Hello Burak,

I think you are making valid points for real business applications involving
sensitive data. However, I wonder how many times a javascript client will be
used for such applications. In the traditional HTML model, the client will
only expose data that the server feels safe enough to expose - all other
data remains on the (middleware) server. But I don't know enough about these
issues - the data I am dealing with (bibliographic data) is not really
sensitive - no need to encrypt anything except passwords. 

What I would propose, however, concerning your SOAP contribution, is to
reorganize the contribution structure according to the new contribution
skeleton, i.e., separationg library code from demo application,. Also, I
would propose to put the backend in a dedicated directory (I use "services"
for the PHP and Python jsonrpc servers on the same level as the "source"
directory). I think this would make including and using your contrib much
easier.

Best,

Christian 


Burak Arslan-2 wrote:
> 
> panyasan wrote:
>> But I would be interested why you think that authentication etc. should
>> be
>> part of the transport? I like the idea of having a transport that is as
>> transparent and simple as possible, an leave all the rest to the
>> application
>> and the backend. 
> 
> hi christian,
> 
> having authentication as part of the transport as a standard will let us 
> put the authentication implementation inside the core library -- saving 
> some developer time, thus lowering the cost of qooxdoo adoption.
> 
> having integrity and confidentiality as part of the transport is a 
> direct result of the routability property. SSL/TLS only works between 
> two points. if your message is supposed to be traveling through multiple 
> nodes, you need to encrypt the message itself -- tunneling it as 
> plaintext through an encrypted tunnel won't be enough. it's just like ip 
> vs ipsec.
> 
>> It seems to me that security and authentication
>> requirements are just too diverse - no one single implementation can
>> satisfy
>> individual use cases. The simpler the transport, the simpler you can
>> debug
>> it without having to use external tools.  
>>
>>   
> 
> i think you're somewhat right for authorization and audit 
> implementations, but authentication, along with integrity and 
> confidentiality of the communication between the client and the server 
> are common requirements. i think those requirements can only be 
> implemented using a well-defined sequences of cryptographic operations. 
> so why not define a standard, implement them in the core library and 
> save everybody else the hassle?
> 
> best regards,
> burak
> 
> 
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008
> 30-Day 
> trial. Simplify your report design, integration and deployment - and focus
> on 
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> qooxdoo-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel
> 
> 

-- 
View this message in context: 
http://n2.nabble.com/rpc-in-qooxdoo-tp4058077p4064489.html
Sent from the qooxdoo mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
qooxdoo-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel

Reply via email to