[M3devel] Is CM3 Modula-3 dynamic binding done at RT being formalized?

Jay K jay.krell at cornell.edu
Thu Sep 22 10:52:26 CEST 2011


GWT: Yeah, I've been meaning to learn this. Looks like one of the better "reachy" client platforms.I've been super busy, and will probably remain so for the long-term. Sorry.  - Jay
 Date: Thu, 22 Sep 2011 10:31:35 +1000
From: peter.mckinna at gmail.com
To: m3devel at elegosoft.com
Subject: Re: [M3devel] Is CM3 Modula-3 dynamic binding done at RT being	formalized?

We use GWT for browser based apps. You write the code in Java (in Eclipse) it gets compiled to Javascript with all the cross browser incompatibilities resolved (hopefully). As well if you want some nice widgets use SmartGwt which is a Java layer on some cool Javascript libraries. So you can get some safety from the Java and the code runs in the client. Removes a few headaches.


Better if it was M3 but its not a perfect world.

As Daniel says where is Jay? This forum is very quiet without him.

Regards Peter
 

On Thu, Sep 22, 2011 at 9:47 AM, Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es> wrote:


Hi all:
I think this is the question, I really don't like to quote, but as someone said, This is it!
It turns out that the history of Internet was marked by SPIN OS, because of several facts.
1. As the long time M3 friend, the birth of Modula-3 was from the Unix school (somehow ironical because of its C dependence), but then the Pascal progeny was that of a P-Machine with its almost OS.

2. Unix birth brought (as its Mach and SPIN representative newer descendants, among others SPINE, etc) had TCP technology (envisioned Internet as the main communication avenue, e.g SPINE, SPIN http server)
3. Java was planed to be the language of the Era (and it certainly was at the moment of its birth, objects, threads, exceptions, all from old school and newer ones like Modula-3, and nearer of NetObj)

4. Java sadly failed unfortunately due big birth of
 Internet mainly due safety* but mainly by performance, let's be honest here (JVM were planeed to be in everywhere if you read its books)
5. JavaScript was born or brought to life, unfortunately, and this is my personal opinion,  by not acknowledging the its roots, that's it, it didn't acknowledge the real influences but just by informal comments which hide precisely, one of the most important questions: the computation Model (Modula-3, later I explain why), and secondly the characteristics of it, mainly safety (I don't mean explicit typing I left this discussion to others who are before me).

6. JavaScript code is becoming critical when it comes to smart TVs, low-end phones, etc, as they are appearing in the market, but let's not forget that the requirements where more or less the same for Java Applets at its time (so this is why JavaScript needs to be serious and when the contra-revolution comes) as it was before its appearing. Then the
 question is why it failed, and if so, why do they can fix it still (Google V8, etc), well, the idea is perhaps the one more or less of Modula-3, you can run code efficiently in UNSAFE code, but most inner parts are safe, then, this is it, finally acknowledge that it's the Modula-3 security model, and secondly the computation model it's still not sound in terms of network models, Modula-3 Obliq style is more oriented to LANs, but fortunately this is what I meant by Cloud VM that has fixed this two models.

OK, if I am correct then this what happened with Modula-3, Java, JavaScript and that if completely finally acknowledge it (Obliq deserves and its precursors of course, Modula-3, mainly, but let's say Java in one way or another) of the most important credit. By denying it, they will cave more depth than of what they are now as they have brought us "their revolution", more way down.

Now, it was acknowledge the importance and recently more and more
 of SPIN, specially its dynamic linker to save the Modularity of JVM, but it was certainly 
special to correct their lack of security, and use its fixed loader as its application (that is linker-loader):
http://www.cs.cornell.edu/slk/papers/slk.pdf


I would like to see a port and working of m3loader just to see where did they arrive to in DEC-SRC days, Jay are you still there? :)
I can bring it compile it but not too much after that (I'm a big fan of compilers, but missing interfaces and Modules problems is harder than to fix simple bugs). Jay is pretty fairly the guy at the dark corners of MS linkers and loaders, interesting as some one suggested recently to see it in a platform independent "dynamic-network-aware linker caller", uhm m3linker, well that's harder work than just to recompile in its native platform. But could be certainly a good work to do later.


Thanks in advance

PS: There is a important fact about SPIN OS,
 that motivated by the above url of this description

--- El mié, 21/9/11, felipe valdez <felipevaldez at gmail.com> escribió:


De: felipe valdez <felipevaldez at gmail.com>
Asunto: Re: [M3devel] Is CM3 Modula-3 dynamic binding done at RT being formalized?
Para: "Daniel Alejandro Benavides D." <dabenavidesd at yahoo.es>

Fecha: miércoles, 21 de septiembre, 2011 15:23

Daniel, conectate al network.


On Fri, Sep 16, 2011 at 2:01 PM, Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es> wrote:



Hi all:

If it isn't it must be worked out I offer myself to do that but anyway, if folks don't dare to comment on that ... but will do my best possible effort to have such thing in a proposal fashion for the public, as nobody seemed very interested in the topic as for now.




In other things I want to talk also given that Obliq has the best "testbed" for the current times, an own calculus, and own Cloud VM implementation, and lots of research in ambients, languages and platforms, just to believe is incredible If I may say so. I spoke recently by email with DR Luca Cardelli, he told me it could be good for the language to have some rework and make it free to see what happens next, I believe that we could really make the contra-revolution of this times, with Modula-3 and its derivatives as for example Obliq. Or better to say, and re-re-evolution, that is the next one.






Thanks in advance





--- El jue, 15/9/11, Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es> escribió:



> De: Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es>

> Asunto: [M3devel] Is CM3 Modula-3 dynamic binding done at RT being formalized?

> Para: m3devel at elegosoft.com

> Fecha: jueves, 15 de septiembre, 2011 22:46

> Hi all:

> I have read the addition to the Modula-3 CM· system was

> done by effect of CM-JVM dynamic linking and loading, has

> this facility being formalized (i.e is part of the current

> language definition, why it hasn't someone has asked that

> before but no clues been actually).

> Somehow one of the open problems on Larch project at DEC

> was the creation of sub-sorting facility (<:) simplifying

> the Modula-3 type system, I would some kind of different

> symbol for that operator (perhaps [:  or say '(:' ) but

> Horning proposes to deal with it syntactically with it so

> more research would be needed (I don't know why he can say

> that, I guess is because is supposed to simplify the type

> system, which could be good of course, but I don't know

> whether '(:' could mess the syntax of the LM3 language ):

> http://nms.lcs.mit.edu/larch/pub/wadt11-slides.ps

>

> By the way we should recover the LM3 semantics manual

> because that way we can show the model intended to be in the

> ESC way and railroad diagrams, etc, kind of more formal, I

> believe they did quite an amount of work on the Larch

> project but as ESC began got somehow abandoned after that,

> but still quite advanced work for the time.

>

> More interestingly there are some efforts to revive this

> old dilemma of Computation and deduction chicken and egg

> problem, that specification languages are Turing-computable

> or not, etc. I believe they are in DEC cases they are

> because of they experience with DECspec, etc and others as

> well are the SPEC language by Lampson, which has a lot

> attention right now out there (i guess there is some

> prototype tools for handling it but nothing of verification

> that I know).

>

> Also this way we can offer some more knowledge to the

> readers interested in DEC's initiatives and its roots and

> effects in the time being passed, after they became

> important (like ten years or more later just as they planed

> things being in current use, but as unfinished they due

> unsatisfiability discarded before its time). Anyway CM3

> source still has still lm3 interface code files I believe

> one could ask why are there those files if they weren't

> used, which I can think they were but because of state of

> the art tools they were not released to the big public now

> but are there still:

> http://modula3.elegosoft.com/cgi-bin/cvsweb.cgi/cm3/m3-libs/m3core/src/thread/Common/larch/#dirlist




> http://modula3.elegosoft.com/cgi-bin/cvsweb.cgi/cm3/m3-libs/m3core/src/thread/Common/set.lsl




>

> Thanks in advance

>



-- 
312-444-2124Skype: f3l.headhunterCasa: 8043901





 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20110922/69a3115e/attachment-0002.html>


More information about the M3devel mailing list