[cap-talk] network level designation and authorization (RANT)

Sandro Magi smagi at naasking.homeip.net
Thu Jun 8 09:25:51 EDT 2006


John Carlson wrote:
> On Jun 7, 2006, at 5:31 PM, Karp, Alan H wrote:
> 
>> Far better and far simpler
>> is using unguessable addresses as with web-calculus.
> 
> Okay, let's use web-calculus.  How do I write a client that can
> use web-calculus to manage a capability passing system.
> Say I want to exchange capabilities and build trust between
> people?  And based upon this, how would I write a multiuser
> whiteboard?  So far, what I've seen of the client side of web-calculus,
> it involves using the  POST and GET classes, and the parameters names  
> that
> you are supposed to pass are extremely hard to get ahold of
> (requiring knowledge of XSLT). 

Only XML is required if you're using custom client software.

> How do we make it easier for
> people to write the client side of web-calculus.  Perhaps I am
> missing something from the documentation?

The documentation is extremely thin, so you're probably not missing
anything because it's likely not there. Just like WSDL, the service
interface is described via XML, and you can inspect (GET) or invoke
(POST) the reference graph as described the XML document at the link.

> Maybe someone can tell me how to force the waterken server
> to process XSLT on the server side, so I can get HTML to work with?

You will need to write XSLT yourself if you want a customized HTML page
for a given object; XML is the default wire format.

> I am open to people pointing me to places on one of the web-calculus
> web sites.
> 
> I am not open to writing a bunch of XML and XSLT processing steps.
> Those should be provided.  I am anti-XML because I feel that Scheme
> or Lisp offer a better solution.  Basically, I want to program in  
> Java on
> the client side.  How do I do this?

You would still need to process the returned XML, OR define another
serialization format that you like better (JSON has been mentioned for
instance).

> Basically, I have heard a lot of hot air about web-calculus, and
> only two people digging into the details.  What walls are people hitting
> when they approach web-calculus?  Are people just reading web-sites,
> and not digging into the code?  I think I've tried to explain my
> issues with web-calculus, but no one seems to be helping me.
> 
> I HAVEN'T been reading the web-sites, I've been digging into the
> code. 

The only good intro is the series of "Hello World!" examples Tyler wrote:

https://yurl.net/blog/tutorial/home

(appears to be down at the moment)

> I wouldn't declare something useable just because you've
> read a few web sites, and think lambda-calculus is wonderful.  Just
> because you see a web site, you fall in love with it?  Get real!
> 
> I've also tried to develop server side code with web-calculus,
> and it requires knowledge of XSLT.  What a load of crap!

Actually, XSLT is only needed if you're using a web browser as a client;
it transforms the standard XML format into HTML so you can browse it.

Sandro

> People really have bought off on this XML/XSLT crap, when it
> would be just as easy to use Java, C++ or some other more
> reasonable language.  Are people too scared to write a C++
> parser/generator for global variables?  I've done it, it's not that
> hard.  And you get  a true graph out of it, not some tree-like
> contraption.
> 
> For someone in my shoes, I offer the first few steps of XML
> independence.  On my site, I have written an XML->Java->XML
> converter.  This will allow you to step away from the XML
> monstrosity, and reenter the world of java. It's not pretty,
> but it certainly is a step.  One reason why I dislike Java is
> because it doesn't offer the same succinctness yet verbosity
> that XML offers.  I suggest that Sun start working on Java
> to incorporate XML features into the base Java language.
> 
> http://schizophrenics.net/javahier.tar.gz
> 
> Are there any sane computer scientists left out there?
> 
> John
> _______________________________________________
> cap-talk mailing list
> cap-talk at mail.eros-os.org
> http://www.eros-os.org/mailman/listinfo/cap-talk



More information about the cap-talk mailing list