James Gosling: on the Java Road

« Compiler fun | Main | No T-Shirt hurling... »

20070109 Tuesday January 09, 2007


The language debate

Posted by barspi on January 09, 2007 at 08:54 AM PST #

James, you say that you never have been happy with the debate about languages features. I totaly agree with you. But with the release of Java under open source these kind of debates will be come more often. This is the top reason I would hoped Java would stayed non open source.

Sun should stay in control on the language feature, but others may vote for there favorite language feature.

Giving everyone control over the languages features is a bad idea and may lead to the end of the langue.

Welcome to the dark side of open source.

I didn't mean that I don't like the debate. What I don't like is debate divorced from experiment and data. Language arguments can get into all sorts of handwaving without building sample implementations. What ksl.dev.java.net is about is trying to provide a scientific basis for the debate. Throw stuff into the kitchen sink without thinking too hard about whether or not its a good idea. Let folks kick the tires. Those experiences then inform the choice of which features go into the standard. Permalink Comments [6]

Comments:

There is experimentation going on all over the Java landscape. I'd be curious to hear your reaction to some of these efforts:

Posted by java guy on January 10, 2007 at 10:00 PM PST #

Oops, typo. That should be JavaGI (uppercase-i), not JavaGl.

Posted by java guy on January 10, 2007 at 10:26 PM PST #

Cool. Might be an excellent idea to try out the BGGA proposal at the kitchen sink and get some clarity. Any volunteers?

Posted by Bharath R on January 10, 2007 at 11:53 PM PST #

Hi James,

This is regardling right-to-left languages on DOCWeb, again through the comments, sorry.
If you'd rather this is not done through the comments, it will be my pleasure if you email me to redbeard (at) gmail (dot) com.

I'm glad you implemented dir=rtl and it works well for the editing mode, however not so for the reading/printing/voting mode.
The simple reason is that the dir attribute is not supported on SPAN elements which what goes around the text in the reading mode. It is however supported on DIV elements.

I played with the document using Firefox's DOM inspector and the pages look good if the DIV element that wraps each translated text bit is given the dir=rtl attribute.

If you want I can also provide you with Hebrew translations for the 'version', 'since' and 'see also' texts.

Thanks,
Tal Rotbart

Posted by Tal Rotbart on January 11, 2007 at 02:59 PM PST #

Another issue, in editing mode, the dir=rtl should be applied to the edit field based on the language being translated into, not based on the language it is currently from.

Steps to reproduce:
1. Go to to an untranslated (to Hebrew) page in translate mode.
2. Because the text provided is still in English, the system doesn't tag it as dir=rtl and thus one can't see it in the right direction.

The workaround is to submit the translation and re-edit it, then the text is tagged as Hebrew and the dir=rtl is added to the text-edit box.

Posted by Tal Rotbart on January 11, 2007 at 03:17 PM PST #

Hi James, the comment you reply on is from me (Evert Tigchelaar jr, not barspi). I agree that arguments should be based on data and this should be more than just "it saves you a few keystrokes". Adding new language constructs should be limited but sometimes its good to add new things like generics (great to prevent lots a cast exceptions) and the enchanged for loop (but to bad that not the same syntax was used as in C# was is more readable). Readablity is very important, I hope you will keep that good in mind to prevent that Java will turn into the new C/C++ (power above simplicity).

Posted by Evert Tigchelaar jr on January 14, 2007 at 12:35 PM PST #

Post a Comment:

Comments are closed for this entry.