And here are some more pictures taken by Katrien earlier on in the project. You can see Nicholas (the print head designer) and James (2-axis, software). Oh, and Richard (support). We managed to do this in a three weeks and as we talk James is demoing this at FooCamp2 (assuming he managed to assemble it again). The best bit about this is having the meeting room full of Lego...
I disagree. It's fully possible for interpreted code to go faster than natively compiled code as you have far more information at runtime than at compile time. You can specialise for the currently-common case and do lots more clever JIT tricks.
Recently some JIT improvements have made interpreted Parrot faster than GCC-compiled code for a simple primes benchmark. See, it's entirely possible;-)
YAPC is an umbrella organisation which helps organise cheap, community Perl conferences. This year we have conferences in locations such as Paris, Israel, Boca Raton and Ottowa, and releated workshops in Bonn and Copenhagen. The namespace collision is intentional.
The Gentoo ebuilds come with a MD5 sum of the package. When the package is downloaded, portage checks that the package's MD5 is the same. If it had downloaded the trojaned package, it would have a different MD5 sum, so portage would have complained and told you that something is wrong. Note that the MD5 sum is kept on the Gentoo mirrors, not the package's mirror.
For more information on Parrot, which will be the Perl 6 virtual machine, and which is register-based, you may want to check out
http://www.parrotcode.org/.
We might as well get the terminology right. Perl 5 compiles the source code at compile time (at sometimes at runtime for eval STRING) into opcodes for the terribly-undocumented and very-well-hidden Perl 5 virtual machine. You just don't notice it 'cos it compiles it and runs it in one step.
Perl 6 (not finished yet) aims to seperate the virtual machine (http://www.parrotcode.org/) from rest of the language.
Oh my god, you're right. Perl5 has a horrible syntax, a really slow runtime, and had lots of really bad design decisions as it evolved. You know what we should do? We should start again from scratch design a new language and a really fast virtual machine. Oh wait, we already did...
It's not hard to write good CGI programs in Perl if you know a little about Perl, and this book would help you do that. Let's not get into language flames. This website runs on Perl and slashcode scales pretty well indeed.
For a bunch of secure well-written CGI programs for specific tasks, you can't beat nms. Hope this helps.
I disagree. YAPC::Europe has been planning so far ahead because people have been complaining that they needed this much time to convince their bosses to sponsor them to go. You don't have to go to the beer events if you don't want to. It just happened to be a nice tie in as most of the Perl people I know are quite happy to take part in the Oktoberfest. Accomodation and flights are cheap at the moment (the yapc::europe crew have even organised group accomodation bookings, check the website) but as is often the case in Europe, yes, flights will get more expensive towards the event.
"The Science of Perl" is a great topic which hasn't been explored already. Granted, there will be Perl 6 talks (hey, it looks like Larry will be there!), but it is nice to make a change.
Slashdot Mirror
god speaks thus: acme is not evil, for he likes the One True Colour: orange ;-)
And here are some more pictures taken by Katrien earlier on in the project. You can see Nicholas (the print head designer) and James (2-axis, software). Oh, and Richard (support). We managed to do this in a three weeks and as we talk James is demoing this at FooCamp2 (assuming he managed to assemble it again). The best bit about this is having the meeting room full of Lego...
I disagree. It's fully possible for interpreted code to go faster than natively compiled code as you have far more information at runtime than at compile time. You can specialise for the currently-common case and do lots more clever JIT tricks.
;-)
Recently some JIT improvements have made interpreted Parrot faster than GCC-compiled code for a simple primes benchmark. See, it's entirely possible
YAPC is an umbrella organisation which helps organise cheap, community Perl conferences. This year we have conferences in locations such as Paris, Israel, Boca Raton and Ottowa, and releated workshops in Bonn and Copenhagen. The namespace collision is intentional.
The Gentoo ebuilds come with a MD5 sum of the package. When the package is downloaded, portage checks that the package's MD5 is the same. If it had downloaded the trojaned package, it would have a different MD5 sum, so portage would have complained and told you that something is wrong. Note that the MD5 sum is kept on the Gentoo mirrors, not the package's mirror.
For more information on Parrot, which will be the Perl 6 virtual machine, and which is register-based, you may want to check out http://www.parrotcode.org/.
We might as well get the terminology right. Perl 5 compiles the source code at compile time (at sometimes at runtime for eval STRING) into opcodes for the terribly-undocumented and very-well-hidden Perl 5 virtual machine. You just don't notice it 'cos it compiles it and runs it in one step.
Perl 6 (not finished yet) aims to seperate the virtual machine (http://www.parrotcode.org/) from rest of the language.
Hopefully this clears things up.
Oh my god, you're right. Perl5 has a horrible syntax, a really slow runtime, and had lots of really bad design decisions as it evolved. You know what we should do? We should start again from scratch design a new language and a really fast virtual machine. Oh wait, we already did...
For a bunch of secure well-written CGI programs for specific tasks, you can't beat nms. Hope this helps.
"The Science of Perl" is a great topic which hasn't been explored already. Granted, there will be Perl 6 talks (hey, it looks like Larry will be there!), but it is nice to make a change.