www.digitalmars.com         C & C++   DMDScript  

D - Paradigms and Aspects (was Eon Language)

reply Mark Evans <Mark_member pathlink.com> writes:
A final comment on 'paradigms' - by the time they seep into the mainstream they
are often dumbed-down and then over-hyped.  That happened with objects and is
now happening with aspects.  Read this from Peter Van Roy:

http://www.mozart-oz.org/lists/oz-hackers/0201.html

"Dammit, those AOP [Aspect-Oriented Programming] guys are masters of media hype.
With epsilon work they get their own special issue of CACM. We have actually
SOLVED some of the problems that they are still cracking their heads over. The
hordes of AOP groupies have a right to know this."

So my point is this:  Working programmers are much better off doing their own
background study than trusting mainstream wisdom du jour.

-Mark
Aug 28 2003
next sibling parent reply "Daniel Yokomiso" <daniel_yokomiso yahoo.com.br> writes:
"Mark Evans" <Mark_member pathlink.com> escreveu na mensagem
news:bilo15$24va$1 digitaldaemon.com...

 A final comment on 'paradigms' - by the time they seep into the mainstream


they

 are often dumbed-down and then over-hyped.  That happened with objects and


is

 now happening with aspects.  Read this from Peter Van Roy:

 http://www.mozart-oz.org/lists/oz-hackers/0201.html

 "Dammit, those AOP [Aspect-Oriented Programming] guys are masters of media


hype.

 With epsilon work they get their own special issue of CACM. We have


actually

 SOLVED some of the problems that they are still cracking their heads over.


The

 hordes of AOP groupies have a right to know this."

 So my point is this:  Working programmers are much better off doing their


own

 background study than trusting mainstream wisdom du jour.

 -Mark


<rant>
There's a recent thread on LtU where I talk about this subject:
http://lambda.weblogs.com/discuss/msgReader$8192?mode=topic

IMO it's the language's community fault. If the Mozart/Oz community solve
these Aspect problems (I'm not saying they didn't, just trying to make a
point) they should publish articles like: "Aspects make your head hurt? Take
a pill of good Oz medicine!" talking about how to use Mozart/Oz instead of
AspectJ and such. What they do instead? Write a enormous compendium of
everything you wanted to know about Oz but was too afraid to ask, several
academic papers and some library tutorials. They should (someone in the Oz
community) write a book like: "How to learn Oz in 24 hours" or "Mozart/Oz
for dummies" and some like "Webservices in Mozart/Oz", "Mastering XML in
Mozart/Oz". These kind of books are easy to write, can be written by several
people, one per chapter (Wrox published lots of these kind of books) and
they'll make Mozart/Oz accessible and atractive to "the masses". Until
there's a "MozartWorld" site publishing practical articles with
code-examples, it's their fault that the other guys get "their own special
issue of CACM". They aren't "masters of media hype" they just know how to
"dumb-down" their work and talk in the language common programmers talk.
</rant>
"Mark Evans" <Mark_member pathlink.com> escreveu na mensagem
news:bilo15$24va$1 digitaldaemon.com...

 A final comment on 'paradigms' - by the time they seep into the mainstream


they

 are often dumbed-down and then over-hyped.  That happened with objects and


is

 now happening with aspects.  Read this from Peter Van Roy:

 http://www.mozart-oz.org/lists/oz-hackers/0201.html

 "Dammit, those AOP [Aspect-Oriented Programming] guys are masters of media


hype.

 With epsilon work they get their own special issue of CACM. We have


actually

 SOLVED some of the problems that they are still cracking their heads over.


The

 hordes of AOP groupies have a right to know this."

 So my point is this:  Working programmers are much better off doing their


own

 background study than trusting mainstream wisdom du jour.

 -Mark


<rant>
There's a recent thread on LtU where I talk about this subject:
http://lambda.weblogs.com/discuss/msgReader$8192?mode=topic

IMO it's the language's community fault. If the Mozart/Oz community solve
these Aspect problems (I'm not saying they didn't, just trying to make a
point) they should publish articles like: "Aspects make your head hurt? Take
a pill of good Oz medicine!" talking about how to use Mozart/Oz instead of
AspectJ and such. What they do instead? Write a enormous compendium of
everything you wanted to know about Oz but was too afraid to ask, several
academic papers and some library tutorials. They should (someone in the Oz
community) write a book like: "How to learn Oz in 24 hours" or "Mozart/Oz
for dummies" and some like "Webservices in Mozart/Oz", "Mastering XML in
Mozart/Oz". These kind of books are easy to write, can be written by several
people, one per chapter (Wrox published lots of these kind of books) and
they'll make Mozart/Oz accessible and atractive to "the masses". Until
there's a "MozartWorld" site publishing practical articles with
code-examples, it's their fault that the other guys get "their own special
issue of CACM". They aren't "masters of media hype" they just know how to
"dumb-down" their work and talk in the language common programmers talk.
</rant>
Aug 29 2003
parent Mark Evans <Mark_member pathlink.com> writes:
Daniel,

Give them credit for a decade of research and a thousand-page book donated for
free online.  Maybe after that goes to press, they'll have time for this
marketing stuff. After a well-deserved vacation, that is.

Heaven help us if our design decisions hinge on who has better marketing...

-Mark
Aug 29 2003
prev sibling parent reply "Jeroen van Bemmel" <someone somewhere.com> writes:
Well, those Mozart/OS guys (never heard of them before) seem to be making a
common mistake: they focus on the means (programming language) rather than
the goal (problem). While it might be great to have a language that supports
everything and everybody, if it does not help people to solve an actual
problem it will not be used. Same is true when nobody knows about it, which
also seems to be the case here

Both Mozart/OS and AOP people think they've come up with the silver bullet
that will put an end to all programming problems. Everybody should know that
such a thing simply does not exist. The best you can do is offer a tool that
can play a small part in solving a problem you have, but partly and with
limitations. The more generic such a tool becomes, the less useful it
becomes too.

The good old KISS paradigm still works fine for me :)

PS Walter if you're reading this, it's also true for D
Aug 30 2003
next sibling parent reply Mark Evans <Mark_member pathlink.com> writes:
Jeroen-

That's just about completely backward - like a photographic negative.  Where do
you get such a picture.  That summary misreads them on every count.  Somehow I
feel like I'm in 1984's Ministry of Truth around here...black is white, up is
down...

1. Mozart-Oz says that there is NO silver bullet, which is exactly why they
advocate multiparadigm languages.  2. Their whole semantic model is based on
KISS - "what are the minimum concepts required to support paradigm X."  3. They
solve several practical problems (like concurrency) which cause other languages
to fall apart miserably.

My own point was that programmers must shoulder the responsibility of finding
the right tools for their job.  If we just accept the latest programming fads,
then we're not the best we can be.  We are like journalists who type up
government press releases instead of doing our own investigative footwork.  No
one but the government benefits from that.

Part of the reason I contribute here is to help with the footwork.  Maybe now's
the time for a good vacation though.

Mark
Aug 30 2003
next sibling parent "Peter Hercek" <vvp no.post.spam.sk> writes:
"Mark Evans" <Mark_member pathlink.com> wrote in message
news:bircmm$1jl8$1 digitaldaemon.com...
[cut]

 Part of the reason I contribute here is to help with the footwork.  Maybe now's
 the time for a good vacation though.


I like to read your posts, and Daniel's. Especially when you talk to each other
:)
Aug 30 2003
prev sibling parent reply "Jeroen van Bemmel" <someone somewhere.com> writes:
Sorry Mark, I guess I was jumping to conclusions a bit here. I did not know
Mozart-Os and I based my posting on the e-mail thread alone.

I just get so tired of people claiming to have solved the problems of the
world. AOP is very active in this, and "we have solved concurrency" gave me
the impression that Mozart-Os was doing the same. It's just not that simple,
and I would like to see a bit more modesty and perspective.

Of course, modesty does not sell.

About KISS - it does seem like a paradox to me: they try to keep things
simple so they come up with a language that supports multiple paradigms.
Wouldn't 'simple' imply sticking to just one paradigm? I'll have to study
their material a bit more.

As for each programmer taking responsibility to find the right tools for the
job, I totally agree. It means that people need to see things in
perspective, and see through the marketing hypes. My critizising AOP was
aimed at that


"Mark Evans" <Mark_member pathlink.com> wrote in message
news:bircmm$1jl8$1 digitaldaemon.com...

 Jeroen-

 That's just about completely backward - like a photographic negative.


Where do

 you get such a picture.  That summary misreads them on every count.


Somehow I

 feel like I'm in 1984's Ministry of Truth around here...black is white, up


is

 down...

 1. Mozart-Oz says that there is NO silver bullet, which is exactly why


they

 advocate multiparadigm languages.  2. Their whole semantic model is based


on

 KISS - "what are the minimum concepts required to support paradigm X."  3.


They

 solve several practical problems (like concurrency) which cause other


languages

 to fall apart miserably.

 My own point was that programmers must shoulder the responsibility of


finding

 the right tools for their job.  If we just accept the latest programming


fads,

 then we're not the best we can be.  We are like journalists who type up
 government press releases instead of doing our own investigative footwork.


No

 one but the government benefits from that.

 Part of the reason I contribute here is to help with the footwork.  Maybe


now's

 the time for a good vacation though.

 Mark
Aug 31 2003
parent Antti =?iso-8859-1?Q?Syk=E4ri?= <jsykari gamma.hut.fi> writes:
In article <bisd54$353$1 digitaldaemon.com>, Jeroen van Bemmel wrote:

 Sorry Mark, I guess I was jumping to conclusions a bit here. I did not know
 Mozart-Os and I based my posting on the e-mail thread alone.
 
 I just get so tired of people claiming to have solved the problems of the
 world. AOP is very active in this, and "we have solved concurrency" gave me
 the impression that Mozart-Os was doing the same. It's just not that simple,
 and I would like to see a bit more modesty and perspective.
 
 Of course, modesty does not sell.


Consider the possibility that they actually *have* something
revolutionary in their hands and they know it; I would, for one, make
one hell of a noise in that situation.

I don't know what I'm talking about, as of yet -- but I'm planning to
find out. A good place to start, I suppose, would be their book,

http://www.info.ucl.ac.be/people/PVR/booksingle.pdf

Chapter 4: Declarative Concurrency
Chapter 5: Message-Passing Concurrency
Chapter 8: Shared-State Concurrency

Probably the chapters 1 and 2 are worth reading too, particularly for
getting familiar with the syntax (section 2.3).

-Antti
Aug 31 2003
prev sibling parent "Walter" <walter digitalmars.com> writes:
"Jeroen van Bemmel" <someone somewhere.com> wrote in message
news:biqoac$n5l$1 digitaldaemon.com...

 The more generic such a tool becomes, the less useful it
 becomes too.
 The good old KISS paradigm still works fine for me :)
 PS Walter if you're reading this, it's also true for D


I know. I've been accused many times of putting too many features into D,
instead of making it more generic.
Sep 01 2003