Weblogs Forum - A (Brief) History of Object-Functional Programming

Articles |
News |
Weblogs |
Books |
Forums

Artima Forums | Articles | Weblogs | Java Answers | News

Sponsored Link •

Weblogs Forum
A (Brief) History of Object-Functional Programming

59 replies on 4 pages. Most recent reply: Feb 15, 2010 12:02 PM by robert young

Welcome Guest
Sign In

Back to Topic List

Reply to this Topic

Search Forum

Threaded View


Previous Topic		Next Topic

Flat View: This topic has 59 replies on 4 pages [ « | 1 2 3 4 | » ]

James Watson

Posts: 2024
Nickname: watson
Registered: Sep, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 8, 2009 8:16 AM

Reply

Advertisement

> We need, at minimum, compilers (not new languages) which
> parallelize linear client code. We need, fully, a new
> hardware architecture which accepts linear code and
> executes it in a parallel fashion. So far, no volunteers.

This isn't going to amount to much. There's not enough information in the code to allow for large increases in speed. In addition, many algorithms that are commonly used cannot be parallelized. Different algorithms are required. This is even true for FP.

Mark Thornton

Posts: 275
Nickname: mthornton
Registered: Oct, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 8, 2009 8:19 AM

Reply

> The compiler's job would be to analyze the work, decide
> where the data dependencies are, and generate appropriate
> code.

Unfortunately the code we usually write in traditional languages tends to have too many dependencies.

Kevin Wright

Posts: 38
Nickname: kevwright
Registered: Jul, 2009

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 8, 2009 9:15 AM

Reply

It's coming though, mostly via libraries instead of compilers.

We already have, or are soon getting:
- LINQ
- Actors
- atomic and concurrent primitive operations and collections
- Software Transactional Memory
- Fork-Join
- Parallel Arrays
- OpenCL

David Rozenberg

Posts: 15
Nickname: drozenbe
Registered: Nov, 2009

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 8, 2009 2:38 PM

Reply

To the last comment. It is already over 25 years since parallel computers emerged as the means of solving complex tasks that can be split into parallel processes (finite element techniques, for example). There is not much progress in this area since then. We still do not have any decent languages, compilers, debuggers, etc. We still struggle with tasks to determine whether they can be concurrent, etc. It seems that all this is still in the early stages of understanding and development and is more for college professors and post-graduates than for real software practitioners. Same thing seems to be the statements that Scala will eventually take over.

Kevin Wright

Posts: 38
Nickname: kevwright
Registered: Jul, 2009

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 8, 2009 3:02 PM

Reply

> Same thing seems to be the statements that Scala will eventually take over.

I'd love to think that Scala will take over, but doubt this will actually happen.

To quote from: http://www.paulgraham.com/javacover.html

- "Java's designers were consciously designing a product for people not as smart as them."

Scala seems to have the goal of providing the developer with a lot more power, and with it the responsibility to learn how to use that power effectively, so in some ways it's a bit like a Jedi training school...

The truth is that Scala presents a learning curve, but many programmers don't much care for learning outside of university, and simply want to get their job done and go home to see the kids. Fortunately for me, coders of this sort tend not read Artima :)

By integrating so closely with Java, Scala will readily lend itself to writing libraries, system-level code, and in situations where FP is obviously the right approach. This functionality can then be easily consumed by the non-Jedi Java coders that still make up the bulk of the IT industry.

I'm sure that Scala will have a growing place on the JVM, and that it will continue to draw the brightest, most driven and most passionate developers away from Java. But as a replacement? That would likely only happen if it was taught more than Java in colleges and universities.

Jeremy Huiskamp

Posts: 3
Nickname: jeremyh
Registered: Dec, 2008

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 8, 2009 8:16 PM

Reply

> Functional Programming has had a shot at mainstream
> acceptance many times, and it has failed every time.

Except for that one time, when they made javascript. I hear that got fairly popular.

Achilleas Margaritis

Posts: 674
Nickname: achilleas
Registered: Feb, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 4:03 AM

Reply

Doesn't Haskell cover all the parallelism issues, plus it provides the very best FP out there?

Is there something more that it is required than Haskell?

Kevin Wright

Posts: 38
Nickname: kevwright
Registered: Jul, 2009

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 5:46 AM

Reply

> Doesn't Haskell cover all the parallelism issues, plus it
> provides the very best FP out there?
>
> Is there something more that it is required than Haskell?

Haskell is mind-bogglingly well architected, but it won't run on existing virtual machines, or make use of OSGi for runtime management, or be amenable to existing techniques that the operations team have already learned for dealing with Java deployments.

The barrier is in the ecosystem, and the steep steep learning curve, and the support infrastructure, and the available optimised database connectors, and vendor support. Everything that's needed for a large-scale enterprise deployment basically. The actual language syntax is a minor consideration.

Scala can be deployed and managed as though it was Java (after all, it's just classfiles and bytecode by the time it hits a server).

Compatibility is considered paramount. Although this has caused some challenges in designing the language, it must surely also be the greatest strength in terms of encouraging adoption.

James Watson

Posts: 2024
Nickname: watson
Registered: Sep, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 6:13 AM

Reply

> To quote from: http://www.paulgraham.com/javacover.html
>
> - "Java's designers were consciously designing a product
> for people not as smart as them."
>
> Scala seems to have the goal of providing the developer
> with a lot more power, and with it the responsibility to
> learn how to use that power effectively, so in some ways
> it's a bit like a Jedi training school...

I think the idea that Java is for 'not as smart' people is really wrong-headed. I would rephrase it this way: Java's designers were consciously designing a product for people with different priorities than them.

Java is far from perfect, even for the target audience but the above idea was a stroke of genius. The main reason so much software is unusable garbage is that the designers don't think about what the users need. They dismiss everyone who doesn't think like them as stupid. If they think of the users at all, they treat them like children.

I agree with you in general. Scala addresses some major flaws in Java but isn't designed for the same target audience as Java. That doesn't make it bad, it just makes it less than a full replacement for Java.

Achilleas Margaritis

Posts: 674
Nickname: achilleas
Registered: Feb, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 6:39 AM

Reply

> Haskell is mind-bogglingly well architected, but it won't
> run on existing virtual machines,

Why does that matter? Haskell has its own runtime environment.

> or make use of OSGi for runtime management,

I have no idea if such a beast exists for Haskell, but remote installation/uninstallation/start/stop/update of services may not be that difficult; certainly no more difficult than native services, since Haskell code is compiled to native executables.

But why is this important? I don't know anyone that uses OSGi.

> or be amenable to existing techniques
> that the operations team have already learned for dealing
> with Java deployments.

Maybe these techniques are redundant with Haskell.

> The barrier is in the ecosystem,

Haskell, despite what you or I think, has a lot of libraries.

> and the steep steep learning curve

It's actually easier than Java.

> and the support infrastructure,

I think the support infrastructure argument is overestimated for its importance.

>and the available optimised database connectors

I think Haskell can use the native ones. I have no idea if there are connectors in Haskell.

> and vendor support. Everything that's needed for a large-scale
> enterprise deployment basically. The actual language
> syntax is a minor consideration.

You have a point, but the things you mention are not that critical for most businesses. It's not that every business has that big needs.

> Compatibility is considered paramount. Although this has
> caused some challenges in designing the language, it must
> surely also be the greatest strength in terms of
> encouraging adoption.

I think it depends on the task at hand. If you write a module that only communicates with other modules via messages, then you can do it in any language. In this case, the language is more important than the infrastructure.

Kevin Wright

Posts: 38
Nickname: kevwright
Registered: Jul, 2009

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 6:40 AM

Reply

> Scala addresses some major
> flaws in Java but isn't designed for the same target
> audience as Java. That doesn't make it bad, it just makes
> it less than a full replacement for Java.

Oh yes, I agree with you 100% on this!

It's a companion, and a strong and necessary one at that, but I think it would be foolish to push for a full replacement :)

Mark Thornton

Posts: 275
Nickname: mthornton
Registered: Oct, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 6:42 AM

Reply

> > Haskell is mind-bogglingly well architected, but it
> won't
> > run on existing virtual machines,
>
> Why does that matter? Haskell has its own runtime
> environment.

I can easily use Scala with our existing Java code.

Achilleas Margaritis

Posts: 674
Nickname: achilleas
Registered: Feb, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 6:52 AM

Reply

> > > Haskell is mind-bogglingly well architected, but it
> > won't
> > > run on existing virtual machines,
> >
> > Why does that matter? Haskell has its own runtime
> > environment.
>
> I can easily use Scala with our existing Java code.

Depending on case, it may be more profitable to use another language though.

Mark Thornton

Posts: 275
Nickname: mthornton
Registered: Oct, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 6:54 AM

Reply

> > I can easily use Scala with our existing Java code.
>
> Depending on case, it may be more profitable to use
> another language though.

There was over 500,000 lines when it was last counted (not counting 3rd party libraries).

James Watson

Posts: 2024
Nickname: watson
Registered: Sep, 2005

Re: A (Brief) History of Object-Functional Programming

Posted: Dec 9, 2009 7:02 AM

Reply

> > Scala addresses some major
> > flaws in Java but isn't designed for the same target
> > audience as Java. That doesn't make it bad, it just
> makes
> > it less than a full replacement for Java.
>
> Oh yes, I agree with you 100% on this!
>
> It's a companion, and a strong and necessary one at that,
> but I think it would be foolish to push for a full
> replacement :)

I actually think that the idea of one language for all purposes is a very naive idea. In my current role, I can't see us using Scala but in a different context, it would probably be at the top of my list.

Flat View: This topic has 59 replies on 4 pages [ « | 1 2 3 4 | » ]

Previous Topic

Next Topic

Sponsored Links

Copyright © 1996-2019 Artima, Inc. All Rights Reserved. - Privacy Policy - Terms of Use