7.0 wishlist?

[Hendrik Maryns]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Harold Yarmouth schreef:

Insulting my ability to insult people now? Tsk, tsk. I wasn't trying to
insult you. If I had been, I would not have written "no offense". And
you WOULD have been insulted.



But this, too, has drifted away from being relevant to Java.

Good, so from now on I just finish every post I write, which is maybe
marginally directed to you with ‘no offense’ and we can stop all the
bullshit about being insulted. Please. No offense.

H.
- --
Hendrik Maryns
http://tcl.sfs.uni-tuebingen.de/~hendrik/
==================
Ask smart questions, get good answers:
http://www.catb.org/~esr/faqs/smart-questions.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iEYEARECAAYFAkkYJOMACgkQBGFP0CTku6PQkQCfRsOR/Tuudm9JYriXmYuAEi1v
3nwAn0bUF2vOFeYGlXGlN7OXtlh/qgN5
=76h2
-----END PGP SIGNATURE-----

 

[bbound]

(e-mail address removed) schreef:

Does anybody understand this?

Yes -- I, for one.

No. None of the nasty things that either of you have said or implied
about me are at all true.

I wonder whether he has set this as a signature in some mail program.
Is G2 google groups?

[implied insult deleted], H.

No. None of the nasty things that you have said or implied about me
are at all true.

 

[Harold Yarmouth]

Harold Yarmouth schreef:

Good, so from now on I just finish every post I write, which is maybe
marginally directed to you with ‘no offense’ and we can stop all the
bullshit about being insulted. Please. No offense.

Oh, but by calling anything that I've written "bullshit" you've just
insulted me again, not to mention been potty-mouthed ...

I recommend just leaving anything out of a post that you might otherwise
try to "patch up" by slapping a "no offense" at the end. Whatever it was
was about a person rather than about Java anyway.

Perhaps the newsgroup ought to be renamed, too, from
comp.lang.java.programmer to comp.lang.java.programming. The present
name seems to have created some confusion among certain posters, such
that they think discussing a Java programmer (even in non-Java-related
ways) is within the topic bounds.

 

[Harold Yarmouth]

We are discussing

We are discussing the best Usenet newsgroup for this topic. Non-Usenet
fora are not relevant to such a discussion.

Your claim was that my post was off-charter for cljp, despite its
obviously being within the topic bounds. That can only be true if there
is another newsgroup that is a better topic fit.

Even then, I notice that although there is a comp.lang.java.gui a lot of
AWT or Swing related stuff gets posted here and there is no hue and cry
that such posts are off-charter.

I think you are just grasping for an excuse why my post was
inappropriate, for the simple reason that a) it was NOT inappropriate
and b) you, personally, didn't like it. Oh, and c) apparently you can't
think up a simple, logical objection, so you're forced to make something
up. If that last item weren't true I'd have expected you to post a
simple, logical objection instead of browbeating me and making vague
assertions that my post was not in the appropriate newsgroup. Especially
once a logical explanation for your objection was requested.

Mailing lists are where the real action happens

Fine, then feel free to unsubscribe comp.lang.java.programmer and go
bother one of these exalted mailing lists.

However I see no badge of authority by which you can order me to do so.

If you prefer mailing lists to newsgroups, that's your own personal
choice, and there is an appropriate way to act on it (go there yourself)
and an inappropriate way (try to force your preferences upon others).

Yes you do.

Are you hard of hearing? I just said "I'm not a Sun engineer or bigwig
...." and it's quoted right there above. I should think that I know more
than you do about whether or not I'm a Sun engineer!

/me whacks Harold

And now you reach rock bottom: silly expressions of desires to commit
acts of physical violence.

Perhaps you should stop typing this stuff here and start typing it into
Eliza.class* -- I think she may be able to help you more than I can.

(* Transparent and pathetic attempt to get back on topic.)

<http://hg.openjdk.java.net/jdk7/jdk...883246/src/share/classes/com/sun/tools/javac/>

Nor am I interested in personally acquiring that expertise. I have lots
of other fish to fry right now, some of which are actually urgent. If
the stuff at the URL won't help me make my deadlines or directly and
quickly earn me some money, then I don't have time for it at the moment.
I barely have time to defend myself against your latest round of public
insinuations about me.

From my discussions with actual spec writers,

Your discussions with spec writers are not relevant here. Try
alt.writing.fiction or similar.

Right.

So will you stop harassing me now?

Good programmers can learn by example.

Good TEAM programmers delegate some jobs instead of trying to do
everything and be everything themselves.

I just asked my roommate. He agrees with me.

Then he's as thick-skinned (and rude) a person as you are. Perhaps
that's a birds-of-a-feather thing -- or simply because nobody else could
stand having you for a roommate.

In terms of personality aspects, it seems you're in the minority here.

I find that unlikely. My personality is, in a nutshell, "civilized".
Yours appears to be "rude American, probably from New York or somewhere
similar". Either because that's your actual ancestry, or because that's
the sort of role model you emulated when learning social skills. At my
count, I estimate that there are four or so other people with social
skills as shoddy as yours or worse here: Lew (as shoddy), Arne (a lot
worse), RedGrittyBrick (as bad; hasn't locked horns with me, but I've
read some of his recent posts here in reply to other people), and Peter
Duniho (worse, but not as bad as Arne; ditto). There're probably one or
two others. A quick grep shows around* 117 unique names in this
newsgroup for the past month or so.

You're in a 4-5% minority, by those numbers.

(* Mileage may vary depending on newsserver -- retention policy, spam
filtering, and so forth -- and your local filtering. If you're not
filtering that guy that keeps screaming about feedback loops, you'll
probably see more. If you're actually successfully filtering that very
persistent shoe salesman, you'll probably see fewer. If you're using
Goggle Gropes, you'll definitely see more, none of the added ones good.
If you're using any server that actually employs CleanFeed, you'll see
fewer.)

 

[Harold Yarmouth]

Discussing whether or not I insulted you or you insulted me is off-topic
for this forum.

Yeah, so cut it out.

Discussing whether or not this subject matter is on- or off-topic for
this newsgroup is off-topic for this forum.

Yeah, so cut it out.

Math expressions rarely have side effects, and math is the primary
target for operator overloading.

I gave you an example.

((array[i++] << 8) & 0xff) | (array[i++] & 0xff)

That is side effects.

That is bit-twiddling, not matrix multiplication.

In fact, that is the kind of code I used to fire people for writing in
C, before I decided that I preferred programming to managing programmers
and went back to basics.

Stop insulting me!

Eh, I didn't -- YOU insulted ME, remember?

/me hooks up a meeting between Harold, Twisted, zerg, and JSH.

Me, who, who, and who?

Anyone else wishing to attend, please bring along your LARTs. You may
also invite a friend, like those dinners in /Le Dîner de Cons/.

Whatever this is, it appears to be completely unrelated to the topic of
Java programming.

 

[Harold Yarmouth]

Harold Yarmouth schreef:

[some attributions are missing here, but I think it's just me and
Hendrik alternating]

I see no reason why [a,b] would not create a new pair in the
background each time.
No existing literal does so, other than primitive literals, and those
only because primitives are passed by value rather than by reference.
Well, yes, but that is an easy claim, since the only non-primitive
literals allowed at this point are Strings.

True, but my point was that the division logically should be based on
whether the object is passed by value or by reference.

Makes sense, but already the fact that you have to think about the
pass-by method makes the concept more difficult to grasp.

No more than you already do, with String vs. primitive literals, or with
objects and arrays vs. primitives in general. Arrays of primitives being
passed by reference seems to trip a few students up in every Java class.
Should we change to passing arrays of primitives by value, then? Surely not.

I shudder at the thought having to explain this to newcomers.

Tell them "it's like a String literal". Or just remind them that
primitives and Objects live by different rules.

It does, but let me think about this a bit:

{ 1, 3 } is an array, so should, like any array in Java, be passed by
reference.

Now, is it mutable? Take the method

public void nasty(int[] ints) {
ints[0] = 0;
// other stuff
}

What do you propose would happen if you did nasty({ 1, 3 })? Without
the comment, nothing of course, but let’s just assume the ‘other stuff’
does more with the array, like use its values.

What does it currently do if passed

static final int[] array = {1, 3};

If that array can actually be modified, then there's a bit of ugliness
in Java, but it does not stem from any proposal of mine.

So I’d say: even you literal is mutable. I don’t think it is desirable
to introduce new behavior here.

But are we? See above.

Actually, had I been on the committee that came up with the original
design for Java, non-nullable references and casts would have been in
from the get-go, and there'd have been some kind of const notion.
Probably even a complete split between "value types" (immutable, derive
from Value, Value has equals() and hashCode overridable) and
"actor/accumulator types" (mutable, have enforced Object equals() and
hashCode()), and parallel sets of collection classes. Collections as
they are now are schizophrenic -- sometimes used in algorithms as
accumulators and the like, and sometimes used as values.
Collections.unmodifiableFoo is a band-aid that's only somewhat useful
(what if a reference to the mutable delegate is still running around
loose?). Having a separate set of unmodifiable collections, with some of
the modifiable ones able to act as builders for these, would perhaps
have been better.

Generics from the get-go, and for-each, would have been nice too. But
though they could be added later, there's no way the core collections
classes will get such a massive overhaul, let alone messing with
java.lang to the point of String deriving from Value etc.

(Now THERE are proposals that, had I made them, you'd have been right to
shoot down with "nobody will ever go for that! It would break too much
existing code!" I didn't dare mention such ideas before, but -- might as
well hang for a sheep as for a lamb.)

 

[Mike Schilling]

(e-mail address removed) schreef:

Does anybody understand this?

Yes -- I, for one.

No. None of the nasty things that either of you have said or
implied
about me are at all true.

I wonder whether he has set this as a signature in some mail
program.
Is G2 google groups?

[implied insult deleted], H.

No. None of the nasty things that you have said or implied about me
are at all true.

Hold on. Wasn't "Harold" denying being yet another alias for Twisted?

 

[Andreas Leitgeb]

What does it currently do if passed
static final int[] array = {1, 3};

static void fubar(int[] a,int i) {
if (i==0) { a[0] = 0; }
System.out.println(java.util.Arrays.toString(a));
}

sample1:
static final int[] array = {1, 3};
public static void main(String[] args) {
for (int i=0; i<2; i++) {
fubar( array, i);
}
}

sample2:
public static void main(String[] args) {
for (int i=0; i<2; i++) {
fubar( {1,3}, i);
}
}

In sample1 it is perfectly clear, that the call will
change that array, so second iteration will see 0 in
first position. It is also possible to write it
differently such that a new array is created for
each usage, but that will be just as obvious:
for (int i=0; i<2; i++) { fubar( array.clone(), i); }

In sample2 there are two interpretations:
1.) on entry, the method *always* sees an array {1,3}
2.) the literal array is stored somewhere, and passed
each time, whereever it is used, thusly mimicking
the behaviour from sample1 except for the lack of
any visible field/variable to hold it.
3.) fubar would throw some exception on trying to write
to the (immutable) array. In that case it would
make sense to keep it in the ConstantPool.

In the first case the array (if it was indeed stored in the
constant pool) needs to be at least cloned at each use, thus
thwarting all the savings you suggested.

The second case is much worse... Let's add another sample:
fubar( {1,3} , 0);
and 100 lines down in the same class:
fubar( {1,3}, 1);
Aren't literals usually collected? Does the second have its
own literal to modify, or will it share the first one's?

PS:
My personal order of preference for the interpretations
would be: 0,(long pause),3,(even longer pause),1,2
with 0 being the status quo with respect to array-literals.

PPS:
I may of course just have missed the obvious and brilliant
idea that solves all the concerns.

 

[Hendrik Maryns]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Harold Yarmouth schreef:

Perhaps the newsgroup ought to be renamed, too, from
comp.lang.java.programmer to comp.lang.java.programming. The present
name seems to have created some confusion among certain posters, such
that they think discussing a Java programmer (even in non-Java-related
ways) is within the topic bounds.

LOL, good one.

H.
- --
Hendrik Maryns
http://tcl.sfs.uni-tuebingen.de/~hendrik/
==================
Ask smart questions, get good answers:
http://www.catb.org/~esr/faqs/smart-questions.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iEYEARECAAYFAkkay5AACgkQBGFP0CTku6MVBQCeJd4m9LOMYNZvajYHsaD9mCf+
56AAoJgxgSGYV7e2VwiWEtF/u5c9I/fb
=YSd4
-----END PGP SIGNATURE-----

 

[Hendrik Maryns]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Harold Yarmouth schreef:

Harold Yarmouth schreef:

[some attributions are missing here, but I think it's just me and
Hendrik alternating]

I am sorry for that.

It does, but let me think about this a bit:

{ 1, 3 } is an array, so should, like any array in Java, be passed by
reference.

Now, is it mutable? Take the method

public void nasty(int[] ints) {
ints[0] = 0;
// other stuff
}

What do you propose would happen if you did nasty({ 1, 3 })? Without
the comment, nothing of course, but let’s just assume the ‘other stuff’
does more with the array, like use its values.

What does it currently do if passed

static final int[] array = {1, 3};

If that array can actually be modified, then there's a bit of ugliness
in Java, but it does not stem from any proposal of mine.

I think it is modifiable. AFAIK, there is no way in (current) Java to
make an array immutable.

I tend to agree that this is a bit of ugliness.

But are we? See above.

Which above? If you mean the array should be immutable, then AFAIK you
are introducing new behavior, since there is no way to do that now.

H.
- --
Hendrik Maryns
http://tcl.sfs.uni-tuebingen.de/~hendrik/
==================
Ask smart questions, get good answers:
http://www.catb.org/~esr/faqs/smart-questions.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iEYEARECAAYFAkkazNAACgkQBGFP0CTku6P+zACgtrjzPjoK7bZfA8A0MgtVlk95
JdwAniVkvXiNg6Di6/8NHercc3pTzhXP
=4LQ8
-----END PGP SIGNATURE-----

 

[Joshua Cranmer]

((array[i++] << 8) & 0xff) | (array[i++] & 0xff)

That is side effects.

That is bit-twiddling, not matrix multiplication.

It is very much the same vein. Face it: expressions can, and will, have
side effects. Even if addition is commutative in mathematics, you can't
assume it's true in computer science.

In fact, that is the kind of code I used to fire people for writing in
C, before I decided that I preferred programming to managing programmers
and went back to basics.

Well, its outcome is undefined in C since when i is incremented is
unspecified.

 

[Roedy Green]

You can't even type "circle plus" on any normal keyboard, so that's
completely out of the question.

Any IDE could deal with that.
--
Roedy Green Canadian Mind Products
http://mindprod.com
Your old road is
Rapidly agin'.
Please get out of the new one
If you can't lend your hand
For the times they are a-changin'.

 

[Roedy Green]

You can't even type "circle plus" on any normal keyboard, so that's
completely out of the question.

Any IDE could deal with that. The IDE could display the operator iv a
special font, colour, size, or with one of the Unicode symbols.

You might not even need to key it differently. The IDE might be smart
enough to discriminate.

You might use a scheme similar to AllChars for keying the chars in
pure text editors.

see http://mindprod.com/jgloss/allchars.html


See http://mindprod.com/project/scid.html
for even more far out ways to deal with these sorts of problem.

"I mean, source code in files; how quaint, how seventies!"
~ Kent Beck
--
Roedy Green Canadian Mind Products
http://mindprod.com
Your old road is
Rapidly agin'.
Please get out of the new one
If you can't lend your hand
For the times they are a-changin'.

 

[Arne Vajhøj]

Any IDE could deal with that. The IDE could display the operator iv a
special font, colour, size, or with one of the Unicode symbols.

You might not even need to key it differently. The IDE might be smart
enough to discriminate.

A language that can only be coded in an IDE, because it can not
be typed in using a standard text editor sounds as a very bad
idea to me.

Arne

 

[Arne Vajhøj]

Yeah, so cut it out.

According to the quoted material, then you started that topic,
so tell it to yourself.

Yeah, so cut it out.

You missed the point.

Arne

 

[Arne Vajhøj]

Harold Yarmouth schreef:

static final int[] array = {1, 3};

If that array can actually be modified, then there's a bit of ugliness
in Java, but it does not stem from any proposal of mine.

I think it is modifiable. AFAIK, there is no way in (current) Java to
make an array immutable.

I tend to agree that this is a bit of ugliness.

It is modifiable.

And there is nothing ugly about it.

It is equivalent to C++:

int *const array

But Java is missing and equivalent of:

const int *array

Arne

 

[Arne Vajhøj]

Hendrik Maryns wrote:
No more than you already do, with String vs. primitive literals, or with
objects and arrays vs. primitives in general. Arrays of primitives being
passed by reference seems to trip a few students up in every Java class.
Should we change to passing arrays of primitives by value, then? Surely
not.

Arrays in Java *are* passed by value !

Arne

 

[Arne Vajhøj]

We are discussing the best Usenet newsgroup for this topic. Non-Usenet
fora are not relevant to such a discussion.

Only for people with a narrow mind set.

I find that unlikely. My personality is, in a nutshell, "civilized".
Yours appears to be "rude American, probably from New York or somewhere
similar". Either because that's your actual ancestry, or because that's
the sort of role model you emulated when learning social skills. At my
count, I estimate that there are four or so other people with social
skills as shoddy as yours or worse here: Lew (as shoddy), Arne (a lot
worse), RedGrittyBrick (as bad; hasn't locked horns with me, but I've
read some of his recent posts here in reply to other people), and Peter
Duniho (worse, but not as bad as Arne; ditto). There're probably one or
two others. A quick grep shows around* 117 unique names in this
newsgroup for the past month or so.

You're in a 4-5% minority, by those numbers.

Weird conclusion.

Everybody posting agrees with Joshua except for you 6-7 identities.

Arne

 

[Arne Vajhøj]

Mailing lists are where the real action happens for all kinds of
specification development (W3C uses them almost exclusively). If you
seriously want this features, you will have to go there.

He has a huge problem with mailing lists.

They don't like anonymous users and trolls get thrown out
very quickly.

Arne

 

[Arne Vajhøj]

No, as I just demonstrated, you can often predict in advance that
certain searches will be fruitless.

It has been proven that this search would not be fruitless, so
your prediction was rong.

Arne