> > But if you say you won't require people write generic > > classes, you essentially require that no one use > generic > > interfaces as parameters. > > I don't see that.
How can you implement a generic interface without writing a generic class?
> Our architecture team is very hands on. You seem to have a > preconception associated with the title "architect." Our > Architecture team is not your typical. They are our best > developers with the deepest language knowledge. They don't > do paper architecture or ivory tower architecture. The > teams like to run their questions past them because of the > value they provide. Staying in touch is not a problem. > They are the most interactive people in the organization.
That's great and I applaud you and your organization for thinking in this way. However, I think you need to understand that a lot of organizations are not this way and that the simplicity of Java has a very particular benefit for less benign organizations. Java made it's name during the height of the dot-com boom. It's hard to remember but back then, finding qualified programmers was a tall order.
Things have changed, to be sure, but not everyone has caught up. Scala is a very different language than Java, not just in the obvious ways but in it's goals.
> > I say Scala is complex because it's complex. > > A tautology?
No, I'm just not confusing 'different' with 'complex'.
> > Complexity > > can be evaluated by the number of distinct atomic rules > > required to describe something. > > Do you have those counts for Java and Scala?
1.4 Java versus Scala? No but would you be willing to wager real money otherwise?
> I think one of the interesting things about Scala is how > the pieces work naturally together. That seems to reduce > some of the complexity. Whereas, some additions to Java > feel like additions in the sense of adding a carport to a > luxury home.
Absolutely. I think Java has lost it's sense of self.
> Our team is definitely atypical by design. We would fail > otherwise. We have very low turnover - by design, we avoid > contractors - by design.
That's wonderful. Are you hiring ;)
> > They don't seem to get that Scala is an OO > > language that act's as a bridge to the functional world > > not a pure functional language. > > I do. The original research money was for exploring the > hypothesis that OO and FP were a natural fit together.
Right. A lot of people get it. But some don't want accept this.
> You were saying you hadn't heard much recently. I was > responding to that. I've heard a lot but not the kind of > hype associated with other technology road kills.
I was? I don't recall writing that or thinking that.
> > You've made it clear that you do. You have the > > 'Architects' and the 'others'. > > I definitely sense a prejudice toward the title > 'architect.'
Actually, my current title is '... architect'. Not that that means anything in particular. It's not the title but your description that I honed in on. I probably misunderstood. You are absolutely right that I am the context of my own experiences though. You've convinced me that what you are talking about is different.
> > Like I said above and in previous posts, I don't want > to > > play feature cop. It's not worth it. I also don't > want > > any implicit conversions in my code. > > Then don't do it. Java lets you. What's your position on > using them in Java?
Sorry. I'm pretty sure implicit conversions aren't a feature of Java unless you are talking about conversions to String. That's just sugar with very small scope. It's not a new feature or something is it?
> We'll see. I don't see any other options out there as the > next Java. Maybe we are still a decade away. The forces > that drive these things are very unpredictable and > capricious.
Maybe we don't need the next Java. Maybe we need Scala. I guess my feeling is that it's not right for where I work. But it sounds like you are in (running?) a very mature operation and maybe it makes sense for you. But Scala isn't like Java, sorry.
> (uh, gosh, you 2 want to actually talk about invoke > dynamic *at all*? :-) :-) > > re: scala - how about > http://www.google.com/search?q=guidewire+gscript as a > viable alternative, meant to be less complex? i don't work > there and i've never used it, but i would like to get to > take it for a spin some day.
The conversation has gone where it wants to go. No one seemed to mind. Do you?
Is this the language that someone here kept mentioning in different threads?
Does anyone have any guess as to when this feature will be rolled into an official release? Will we see it in 7? 6.X?
When James Gosling was asked at JavaOne when he thought 7 would be released, he said it was up the the OpenJDK and they didn't seem to be in a hurry. Is there even a 7 date?
> Does anyone have any guess as to when this feature will be > rolled into an official release? Will we see it in 7? > 6.X? > > When James Gosling was asked at JavaOne when he thought 7 > would be released, he said it was up the the OpenJDK and > they didn't seem to be in a hurry. Is there even a 7 date?
If by "feature" you mean the changes related to the invoke dynamic JSR, I believe they can only ship with a new version of the JDK (since they are part of a JSR), and the aim (as far as I can tell from following the mailing lists) is to make this part of Java 7. It's unclear how that will work out in practice, since in this case it's actual changes to the JVM which are involved. I assume Sun will lead the rollout with HotSpot and other JVM maintainers will follow sometime afterwards.
What exactly gets included seems to be an open question at this point, although the invoke dynamic bytecode + some additional features (like method handles). It's up to the expert group--there's a review of an EG meeting at JavaOne this year over at http://tinyurl.com/3podxf. There seemed to be preliminary buy-in from the EG (which include JVM implementors) on the current proposal.
FYI, both the Multi-language VM Project (aka DaVinci Machine Project) mailing list and the JVM Languages Google Group are the places to follow progress on this. John Rose also blogs (in technical detail) on different proposals. The MLVM has a broader scope than just invoke dynamic -- there are several subprojects brewing -- and it's possible some of those features (tailcalls, some type of continuations support) might be bundled in with the JSR, though it's too early to say.
The sense I get, though, is that Sun is prioritizing this at this point. We should be able to see some real-world results in the next few months as the JRuby (and Jython?) developers start testing integration.
> We'll see. I don't see any other options out there as the > next Java. Maybe we are still a decade away. Why does nobody mention C#? Because it doesn't run on the JVM? Or because it's from MS? In terms of language features, this is irrelevant.
> No one mentions C# because it is as much a step backwards > as it is forwards. Yes, in terms of language evolution it occupies almost exactly the same territory as Java, just with a different owner.
Incidentally, in terms of Java that few understood:
class A { protected class B {} }
class C extends A { class D extends B {} }
This was theoretically possible in Java 1.1, but many compilers barfed on these structures (or even silently generated incorrect code). The value of this scheme was that class B had access to private members of A (and D likewise for C).
This code is still legal today, but I wouldn't recommend its use.
> The conversation has gone where it wants to go. No one > seemed to mind. Do you?
check out the original smiley + the text that wasn't about invoke dynamic and infer the answer. :-)
> Is this the language that someone here kept mentioning in > different threads?
yup. the thing i don't get is why the folks making it seem to still be using java anywhere at all when gscript looks a lot nicer. hopefully they'll have the resources and desire to release it as FOSS some day soonish.
Flat View: This topic has 22 replies
on 2 pages
[
«
|
12
]
Previous Topic
