Ulti

Well the "expertise" is mostly just Python thats sort of the value prop. But yeah building an actual AI product ontop I'd be more worried about the early stage nature of Modular rather than the implementation is closed source.

More of a question of /will/ Mojo eventually be entirely open source, chunks of it already are. The intent from Modular is eventually it will be, just not everything all at once and not whilst they're internally doing loads of dev for their own commercial entity. Which seems fair enough to me. Importantly they have open sourced lots of the stdlib which is probably what anyone external would contribute to or want to change anyway? https://www.modular.com/blog/the-next-big-step-in-mojo-open-...

Feel like Mojo is worth a shoutout in this context https://www.modular.com/mojo Solves the issue of having a superset of Python in syntax where "fn" instead of "def" functions are assumed static typed and compilable with Numba style optimisations.

yeah looks like how they distribute too, basically pre installed and configured pg_search pg_datalake and pgvector

Came here to suggest this should be generic, but I'd also do something like pack in `man <command>` into the prompt if you are one shotting. Then it works for "all" commands that have a man page rather than just the commands GPT knows about before its cut off. Even just trying to scrape out `<command> --help` or something would be good too.

There area also a lot of people naive to how BS publishing has become. Plenty of PhD students get sucked in, a common one is being asked to publish your thesis as a book. Learning hubris comes before an embarrassment is a right of passage to the vast majority of people who think anyone really wants to read their thesis in a bound form from a third rate publisher created last year.

No you should strongly reason and believe in your own power of reasoning, but you should never dogmatically hold on to those beliefs in the light of new evidence that contravenes your previous ideas. Essentially all scientists that are even vaguely good follow this ideal. You have to know what is opinion from fact, facts dont change but your opinion about them does. Opinion guides generation of new facts. If you dont have strong opinions then you wont have the dedication to generate new fact. Look at the more zany side of religion, its strong opinion strongly held. They aren't capable of change and generation of new ideas. But if you were forever lacking in an opinion you also wouldn't action on anything and have some kind of existential nihilism all the time.

Media is a few days lag behind academic publication on this one http://www.nature.com/articles/nenergy201732

I think that says more about you than everyone else. I would walk halfway across town for slightly better bbq. Ultimately what is there to life than improving the quality of your experience. The issue is what consitutes better BBQ is completely subjective. So we might start from opposite sides of town and try and get to each others nearby BBQ ribs place. That's just human nature. With respect to "frameworks" some of this tends to be kind of centric to how you program. For the last decade most of my work isn't within any framework because its algorithmic with only a need for string IO and some parsers. Perl 6 for that is the framework.

Not for performance which is what I think rurban was getting at.

If you come from only a Perl 5 background this is the issue. None of that is obvious. It is to someone who's seen OO languages that are a bit functional with good exception handling. In Perl 5 you would probably add to the top of the file:

  use warnings;
  use autodie;

To get similar functionality as the implicit Perl 6 IO error system.

Well if you are wandering that way again it's a bit faster a year later. About 10x across the board would be a good estimate. That's still super slow for some things.

Thanks for your projection of rudeness... Makes me feel extra special given I spent about ten minutes linking all that stuff.

Baliador is meant to be a take on Dancer from Perl 5 which is somewhere between Sinatra and Rails in complexity. In the Perl 5 world the equivalent of Rails would be Catalyst. But even that isn't really as monolithically well integrated.

The point I'm making is Rails wasn't built in a day and to date Perl 6 has had quite a few people want to kickstart a Rails initiative but they always do it with new code and new ideas cloning a previous paradigm. A better plan is to embed yourself in what's there especially the core language. With grammars you could write a complete DSL as a web framework with all of the power of Perl 6. But it would take someone knowing the core language well to make something that's slick. For behind the scenes there is little reason to reimplement HTTP parsing/serving since there are a couple of modules that are tested against every release of the Rakudo compiler already.

A) It's released with a frozen spec and a large test suite for compliance (https://github.com/perl6/roast), with a compiler available that supports a module ecosystem with over 700 modules (http://modules.perl6.org). What's a complete state in your mind? If not this.

B) All of the below is core language, not a use or import anywhere required. Error reporting is also handled for you without autodie and friends.

  Reading from a file:

  #Efficiently line by line for STDIN
  for lines() -> $line {
      $line.say;
  }

  #Get everything in RAM right now
  my $string = "filename".IO.slurp;

  #Lazy list will do IO as you request into the list
  my @lines = "filename".IO.lines;

  #Listing of directories if the path is a directory
  my @directories = "coolstuff".IO.dir if "coolstuff".IO.d;

  Playing with dates:

  #Get a DateTime for right now
  my $date = DateTime.now;

  #Does what it says on the tin
  say "Yippee" if $date.later(:5years).is-leap-year;

It's hard for me to describe how much Perl 6 doesn't need another web framework, but someone to actually use the ones that already exist! There is even already more than one middleware solution Crust and P6W.

https://github.com/ufobat/Bailador

https://github.com/tokuhirom/p6-Crust

https://github.com/skaji/Frinfon

https://github.com/nunorc/p6-Pekyll

https://github.com/retupmoca/P6-Web-RF

https://github.com/drforr/perl6-App-prancer

https://github.com/sergot/BreakDancer

https://gitlab.com/samcns/uzu

https://github.com/supernovus/perl6-web-app-ballet

https://github.com/zostay/P6W

Testing https://github.com/azawawi/perl6-selenium-webdriver

Even desktop style web applications with Electron https://github.com/azawawi/perl6-electron

It's all there just no one engages with any of the development because its not trending on HN every five seconds or has the eye of the tweener set of developers who blog and vlog.

No one talking about Perl 6 compactness means golfed code! Just naturally translating over say a Java program with the OO design intact will be hugely compact. In a way that means the code is written at a succinct conceptual and declarative level. Perl 6 is incredibly declarative, you say what you want, not how to get there. The core features cover nearly all of what you want out of the box too.

Since Christmas 2015 was the "production" release of the language specification. Hence the 6.C version number, the C stands for Christmas. Rakudo around about now implements that feature set in a stable way. The last year has mostly been optimisation work in the runtime and VM. The whole thing is about as stable and production worth as Swift. That is to say features do still change, but not the ones already specced as 6.C

Because people keep asking why it hasn't taken off rather than using it and posting here about how it was awesome.

I'd say most netbooks or tablets replaced it for all of those things now... But you're right that Psion were super innovative. More important than their series 5 hardware was the operating system. EPOC was a 32bit embedded OS long before anyone else was doing them. It was nice to program for and dealt with multiprogram which even the first iPhones didn't think about.

This article shows a gross ignorance of the reality. Academic sciences especially are woefully under funded and anyone coming out now with a PhD has no position to go into especially if you are half decent. Accepting to only teach in a crap university and struggle to have a research career is not a choice for competent people. Instead they leave once they fully understand the insermountable task of obtaining a decent life in the academic world is. Worse is the younger 40-45 yol academics think people now face the same world they did. Whereas there has been a 10x increase in PhD graduates since they went looking for a position. The number of available jobs has not dramatically increased at all. This and the stochastic nature of funding of grant proposals means no one chooses to remain in academia. Apart from the top 1% or the no hopers who think being a stooge for the next 40 years is admirable and they should take one for the team to pursue "science". The reality is research jobs outside of academia are obviously booming with this sudden availability of talented and highly educated work force.

I want to see Lee Sedol win with just a single neurone from his neocortex then...

This isn't a better view though. The number of Watts is not important. Biology is efficient in energy use, so what? If you tax a human brain it doesn't especially consume vast amounts more energy. I don't work up a sweat thinking. What's more interesting is if the computational "power" has parity. Deep Blue wasn't as computationally powerful as a human brain. It was specialised which is why it's not as interesting a victory as AlphaGo which is based on a more general principal.

There are still the design docs (http://design.perl6.org/) which are essentially spec documents. They are what implementors have been using over the years. Don't forget the tests came from Pugs and now Rakudo is using the same test suite, so it's more portable than you perhaps imagine. The tests are the stamp of approval for a version number conformance. If you can pass all the versioned tests then you are that version of the language. I think that's a lot stronger statement than the engineers think they understood a spec document, even if formally specified. What are your thoughts on why that isn't ok or worse than relying on pure docs?

Guess what you can do both at the same time! :D There is already a module which converts an ANTLR4 grammar into a full Perl 6 grammar which then provides a ready made parser for you :) https://github.com/drforr/perl6-ANTLR4

Ruby and Go for the web?? Really? Not JavaScript and PHP (still). Python is old hat too it's all about Julia now :P and R is making a come back in some weird computing renaissance.

Maybe they'll rewrite Hurd in Perl 6...

Once? This is the only time there has been a release of a versioned specification for the language compilers can target. That's what has been in development for 15 years! The specification of the actual language. It's like the difference between JavaScript existing as a language for a while and then one day ECMAScript 1 was defined. Apart from Perl 6 was always about the evolution of the spec with various implementations showing the way.

The first major implementation of Perl 6 back in the early/mid 00s was Pugs written in Haskell where Perl 6 picked up a tonne of functional heritage both in the compiler at the time and the language spec. The more recent Rakudo compiler has a lead architect who often works in C# where the compiler and the language spec picked up and redefined a reactive programming concurrency model with lots of nice new syntax like the react/whenever block.

Rakudo Perl 6 and Pugs and the host of other early implementations have been around a while and released often when they were worked on, monthly in the case of Rakudo with the first publicised release in 2010. Rakudo is also releasing a version of the compiler compliant with the newly frozen and released specification.

Quoting is highly generalised in Perl 6. Comments happen to be more feature rich and dynamic than other languages. They aren't simply ignored by the compiler but completely parsed and used in a variety of ways. The flexibility looks odd because it's meant more for quoted string interpolation, but reuses the same mechanism for comments. Generality of concepts is a common theme throughout Perl 6, if something looks round and spins it should be a wheel.

For example all of the literate style POD6 comments are accessible within a variable as a fully structured tree in the $=pod variable. The 'q' quoting context in the main language allows you to use any matched quote/bracket from unicode as the markers of the start and end of a string or regex, or comments as you've picked up on.

This is super helpful if you want all sorts of weird quotes and brackets inside of a string and still have it interpolated for variables for example.

    say q⧚This string has a "'weirdly quoted bit in'"!⧛;

Will print out:

    This string has a "'weirdly quoted bit in'"!

To see some examples of how extreme you can have these characters be from unicode check out http://xahlee.info/comp/unicode_matching_brackets.html

Perl 6 has some of the best unicode support out there as default behaviour. Even supports all unicode digits as numeric literals. So:

    1 + ൭ == 8

Cool I should give it a go then. perl6.io might be the best plan if we're trying to get HN to take notice and try something out >:3