Log in

No account? Create an account

End of Roads

I was thinking about making an end-of-year push to learn Perl6, and if things went well to start to port all my existing projects (that are in Perl; a few are in other languages) to it. Unfortunately, the Perl6 compiler/interpreter that ships with Fedora is terribly slow and can't run a lot of the sample code (that was written to the Perl6 spec; nobody has a compliant compiler/interpreter for it yet). This is the many-th year that Perl6 has been in this state. I popped onto IRC, and found some terribly defensive people with shaky responses like "we don't believe in a crossing-the-finish-line release cycle". After a bit of calming discourse, eventually they (implementers of the leading implementation) admitted that they thought Perl6 was probably not going to go anywhere and that it was likely a waste of time at this point.


Perl6 does look neat; it has the freewheeling feel of Perl combined with an actual spec, a type system if you want it, and all sorts of neat mathy-structures built in. It's also a language whose implementation has become so distracted that it probably will never be usable. I have a lot of ideas for programming languages for various purposes; Perl has long been a language I liked for expressivity in terms of code structure, but it's never been the perfect language (nor have I actually seen languages experimenting much with the things I'm interested in). I don't want to be stuck with a language that's as dead as Fortran (and there's no way I'm going to Python because FOR SRS THAT SYNTAX IS SHITE). I'm sure I could have a good time designing my own language (based around being able to help the compiler prove things about it that allow for crazy optimisation as well as managed transitions between, say, reentrant code and non-reentrant code), but I don't have the skills needed to effectively write the other end that would use that provability for awesome things (whether better warnings/errors or amazing optimisations).

I want a new general-purpose programming language that has lots of great libraries (from curses to apache, threading to multimedia, and so on). Perl is that now. What are my other options? I suspect Go isn't the language (yet). Objectionable C doesn't have nice development libraries except on NeXTStep (but I don't want to develop on a dead platform) and OSX (great platform technically, bloody awful platform strategy though). Ruby.. well, I'm a bit disenchanted with it. I'm not keen to move *all* of my projects back into stone-age C, although that certainly would work and it'd be a very safe move. I guess there's Java. Sigh. Or I can just stick with Perl5 until something better comes along.

I guess Perl5 isn't obscure yet. It's at least still top-ten.

I am really frustrated at how badly Larry Wall screwed over the Perl community by managing Perl6 this badly.

I'm also deeply frustrated at the direction Redhat is taking Fedora Linux. Redhat's engineers are, outside of the GNOME Foundation, the biggest source of bad ideas for Linux in recent times. I want a distribution that still uses RPMs, that wholeheartedly rejects abandoning X11 for Wayland, that doesn't treat traditional Unix policy as a joke (referring to the let-users-install-anything-in-a-registered-repository-without-sysadmin-choice mess), that don't decide to radically rework the filesystem hierarchy, that rejects GNOME3 for being the awful mess it is, that doesn't have Windows-envy (referring to the abandon-dotfiles-for-a-registry bs), and that doesn't have a dev community where I need to worry about the next dumb idea of the month.

So I'm halfheartedly looking for a new distro (or cousin OS) and a new programming language. I am of course very picky about these things.


Linux Mint looks promising; that's what everyone's supposed to be switching to from Ubuntu, in response to the Unity debacle.

Interesting about Perl6; I had already decided I was not going to upgrade, it being a totally different language from Perl5, but I hadn't imagined it was in such a disastrous state.
I've heard someone speak favorably of Mint, but a coworker who tried it disliked it. I don't remember the reason.