Distributed Matter - Blog - Comments

A look at WS-Notification - celynh

celynh — Tue, 23 Jun 2009 15:17:19 +0100

Hi !
You seems to be an expert with the domain of web service !
I need some help, and it would be very grateful if you can help me, even if it's just a little bit!

I work on a project with web service, but all this stuff is NEW for me... And I have to find a solution to make a message manager like this :

A generator write messages at anytime, we receive them in a message generator by web service (so callback?) and we have to notify the client (by web service, so callback again I guess), and send him the new message generated.

In brief, if you can help me about anything from Visual Studio, WS-*, SOAP, Callback, WCF, etc... If you have some books or anything, it would really help me

Thanks very much!

celynh

[ping] jSSLutils: Customisable configuration of SSL in Java - iLinkShare (Web 2.0 linksharing)

iLinkShare (Web 2.0 linksharing) — Wed, 28 Jan 2009 17:48:04 +0000

Foaf Links

Tagged your site as foaf at iLinkShare!...

Comparing Web of Trust and Hierarchical PKI for FOAF+SSL - Dan Brickley

Dan Brickley — Sat, 10 Jan 2009 19:25:34 +0000

Great writeup, I'm very happy to see all this work coming together! My interest currently is more in the document signing area, rather than SSL/TLS, but it's good to see tools like your pgp-to-x509 convertor that bridge some of these systems.

Couple questions -

Have you looked at XML Signature at all? There's a mode where you can treat RDF/XML (or other RDF-related formats, even SPARQL) as a blob, so we can forget about canonicalisation issues as a distraction I think. Here's a mockup of my foaf file signed,

Also - any thoughts on http://danbri.org/foaf.rdf.sigdata ... done with http://java.sun.com/javase/6/docs/t... and bogus key data, ... I'd love to see something like this integrated with your code. Any interest there? My hope is that we can run verifiers when populating a SPARQL db, much as I did in the old FOAF/PGP demos. Only this time we have a query language - SPARQL - that has some nice constructs for asking about provenance of data.

https://moo.cmcl.cs.cmu.edu/perspec... ?

[ping] jSSLutils: Customisable configuration of SSL in Java - Post Saver - Website voting and saving system

Post Saver - Website voting and saving system — Mon, 29 Dec 2008 11:02:25 +0000

Foaf Bookmarks

Bookmarked your page with keywords foaf!...

HTTP authentication mechanisms (and how they could work in Restlet) - Stephan Koops

Stephan Koops — Sat, 27 Dec 2008 10:54:57 +0000

Hi Bruno,

to the sessions with cookies:
If you create a session in the server, than there is a session in the server (of course ...), that means, that the communication is not statless, and that is against REST principles.

If you only put into the cookie, that user xy is logged in, than there is no session for it in the server. No session in the server -> stateless -> ok with REST principle of statelessness.

best regards

 Stephan

Comparing Web of Trust and Hierarchical PKI for FOAF+SSL - Henry Story

Henry Story — Tue, 14 Oct 2008 09:33:42 +0100

Hi Bruno,

thanks for your great help on the PKI and Java side. This has really helped move the project along.

Here is my reason why I think the web of trust model has a huge amount of potential.
---

Recently there has been a hugely successful web of trust model that has gained ground in nearly every application space: Social Networks.

These networks if they use PKI, certainly don't do it for anything more than to help the browser verify that it is connected to the correct web site. The trust you put in what people say about themselves is gained mostly because you trust that people would not be lying in front of all their friends. As with Open Source projects, the more eyes there are the more visible bugs become. Everyone's link to their friends adds a little trust to the system.

Linked in provides privacy, but not by using PKI. Your friends can see your detailed information, and who your friends are, the others get only a minimal view. Who sees what is decided by an algorithm that takes into account:

 + who you are. As per your ability to repeat what you told the system initially
 + your position in the social network (what claims others have made of you)

The problem with all these social networks is that they are centralized. What foaf+ssl offers is the ability to do the same in a distributed, decentralize setting.

see the audio slide show on this:
http://blogs.sun.com/bblfish/entry/...

What may be needed longer term is a way for the servers to describe who will get access to what resources -- which should help people debug broken restrictions. But I don't think there will ever be one or even just a few formulas to evaluate trust.

jSSLutils: Customisable configuration of SSL in Java - Bruno

Bruno — Mon, 09 Jun 2008 19:03:16 +0100

Hi,

Sorry, I'm afraid I haven't tried without OpenSC. I did report this bug (as mentioned on the Apple-CDSA mailing-list) to Apple, but Apple bug reports are under NDA, so you would probably need to report it too so that they can link your bug entry the one I've reported if you want to follow the progress. (You may need to create an account on https://bugreport.apple.com/ if you
don't already have one.)

Regards,

Bruno.

jSSLutils: Customisable configuration of SSL in Java - AndreaB

AndreaB — Mon, 09 Jun 2008 09:43:14 +0100

Hi, greetings from Italy.
I just started to work on OSX trying to sign with a certificate from keychain by java.
Did you find a workaround for the keychain Java API not obtaining the private key reference ? The smarcard I have to use are not supported by OpenSC and I have only a working tokend...
Ciao Andrea

HTTP-based Notification system - Bruno

Bruno — Fri, 10 Aug 2007 22:27:31 +0100

Thanks for the pointers. XMPP is indeed a good candidate for this type of application. HTTP and XMPP could certainly be combined if required, as mentioned in BOSH (or perhaps differently, as Mark Baker suggested a few weeks ago on his blog).

HTTP-based Notification system - Stelios Sfakianakis

Stelios Sfakianakis — Fri, 27 Jul 2007 10:53:03 +0100

Just to add "Using IM for Grid computing", http://www.dehora.net/journal/2006/...

HTTP-based Notification system - Stelios Sfakianakis

Stelios Sfakianakis — Fri, 27 Jul 2007 07:43:00 +0100

The other possibility is to use XMPP/Jabber http://www.jabber.org/

gsiftp URI madness - Bruno

Bruno — Mon, 08 Jan 2007 16:33:14 +0000

Thanks for taking this issue into consideration. I have just added a comment to an existing bugzilla entry about the gsiftp URI scheme (#3413): http://bugzilla.mcs.anl.gov/globus/...

Regards,

Bruno.

gsiftp URI madness - Ian Foster

Ian Foster — Thu, 21 Dec 2006 07:00:26 +0000

Thanks for pointing this out.

Globus is the product of an open source community. Community members do care about consistency and forward planning, but they're not perfect. We would be very appreciative of your input on how to address these problems that you have discovered. The best way to do so is to use the Globus bugzilla at http://bugzilla.globus.org.

Regards -- Ian.

gsiftp URI madness - M Keith

M Keith — Mon, 11 Dec 2006 10:23:33 +0000

Yes, this. I think that it is ridiculous that gsiftp exists without specifying a rigorous URL schema. URLs are a very convenient, application and protocol non-specific, way to describe data files in a grid environment.

This is another case of a lack of sensible forward planning in Globus, leaving a fractured product that is not self consistent.

Experiences with WSRF - Bruno

Bruno — Thu, 07 Dec 2006 16:34:38 +0000

I must admit I was a bit harsh with perl. I guess I was in a bad mood after a few fights with SOAP::Lite.

Experiences with WSRF - Mark Mc Keown

Mark Mc Keown — Wed, 22 Nov 2006 15:34:05 +0000

Perl does have strong support for types: scalar, array, hash etc...Other than that I kinda agree.