development process.
|
|
||||||||||||