ECOOP paper

[sha...@uurda.org]

   Is http://bracha.org/newspeak-101.pdf the most up to date doc on Ns?  It’s from late 2014, which is amazingly recent. 

 

Yes, we have tried to keep it up to date.  The next major update will focus on support for access control.

 

Do you mean control of access to source code in the image?

 

This is already working in the web version, and we have a version of the VM under development  that enforces it as well.  At some point, we'll throw the switch and enforce the semantics.  This will make Newspeak even more different than Smalltalk. It is the basis for security and true modularity.

 

For a deeper understanding, read the ECOOP paper (listed under docs).

 

Is there a title?

---

This email has been checked for viruses by Avast antivirus software. www.avast.com

[Gilad Bracha]

On Sun, Apr 5, 2015 at 4:17 PM <sha...@uurda.org> wrote:

 The next major update will focus on support for access control.

 

Do you mean control of access to source code in the image?

I mean that the access modifiers private or protected are enforced at runtime as dictated by the language spec, so that object can act as true capabilities.

The tricky thing is to provide easy access to non-public members from the programming environment during debugging etc. This is a delicate tooling issue; if it isn't done right, the nice feeling of liveness can be lost. 

I have no doubt that this will be one more barrier to entry for Smalltalkers.  Smalltalk has no concept of modularity and is impossible to make secure.

Is there a title?

 It is listed on the docs page at the Newspeak web site.  Or you can type Newspeak ECOOP into a good search engine and it will find it for you. Here is a link

http://bracha.org/newspeak-modules.pdf

In general, there is a lot of material on the web site - documents, blog posts, audio, video - and it is well worth looking into. 

[sha...@uurda.org]

I mean that the access modifiers private or protected are enforced at runtime as dictated by the language spec, so that object can act as true capabilities.

 

I see.   Are access modifiers actually useful?  I’m not convinced.  I can only recall once example in Smalltalk for namespace imports.  I guess that is okay.   Where do you find you typically need such a compiled-time spec?

 

The tricky thing is to provide easy access to non-public members from the programming environment during debugging etc. This is a delicate tooling issue; if it isn't done right, the nice feeling of liveness can be lost. 

 

What happens if we solve that problem by getting rid of all access specifies?  Why aren’t nested classes and methods good enough for scoping.

 

I have no doubt that this will be one more barrier to entry for Smalltalkers.  Smalltalk has no concept of modularity and is impossible to make secure.

 

Can you define what you mean by “module” in this context?

Is there a title?

 

 It is listed on the docs page at the Newspeak web site.  Or you can type Newspeak ECOOP into a good search engine and it will find it for you. Here is a link

 

http://bracha.org/newspeak-modules.pdf

 

 

I wasn’t sure which topic or paper you were referring to.  I saw three papers “with ECOOP” in the title. 

 

The paper looks interesting.  It will have to be second in the queue.

You want a first-class module in the language, not just file-system-level structures, like parcels.   And you want this as a way of packaging code for applications, such that loading a module works as a parcel does, adding and overriding as needed?

[Gilad Bracha]

On Sun, Apr 5, 2015 at 7:34 PM <sha...@uurda.org> wrote:

I mean that the access modifiers private or protected are enforced at runtime as dictated by the language spec, so that object can act as true capabilities.

 

I see.   Are access modifiers actually useful?  I’m not convinced.

I'm not interested in convincing you.  If you want to know more, read the paper, watch the talks etc.  Sorry, but I can't spend any more time explaining this stuff.  If you understand and appreciate it, good. If not, that's fine too - use Pharo or Ruby or PHP or whatever. 

[sha...@uurda.org]

I see.   Are access modifiers actually useful?  I’m not convinced.

 

I'm not interested in convincing you.  If you want to know more, read the paper, watch the talks etc.  Sorry, but I can't spend any more time explaining this stuff.  If you understand and appreciate it, good. If not, that's fine too - use Pharo or Ruby or PHP or whatever. 

 

I will.  Sorry; it’s in the queue of things to read.  I thought you might want to share a quick justification for such a feature.  I always try to restrain myself when thinking about adding a new constraining feature to a language.  You probably do the same.