Programming is hard, let's go scripting..
perl.com
perl.com
This quote, and what little I've sampled of Perl 6, reminds me of a few lines from the movie Wonder Boys:
Grady, you know how in class you're always telling us that writers make choices? [...] And even though your book is really beautiful, I mean, amazingly beautiful, it's... it's at times... it's... very detailed. You know, with the genealogies of everyone's horses, and the dental records, and so on. And... I could be wrong, but it sort of reads in places like you didn't make any choices. At all.
(yes, I have nothing to add)
I would think that by 2007 Larry Wall would have seen DrScheme (now DrRacket). You can't get a better onramp than that.
I think I disagree with you, but have to understand what you're arguing to be sure.
Moving "QUUX formatted records from BAZ mainframe to new Oracle Windows Java 7.0 Lightweight Enterprise Server" involves these things: a source, a sink and the pipe character. Your source and sink may necessarily be in Java--meh, whatever works. But now you have a decoupled framework for doing these kinds of jobs, and you haven't invested any more time or effort than you would have with your one-off, inflexible solution. And you can now stuff anything in that pipeline. And I regularly use racket for exactly this, using protocol buffers as the pipe character. It's not rocket science, it's the Unix Way. People, embrace it and free yourself from the monolithic millstone!
If everything was a service, then everything would be easy, and our tools would look a lot different. But if you can't change the world, you can at least make living in an imperfect world easy, and that's what Perl tries to do.
Everything is a service, if you make it so. A component is a service, offering the functionality of its interface. The 'wc' tool is a service--its interface is its command-line switches. And you're right, when everything is a service, everything is easy (I'd say easier).
http://planet.racket-lang.org/package-source/ryanc/db.plt/1/...
The difficulty right now for anyone from a Perl 5 background who would otherwise jump into Perl 6, is that running your Perl 6 program is really, really slow. I don't know if it's the compile step or execution or something else.
I do not (only) ask to be a smart mouth -- I'm genuinely interested in how it rates on your personal scale of production-readiness...
My latest understanding is that it's still a single-developer project with better performance on many benchmarks than Rakudo but much worse startup time and significant feature overlap. It has monthly releases but makes no promises of stability or usability across releases.
Again, I haven't used it myself.
http://news.ycombinator.com/item?id=1808568
At that time, you weren't so "perl 6 is not ready" camp. What changed?
Anyway, what I'm getting that is 4 years is a pretty short amount of time now that Larry Wall has made me think about it.
There is a really excellent SF novel 'A Deepness in the Sky' by Vernor Vinge.
It's a few thousand years from now, and one really important job is "Programmer-Archeologist". Their job is to grub around in libraries, find an already existing solution, adapt it to present needs. This is important because without quality automation running your spaceship, you could die.
There are hints that some of these libraries are really old: an off-hand comment is made referencing what must be unix epoch.
Given those goals, Perl as a language has a long bright future. Because flexible and extensible languages evolve their way out of time.
Compared to all languages in its class, Perl hasn't broken backwards compatibility since past 25 years. Python had to break it for some very simple syntax changes. Perl has never done it, not even once in past 25 years. That is quite a feat considering so much syntax has been added to the language since then. Perl 5 has also found a way to add more syntax without dealing with the core by modules like Devel::Declare.
So Perl evolution is set for many years. You will see it evolve grow. New neat ideas and many like Moose will find their way into Perl.
Basically what makes it special is its hackable, flexible and extensible. And such languages generally stay around for a long time.
This is not to say that I don't agree that Perl has perhaps the best back-compat of all the languages that still evolve - but these absolute statements have this slightly sect feeling.
That said, I'm pretty sure you're right that backwards Per 5 broke backwards compatibility with Perl 4 (and presumably 4 with 3, etc.)
"I had a hard time characterizing this wonderful book. It explains the fundamentals, but it's not an introductory book like Learning Perl. It covers almost every feature of the Perl 5 language, but it's not a reference book like Programming Perl. It explains common idioms, but it's not a guide to Perl 5 fluency like Effective Perl Programming. It contains many practical suggestions, but it's not a book of tips like Perl Hacks.
I can only describe it as a "textbook". If I had to pick a single book to teach Perl 5, this is the one I'd choose. As I read it, I was reminded of the first time I read K&R (C Programming Language) and how much learning was packed into it. (It's the only college programming text I still have). In a slim 250 pages, Modern Perl obsoletes most of my shelf of Perl 5 books. It's not intended for a complete novice to programming (any more than K&R was), but in the hands of a competent programmer or a diligent student it will teach everything that one needs to know to write Perl 5 well."
[0] http://onyxneon.com/books/modern_perl/
[1] http://www.amazon.com/review/R1EAHI4GHDLQ96/ref=cm_cr_dp_per...
Anyone aware of a similar book for Python? For when I, you know, come back to Python after straying into Perl for a few years?
At the moment Perl 6 specification gives a good evolution map for Perl 5 development. So you need not to have to ever worry about Perl 6.
Perl 5 is a solid language in itself. Perl has been amazingly useful since almost 25 years now. And anybody who denies this is simply trolling. Its useful for almost anything to everything. From scripting, to application development, to prototyping to building reliable applications in any domain.
Perl adapts to nearly every programming paradigm out there. Object oriented, functional, declarative you say it Perl has it. Its amazingly flexible, easy to learn and extensible.
You also get to benefit from the community experience and knowledge accumulated over years. You get the power of CPAN. There are also amazing resources to learn Perl.
Its a tool that you will never stop being in love with.
Those of us who have used Perl and pushed its boundaries to get amazing stuff done over years, even now can't write down words to describe its greatness. We can only tell you how great it is, but when you see it yourself you will know it.
But the position of any new Perl programmer is enviable to say at the least. He stands at the gates of a city which is going to take him to places he has never seen and experiences he has never been through. Most of them are going to change his life for ever.
Some languages get a special distinction that its users just love to program in it. Languages and tools like Lisp, C, Perl, Vi, Emacs etc they just never go away. Because someone somewhere is always discovering some goodness in them and is using it. And that process is timeless. These languages have managed to last for decades sometimes their adoption drops, sometimes it peaks, sometimes there is a constant line of usage. But they don't go away.
If you read the mailing lists during early days of Perl on comp.lang.perl.* and other writings by people like Tom Christiansen. You will see some unique tone of passion and excitement. The golf competitions, the discussions on one liners, the growth of CPAN, the occasional gems posted on Perl Monks. Great books like Programming Perl, Higher Order Perl and Modern Perl Book and Perl's extreme practical approach and emphasis of 'Getting things done'.
I guess for people on Unix platforms who had no other respite but to use C and a mash of other Unix text processing utilities. Or for Java programmers who have to write 10 classes to open a file and do trivial operations.
Perl is almost like giving Ice water to somebody in hell.
But unfortunately, its almost like going backwards. Perl was invented because C + Unix text processing utilities together were not proving to be good for many tasks like Text processing, fast prototyping etc. Now using another C based language which doesn't do that, takes us back to the old problem again.
IMO, Python doesn't solve those problems. Regex support is really bad in Python and file operations are not trivial. There is too much of Java feel to it and it just doesn't have the feel of bash/perl/awk's scripting powers. Scoping is just not right. There doesn't seem to be a sane ways of evolution without breaking backwards compatibility. No equivalent of CPAN. And in many ways white space based parsing is not good for large and nested code blocks. It still doesn't have multiline lambda's. And its too much do-it-one-true-way kind of language, that sort of is a blocker for quick hacking. Because I like to write what I think, and not the other way around. The language is too verbose, and unlike Perl and Emacs there are no exponential productivity gains over time.
But I agree though, that Python does have some other good things. The OO syntax is neat, there are some great frameworks like Twisted and Django. And there is low barrier to entry which makes new bie feel just at home. The tutorials are just too good. And good standard libraries ship with it.
Coming to asking people forcefully shorten their code blocks. It's none of language's business to dictate its style on me and steal my freedom. A language is supposed to behave the way I want. That's its very purpose.
No multiline lambdas, No CPAN equivalent, Seamless interaction with shell like $output = `command` is absent. No first class regular expressions. Lack of ability to provide the command line power Perl gives. No switch statement, No equivalent of Perl's smart match operators, Nothing like Moose, Nothing like Devel::Declare, Parsing power nowhere even near to Python. And I can go on and on..
I guess these are solid enough things(at least for me)
if True:
x = 1
[copy the above line, paste it 999,998 times]
x = 1
print x
Not exactly the same, but the standard library is extensive:
http://docs.python.org/library/index.html
There is pypi:
whose packages can either be installed directly or with pip.
More verbose than Perl? Yes.
Bad?
for x in re.findall(r'\d+', '12 drummers, 11 pipers'):
print x
> file operations are not trivial.
File operations are trivial and well designed. You know, there is much more to file operations than in-place substitute.
> There is too much of Java feel to it
Bullshit.
> and it just doesn't have the feel of bash/perl/awk's scripting powers.
Want it to feel like Perl? Don't bother - use perl.
> Scoping is just not right.
Scoping doesn't require your testimony.
# foo doesn't exist
if blah:
foo = 'bar'
# foo exists if blah
> There doesn't seem to be a sane ways of evolution without breaking backwards compatibility.
When you change the semantics of the language, bozo syntax tricks don't help. Python 3 has semantic changes - syntax tricks aren't going to handle them.
> No equivalent of CPAN.
A large stdlib and PyPi.
> And in many ways white space based parsing is not good for large and nested code blocks.
Spaghetti will look like spaghetti, irrespective of whether block separators are braces or indentation.
> It still doesn't have multiline lambda's.
"still doesn't" sounds like it is something which is planned but not achieved yet. It is not something that Python designers are planning to incorporate in the language. Perl has multiline lambdas but I only used single line lambdas for sort, grep et al - anything larger than that becomes a named function. Unless you have blocks, multiline lambdas aren't used that much.
> And its too much do-it-one-true-way kind of language, that sort of is a blocker for quick hacking.
There can't be an objective explanation to a subjective statement.
> Because I like to write what I think, and not the other way around.
What's the other way round? I like to think what I write? So in Python you think what you write? That makes zero sense.
> The language is too verbose, and unlike Perl and Emacs there are no exponential productivity gains over time.
Complete, utter bullshit. There isn't a single language in the world where the productivity gains are exponential over a long time - there might be some when you are a beginner which gets capped pretty fast.
>>There isn't a single language in the world where the productivity gains are exponential over a long time - there might be some when you are a beginner which gets capped pretty fast.
Sorry to ruin your party. But there are languages where there are exponential gains in productivity with time. Languages like Lisp, Perk, Haskell etc. But I can't convince that to people who can't look beyond 'readability matters' and consider that as the Pinnacle of programming language design.
Coming to regular expressions they are first class citizens in Perl. That is basically what takes them to the next level. Now scripting on Unix has always been so much about Text processing. That is how it has always been. tr, sed, awk and dozen other text processing utilities. And then there is Perl. If you are not getting 'Text processing' thing correct, then you already loose the scripting race with a very huge margin.
Perl added a lot of new semantics to the language too given/when statements, smart matching operators etc. Python broke it for as simple a thing as a print statement. That is the whole point(The language is too brittle to evolve). And I'm sure any new thing in Python is going to go the same way.
And stdlib is no where near CPAN. Although I agree PyPi is doing a good job.
You are right that I must use Perl,And you are also right that Python is an awesome language. There are great frameworks Twisted, Django, NumPy. I've used Twisted extensively. In many Python ways is a great language for web development, for studying algorithms(because its better psuedocode), and due to its frameworks like Twisted - its good for network programming.
But I am sorry to say Python doesn't do well in scripting areas. And it so happens once you start there and go a little deeper you can just go on and on with Perl.
Saying 'python feels like java' or 'exponential gains from switching language' is bullshit, and augmenting it with "calling it bullshit means you don't have an answer" is more bullshit.
> But I can't convince that to people who can't look beyond 'readability matters' and consider that as the Pinnacle of programming language design.
I don't remember mentioning 'readability matters' as being pinnacle of programming. You are responding to arguments you are imagining I am making.
> Sorry to ruin your party. But there are languages where there are exponential gains in productivity with time. Languages like Lisp, Perk, Haskell etc.
Get off your high horse. Lisp, Perl, Haskell aren't some exclusive club only you are allowed into. I have used Racket, Clojure, Python, Ruby etc. and if your gains are exponential accumulated over a period of 1 month between say Python and Racket, be assured you are a very lousy Python programmer. Let alone exponential, a gain of even 2x is quite a stretch. Why don't you put your money where your mouth is and show me your 'exponential gains'?
In place of spewing this nonsensical shit, you could have quoted me an example of "exponential gains".
Out of curiosity, have you actually programmed in Lisp or Haskell? I am not saying you have, I am not saying you haven't; it's just that people who claim "fuck yeah macros - now I am 100x more productive" are mostly people who haven't actually used them and are regurgitating blogosphere bullshit.
This is one lesson I'm not going to forget.
Sheesh. In this thread, there is nothing but platitude in your posts, and you are constantly trying to put down people(everyone else who doesn't agree with you about perl is a troll, saying that what you know about python is not correct is me defending my religion etc etc), and then you whine about people being impolite. Pot, meet kettle.
> This is one lesson I'm not going to forget.
Another lesson would be to actually learn and use Lisp before claiming exponential gains.
It's not impolite to say you are regurgitating blogosphere bullshit when you claim with utmost confidence "I am sorry to ruin your party but there are languages like lisp and haskell which results in exponential gains" when actually you haven't used either one of them.
The gain you mention (though the exponential part is debatable) is not limited to Perl alone[1]. Scripting languages provide a substantial gain in developer productivity freeing them from things like typing, faster debug cycle etc. and that gain is not limited to using Perl alone.
Check Table 1. in the reference I have mentioned, which compares the time it took to develop the same application in Tcl with time it took for development using in C/C++/Java. I think similar gain can be seen when one uses other scripting languages (say Python or Perl).
1. www.stanford.edu/~ouster/cgi-bin/papers/scripting.pdf
If you're looking for more batteries included then I'd suggest Python or perhaps Ruby. Things are always going on in those communities and they've made some great killer apps like Blender (python), Rails (Ruby), Django (Python) and countless more. These languages are just as flexible as Perl, though they have their strengths and weaknesses in different places than Perl does.
My first programming job was in Perl back in the late 90's. Loved it, but I have to say that the OO facilities left something to be desired. I found them fatiguing. It took Java, then Smalltalk for me to really get OO. Perl requires too much mastery of mechanism just to do objects, so it's not the best environment for learning OO.
After learning Moose a couple of years ago, I don't miss the Perl 5 mechanism at all.
A better way to program is with interfaces and composition via delegation, something that Java and C++ support but have no syntax for making it easy. Perl's Moose object system, on the other hand, has easy support for composition via roles and delegation via ... well, delegation :)
And, you have "extends"-style OO available if you think it's the right solution to a particular problem. So really, I think it's a fine way to learn OO. Doing things right is easier, if you decide to do them right.
(We do Java OO right at Google, but it sure involves a lot of typing and @hacky-attributes. And I've never seen good Java OO anywhere else.)
I have used Java and Python extensively now and I like Perl a lot more. All three are quite usable, however.
And I do want to remind readers that Moose is not the Perl object system, it's an object system added on top of the mess that a huge collection of legacy code presumes.
>And anybody who denies this is simply trolling.
Logical fallacy: Appeal to emotion. "My argument cannot be argued with because " is always a logical fallacy. "amazingly useful" is subjective, so of course it can be argued and doing so is not trolling but perfectly reasonable. Don't make claims you don't wish to back up.
>Its useful for almost anything to everything. From scripting, to application development, to prototyping to building reliable applications in any domain.
Logical fallacy: Misleading vividness. While what you say may be true, you've just described a programming language. There is nothing in this sales pitch that is perl specific.
>Perl adapts to nearly every programming paradigm out there.
This one is a better claim. The proper term is "multi-paradigm". Most (all?) Lisp dialects are also multi-paradigm, among various other languages.
>Its amazingly flexible, easy to learn and extensible.
"Amazingly flexible" is subjective. For example, a Smalltalk or Lisp programmer isn't going to be impressed (nor would programmers of Self, IO, etc., etc.).
Easy to learn can be tested to some degree and I believe you'll find the data against you. It would be an interesting to catalog the kinds of questions that get asked about the various languages. In my anecdotal experience, people start having problems with perl right at the beginning with e.g. how variables work.
Extensible is again "Misleading vividness". I can think of more languages that have a similar level of extensibility than those that do not.
>You also get to benefit from the community experience and knowledge accumulated over years.
Misleading vividness.
>You get the power of CPAN.
Overemphasized. Use a JVM-language (e.g. Clojure), for example, and you'll get the power of any open source JVM-language code ever written.
>There are also amazing resources to learn Perl.
Again, if you look at popular languages today there are few that wouldn't tick this box.
>Its a tool that you will never stop being in love with.
This claim is just ridiculous and wrong. I used perl professionally for over half a decade. At first I didn't care that much one way or the other. I was a C++ programmer and this was easier in some cases (but less flexible in other ways that I was used to exploiting). But as the years went on I grew to hate the language more and more until I just couldn't bare it anymore. I hated the backwards defaults (e.g. dynamic variables are fantastic. But as a default? wtf?), how ugly sigils look in code, the embarrassing "OO" syntax at the time and, one of the worst, how the community was constantly trying to prove they were relevant (obviously still the case). You could never mention Perl online or around Perl people without having to listen to the same tired tautologies about how incredibly powerful it is (wtf have you been using to think this?), etc.
So no, it's not a tool you'll never stop being in love with. This is the "Mind projection fallacy". You feel that way, but the vast sea of perl developers who abandoned ship for Python, Ruby, PHP, etc. obviously don't.
>Those of us who have used Perl and pushed its boundaries to get amazing stuff done over years, even now can't write down words to describe its greatness.
More "Misleading vividness". We all have stories like this in our favorite languages, but practitioners of other languages don't seem to feel compelled to post them anywhere and everywhere.
>We can only tell you how great it is, but when you see it yourself you will know it.
Appeal to mysticism. Sure we will.
>But the position of any new Perl programmer is enviable to say at the least.
Hah! If you had just said "Programmer" I would have found this an inspiring sentiment, but Perl programmer? No. I don't envy several months of "wtf" (e.g. the first time you figure out that your code accidentally applies ++ to a var that has a string in it and instead of crashing so you detect this immediately it does something with it).
Speaking of misleading vividness....
'I don't like Perl, so don't speak good about it. And if you do I will troll you.'
Nothing that you can do will make Perl less enjoyable, less powerful or less useful or any thing.
All I could infer from your post was you hate Perl, and you just can't stand somebody talking good about it.
You are making tautological as if they were meaningful and subjective ones as if they were truth [1]. The "infer"ing you did is called "projection". I would have given you similar treatment if you had made a post like you did on a wide range of subjects. But since you have made Perl part of your identity you take anything said about personally. I don't think that's a mentally healthy thing to do, but one would have to ask a professional to know for sure.
[1] Again in this post with "Nothing that you can do will make Perl less enjoyable, less powerful". Enjoyable is subjective and calling Perl "powerful" is a near tautology since the majority of active languages are exactly as powerful if not more so.
You might find it interesting to use it.
Simply put, a function can map arbitrarily, and a script has the mapping and parameters provided!
At least, that's when you get into the scientific computing paradigm, as I understand it.
Learn Perl 5 if you want to learn how a full-featured, complete, and liberating-to-use programming language works.
Check out Perl 6 if you want to see what some of us are building, and if you value ideas as such, even when they haven't all been put together into a finished, polished product.
Is this a reference to consultant_barbie on reddit, or is this meme more widespread?
I'm sure consultant_barbie is referencing the same.
In my opinion, none of the language will solve the problems with programming or bad programs. The only thing which could help programmers is the way they think of the solution and how they approach to it. Thinking about long term solution for a given problem may help.
People are much harder to deal with. Much more dynamic. I'm confident the computer would not have downvoted me :)
BTW you apparently have a gazillion repo's in github but no blog ... writing is much harder than programming IMHO.
Hey its just my opinion that engineering is an easier problem than other business activities. Programming is more predictable. You can fix bad programmers.
Just a curious question... was writing the book harder than programming?