"Verdict" on Scala

[Ola Bini-3]

If people haven't seen this already:
http://creativekarma.com/ee.php/weblog/comments/my_verdict_on_the_scala_language/

I think he has at least a few good points, especially about documentation.

Cheers

--
 Ola Bini (http://ola-bini.blogspot.com)
 JRuby Core Developer
 Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
 Practical JRuby on Rails (http://apress.com/book/view/9781590598818)

 "Yields falsehood when quined" yields falsehood when quined.

[Jon Pretty]

Ola Bini wrote:
> I think he has at least a few good points, especially about documentation.

What a depressing read...

Jon

--
Jon Pretty | Sygneca Ltd.

[Erik Engbrecht]

I'm really surprised about the comments about the community.

 

On 1/4/08, Jon Pretty <jon.pretty@...> wrote:

Ola Bini wrote:
> I think he has at least a few good points, especially about documentation.

What a depressing read...

Jon

--
Jon Pretty | Sygneca Ltd.

--
http://erikengbrecht.blogspot.com/

[Andrew.Foggin]

Erik Engbrecht wrote:
> I'm really surprised about the comments about the community.
>
>  
The fact that the author's only question posted to the mailing list
(http://www.nabble.com/Iterable.projection-to14503949.html#a14503949)
has gone unanswered may have something to do with that.  (I'm sorry I
wasn't confident enough to attempt an answer)

Personally I cannot be offended by any categorization that is
demonstrably false.

FWIW my experience of the community is the exact opposite, and the
article's other conclusions are in most cases quite different to my
own.  And that's OK.  We aren't all going to see everything the same way :-)

-- Andrew

> On 1/4/08, *Jon Pretty* <jon.pretty@...
> <mailto:jon.pretty@...>> wrote:
>
>     Ola Bini wrote:
>     > I think he has at least a few good points, especially about
>     documentation.
>
>     What a depressing read...
>
>     Jon
>
>     --
>     Jon Pretty | Sygneca Ltd.
>
>
>
>
> --
> http://erikengbrecht.blogspot.com/ 

[Stanislas Klimoff-2]

I don't quite follow his comment about dependency analysis, could someone please point me out what's the difference between javac and scalac?

Thanks

On 1/4/08, Ola Bini <ola.bini@...> wrote:

If people haven't seen this already:
http://creativekarma.com/ee.php/weblog/comments/my_verdict_on_the_scala_language/

I think he has at least a few good points, especially about documentation.

Cheers

--
Ola Bini (http://ola-bini.blogspot.com)
JRuby Core Developer
Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
Practical JRuby on Rails (http://apress.com/book/view/9781590598818 )

"Yields falsehood when quined" yields falsehood when quined.

--
Best wishes,
  Stan Klimoff
  Grid Dynamics Consulting

[Bill Venners]

Hi Jon,

On Jan 4, 2008, at 10:23 AM, Jon Pretty wrote:

> Ola Bini wrote:
>> I think he has at least a few good points, especially about  
>> documentation.
>
> What a depressing read...
>
I think it is good to get some critical feedback. It helps point out  
problems, some of which can be solved. I have long been concerned  
about the documentation. It is a simple thing that can make a big  
difference to people's initial feelings about Scala. The book will  
help, but I would really like to again suggest we try and improve the  
scaladoc documentation. I've been documenting ScalaTest, which will be  
out very soon, and have several suggestions for making the scaladoc  
output more user-friendly. But the main thing is that much of the  
actual documentation in the API docs is missing. That makes Scala much  
harder to learn and also gives a feeling that Scala isn't as cooked as  
it actually is. I wouldn't mind helping document some of that. Perhaps  
the community could help out here. Anyway, this problem can be solved  
fairly easily, and I think we should try and solve it soon. I think it  
should be a high priority.

On the community, I have noticed occasionally some slamming of newbies  
who dare to ask questions that may seem to challenge sacred functional  
cows. But in general I think the list has been very helpful and  
responsive to newbies, including to me as I've been coming up to  
speed. I don't think there's much of a problem here, really.

I'm not sure where the author gets his notion of "properly" in his  
claim that Scala "tends toward being illegible when written  
“properly”." I have found it hard to read other people's code in  
Scala, but I would never consider inscrutable code the proper way to  
write something. Some of that is my own lack of familiarity with  
Scala, which will improve over time, but I think some is just Scala  
itself. I think that questioning Scala's readability is fair and  
healthy, but I don't agree with his notion of proper. In the book I've  
tried to make the examples easy to read, and to try and set a good  
example for people to follow. That's why I keep asking questions to  
the list about what formatting and style is preferred. No matter what,  
though, I'm sure there will be people who write code I find ugly and  
hard to understand in any language. But Scala may make that slightly  
easier than Java does. The only thing I can think of to do here is to  
try and set good examples of style.

Bill
----
Bill Venners
President
Artima, Inc.
http://www.artima.com

[Bill Venners]

Hi Stanislas,

My take was that scalac would know to recompile a Scala class, say,  
that used a Java one that had changed. Say the Scala class called into  
a method that no longer exists. He wants that to be caught at compile  
time without having to recompile every Scala file.

I think this would be a great thing someday, but not urgent. It is  
something I can live with. But I've noticed myself that it is  
cumbersome to combine Java and Scala. In the ScalaTest build I have  
Scala and Java together in both the source and the tests, and I have  
four compiles that have to be done in just the right order. Sometimes  
I have to do a clean build just to make sure everything gets  
recompiled that needs it. I suspect mixed projects like this would be  
common, but we'll have to see how Scala actually gets used. If it is,  
I wonder if it would make sense to let scalac handle all of the  
dependency checking, and let it fire off the Java compiler when  
necessary.

Bill

On Jan 4, 2008, at 11:10 AM, Stanislas Klimoff wrote:

> I don't quite follow his comment about dependency analysis, could  
> someone please point me out what's the difference between javac and  
> scalac?
>
> Thanks
>
> On 1/4/08, Ola Bini <ola.bini@...> wrote: If people haven't  
> seen this already:
> http://creativekarma.com/ee.php/weblog/comments/my_verdict_on_the_scala_language/
>
> I think he has at least a few good points, especially about  
> documentation.
>
> Cheers
>
> --
> Ola Bini (http://ola-bini.blogspot.com)
> JRuby Core Developer
> Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
> Practical JRuby on Rails (http://apress.com/book/view/9781590598818 )
>
> "Yields falsehood when quined" yields falsehood when quined.
>
>
>
>
>
> --
> Best wishes,
>   Stan Klimoff
>   Grid Dynamics Consulting

[Andrew.Foggin]

Stanislas Klimoff wrote:
> I don't quite follow his comment about dependency analysis, could
> someone please point me out what's the difference between javac and
> scalac?
>
I don't think that particular statement is very meaningful.  The point
about not being able to freely intermix Java and Scala code is valid but
hard to solve (see this thread for a discussion:
http://www.nabble.com/-scala--mixing-scala-source-with-java-source-to13129958.html)

BTW I'm pretty sure the paragraph about pattern matching and variable
scoping is incorrect too.

--Andrew

> Thanks
>
> On 1/4/08, *Ola Bini* <ola.bini@... <mailto:ola.bini@...>>
> wrote:
>
>     If people haven't seen this already:
>     http://creativekarma.com/ee.php/weblog/comments/my_verdict_on_the_scala_language/
>
>
>     I think he has at least a few good points, especially about
>     documentation.
>
>     Cheers
>
>     --
>     Ola Bini (http://ola-bini.blogspot.com)
>     JRuby Core Developer
>     Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
>     Practical JRuby on Rails
>     (http://apress.com/book/view/9781590598818
>     <http://apress.com/book/view/9781590598818>)
>
>     "Yields falsehood when quined" yields falsehood when quined.
>
>
>
>
>
> --
> Best wishes,
>   Stan Klimoff
>   Grid Dynamics Consulting

[Amir Michail]

On Jan 4, 2008 10:23 AM, Jon Pretty <jon.pretty@...> wrote:
> Ola Bini wrote:
> > I think he has at least a few good points, especially about documentation.
>
> What a depressing read...
>
> Jon
>

The only thing that matters for me is productivity.  Scala is nice in
that respect.

Sure, things may be different if I were working in a team with
conflicting coding styles.

I do find it strange though that most of the Scala community has a FP
background.  I would have thought Java/Python programmers would be
more interested!

Amir

> --
> Jon Pretty | Sygneca Ltd.
>

[Stanislas Klimoff-2]

Well, my first thought was about managing cyclical dependencies in mixed environment, but he clearly takes on difference between javac and scalac — that amused me.

Stan

On 1/4/08, Andrew Foggin <andy@...> wrote:

Stanislas Klimoff wrote:
> I don't quite follow his comment about dependency analysis, could
> someone please point me out what's the difference between javac and
> scalac?
>
I don't think that particular statement is very meaningful.  The point
about not being able to freely intermix Java and Scala code is valid but
hard to solve (see this thread for a discussion:
http://www.nabble.com/-scala--mixing-scala-source-with-java-source-to13129958.html)

BTW I'm pretty sure the paragraph about pattern matching and variable
scoping is incorrect too.

--Andrew

> Thanks
>
> On 1/4/08, *Ola Bini* <ola.bini@... <mailto:ola.bini@...>>
> wrote:
>
>     If people haven't seen this already:
>     http://creativekarma.com/ee.php/weblog/comments/my_verdict_on_the_scala_language/
>
>
>     I think he has at least a few good points, especially about
>     documentation.
>
>     Cheers
>
>     --
>     Ola Bini (http://ola-bini.blogspot.com)
>     JRuby Core Developer
>     Developer, ThoughtWorks Studios ( http://studios.thoughtworks.com)
>     Practical JRuby on Rails
>     (http://apress.com/book/view/9781590598818
>     <http://apress.com/book/view/9781590598818>)
>
>     "Yields falsehood when quined" yields falsehood when quined.
>
>
>
>
>
> --
> Best wishes,
>   Stan Klimoff
>   Grid Dynamics Consulting

--
Best wishes,
  Stan Klimoff
  Grid Dynamics Consulting

[Stanislas Klimoff-2]

I'm one of those, and I perfectly understand the moans about the documentation quality and the build process :)

Stan

On 1/4/08, Amir Michail <amichail@...> wrote:

I do find it strange though that most of the Scala community has a FP
background.  I would have thought Java/Python programmers would be
more interested!

Amir

> --
> Jon Pretty | Sygneca Ltd.
>

--
Best wishes,
  Stan Klimoff
  Grid Dynamics Consulting

[Randall Schulz]

On Friday 04 January 2008 08:41, Amir Michail wrote:
> ...
>
> The only thing that matters for me is productivity.  Scala is nice in
> that respect.

Individual productivity is highly idiosyncratic, is it not? I'm not
saying you're an outlier or that you're highly typical, but anecdotal
information about productivity is not something that constitutes real
evidence.

Also, I wonder how easily one can reacclimate to one's own code after
not working with it for some time (let alone comprehend code never seen
before). It's a very common experience for programmers across many
languages and with a wide range of experience and skill to find their
own code perplexing after they've moved on to other work and then later
revisit that code. It can take only a matter of a couple of weeks for
this phenomenon to occur.


> Sure, things may be different if I were working in a team with
> conflicting coding styles.
>
> I do find it strange though that most of the Scala community has a FP
> background.  I would have thought Java/Python programmers would be
> more interested!

It is, or at least appears to be, a very foreign language. I'm not
surprised that workaday Java programmers are not choosing it, prefering
something more readily comprehensible (if much weaker on the theory and
principle-of-operation side) such as Groovy.


> Amir


Randall Schulz

[Ola Bini-3]

Bill Venners wrote:

> Hi Stanislas,
>
> My take was that scalac would know to recompile a Scala class, say,
> that used a Java one that had changed. Say the Scala class called into
> a method that no longer exists. He wants that to be caught at compile
> time without having to recompile every Scala file.
>
> I think this would be a great thing someday, but not urgent. It is
> something I can live with. But I've noticed myself that it is
> cumbersome to combine Java and Scala. In the ScalaTest build I have
> Scala and Java together in both the source and the tests, and I have
> four compiles that have to be done in just the right order. Sometimes
> I have to do a clean build just to make sure everything gets
> recompiled that needs it. I suspect mixed projects like this would be
> common, but we'll have to see how Scala actually gets used. If it is,
> I wonder if it would make sense to let scalac handle all of the
> dependency checking, and let it fire off the Java compiler when
> necessary.
>
> Bill

Bill!

I totally agree about this. The project I'm working on to learn Scala
uses almost only Scala code, but generates Java through Antlr. This
necessitates two compile phases, which is annoying in itself. Another
problem is that recompilation doesn't always catch that closures should
be rebuilt, so I end up with strange errors at runtime, because not
everything was cleaned. I find it strange such a thing happens with only
one Scalac phase, though.

A combined Javac and Scalac would be a very nice thing.

--
 Ola Bini (http://ola-bini.blogspot.com)
 JRuby Core Developer
 Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
 Practical JRuby on Rails (http://apress.com/book/view/9781590598818)

 "Yields falsehood when quined" yields falsehood when quined.

[olambo]

Hi Ola

Cant help myself being nosy! Feel free to ignore my post, if the  
answer is top secret : )

Your generating Java through Antlr using Scala. Why (must be more than  
learning Scala)?

Oliver

On 05/01/2008, at 6:26 AM, Ola Bini wrote:

> Bill Venners wrote:
>> Hi Stanislas,
>>
>> My take was that scalac would know to recompile a Scala class, say,  
>> that used a Java one that had changed. Say the Scala class called  
>> into a method that no longer exists. He wants that to be caught at  
>> compile time without having to recompile every Scala file.
>>
>> I think this would be a great thing someday, but not urgent. It is  
>> something I can live with. But I've noticed myself that it is  
>> cumbersome to combine Java and Scala. In the ScalaTest build I have  
>> Scala and Java together in both the source and the tests, and I  
>> have four compiles that have to be done in just the right order.  
>> Sometimes I have to do a clean build just to make sure everything  
>> gets recompiled that needs it. I suspect mixed projects like this  
>> would be common, but we'll have to see how Scala actually gets  
>> used. If it is, I wonder if it would make sense to let scalac  
>> handle all of the dependency checking, and let it fire off the Java  
>> compiler when necessary.
>>
>> Bill
> Bill!
>
> I totally agree about this. The project I'm working on to learn  
> Scala uses almost only Scala code, but generates Java through Antlr.  
> This necessitates two compile phases, which is annoying in itself.  
> Another problem is that recompilation doesn't always catch that  
> closures should be rebuilt, so I end up with strange errors at  
> runtime, because not everything was cleaned. I find it strange such  
> a thing happens with only one Scalac phase, though.
>
> A combined Javac and Scalac would be a very nice thing.
>
> --
> Ola Bini (http://ola-bini.blogspot.com) JRuby Core Developer
> Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
> Practical JRuby on Rails (http://apress.com/book/view/9781590598818)
>
> "Yields falsehood when quined" yields falsehood when quined.
>
>

[Bill Venners]

Hi Ola and Oliver,

Another thing while we're on the subject. I have an API here, which  
will soon see the light of day (I know you've heard that before), that  
is all .scala files except for one .java file, because I need to  
create an annotation class in Java to use Java reflection to see it.  
Well, I can generate documentation about the Java class using javadoc,  
and about the Scala class using scaladoc. But then I'm going to have  
to stitch them together by hand. If people end up using Java and Scala  
together a lot, it might be nice if scaladoc could generate  
documentation from the Java classes in the API as well. (It could show  
a Java interface as a trait, for example, which is what people using  
the API documentation could think of it as.) This shouldn't be too  
hard. Javadoc is very extensible.

Bill

On Jan 7, 2008, at 1:05 AM, Oliver Lambert wrote:

> Hi Ola
>
> Cant help myself being nosy! Feel free to ignore my post, if the  
> answer is top secret : )
>
> Your generating Java through Antlr using Scala. Why (must be more  
> than learning Scala)?
>
> Oliver
>
> On 05/01/2008, at 6:26 AM, Ola Bini wrote:
>
>> Bill Venners wrote:
>>> Hi Stanislas,
>>>
>>> My take was that scalac would know to recompile a Scala class,  
>>> say, that used a Java one that had changed. Say the Scala class  
>>> called into a method that no longer exists. He wants that to be  
>>> caught at compile time without having to recompile every Scala file.
>>>
>>> I think this would be a great thing someday, but not urgent. It is  
>>> something I can live with. But I've noticed myself that it is  
>>> cumbersome to combine Java and Scala. In the ScalaTest build I  
>>> have Scala and Java together in both the source and the tests, and  
>>> I have four compiles that have to be done in just the right order.  
>>> Sometimes I have to do a clean build just to make sure everything  
>>> gets recompiled that needs it. I suspect mixed projects like this  
>>> would be common, but we'll have to see how Scala actually gets  
>>> used. If it is, I wonder if it would make sense to let scalac  
>>> handle all of the dependency checking, and let it fire off the  
>>> Java compiler when necessary.
>>>
>>> Bill
>> Bill!
>>
>> I totally agree about this. The project I'm working on to learn  
>> Scala uses almost only Scala code, but generates Java through  
>> Antlr. This necessitates two compile phases, which is annoying in  
>> itself. Another problem is that recompilation doesn't always catch  
>> that closures should be rebuilt, so I end up with strange errors at  
>> runtime, because not everything was cleaned. I find it strange such  
>> a thing happens with only one Scalac phase, though.
>>
>> A combined Javac and Scalac would be a very nice thing.
>>
>> --
>> Ola Bini (http://ola-bini.blogspot.com) JRuby Core Developer
>> Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
>> Practical JRuby on Rails (http://apress.com/book/view/9781590598818)
>>
>> "Yields falsehood when quined" yields falsehood when quined.
>>
>>
>
>

[Ola Bini-3]

Oliver Lambert wrote:

> Hi Ola
>
> Cant help myself being nosy! Feel free to ignore my post, if the
> answer is top secret : )
>
> Your generating Java through Antlr using Scala. Why (must be more than
> learning Scala)?
>
> Oliver
>

It's quite simple. As far as I know, Antlr doesn't have a Scala backend
yet, so I must generate Java. But that doesn't mean I need to use Java
to use the generated the lexer and parser.

Cheers

--
 Ola Bini (http://ola-bini.blogspot.com)
 JRuby Core Developer
 Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
 Practical JRuby on Rails (http://apress.com/book/view/9781590598818)

 "Yields falsehood when quined" yields falsehood when quined.

[Andrew.Foggin]

Ola Bini-3 wrote:

It's quite simple. As far as I know, Antlr doesn't have a Scala backend
yet, so I must generate Java. But that doesn't mean I need to use Java
to use the generated the lexer and parser.

Is there a reason you can't just use the parser combinator library and have a pure Scala solution?

--Andrew

[Meredith Gregory]

Andrew,

The combinator library is nice, but i for one have a lot of investment in a lex/yacc-like presentation of the DSLs i'm developing. i have found that BNFC offers a very nice solution in this regard. First, it provides coverage for generation of

• Haskell
• OCaml
• Java
• C#
• XML

Since the Java solution may also be used with the Scala, this pretty much covers my needs.

Best wishes,

--greg

On Jan 7, 2008 6:24 PM, Andrew Foggin (h) <andy@...> wrote:

Ola Bini-3 wrote:
>
> It's quite simple. As far as I know, Antlr doesn't have a Scala backend
> yet, so I must generate Java. But that doesn't mean I need to use Java
> to use the generated the lexer and parser.
>

Is there a reason you can't just use the parser combinator library and have
a pure Scala solution?

--Andrew

--
View this message in context: http://www.nabble.com/%22Verdict%22-on-Scala-tp14617903p14681172.html
Sent from the Scala - Lounge mailing list archive at Nabble.com.

--
L.G. Meredith
Managing Partner
Biosimilarity LLC
806 55th St NE
Seattle, WA 98105

+1 206.650.3740

http://biosimilarity.blogspot.com