Hello thanks for the answer. Here are some more considerations: - basically my philosophy is KISS. keep it short and simple (or stupid if u prefer :)) ), so merging too many different things in a production environment can lead to a mjor headache when something goes wrong... I will anyway have a look at the products you mentioned because they can be very interesting, so thanks for the advice :) - how do you store session data? I thought there was a project running, but it was not completed.. - I am getting confused with the PTK-now-CMF portal basically because on the PTK-CMF page there are references to an alpha version which comes after a stable release with a smaller number...:) Thanks Stefano -----Messaggio originale----- Da: albert boulanger [mailto:aboulang@ldeo.columbia.edu] Inviato: giovedì 15 febbraio 2001 17.27 A: spinwing@inwind.it Cc: zope@zope.org Oggetto: Re: [Zope] Java and Zope ok, I have managed to read the docs and found all answers to my previous (and unfortunately html tagged, sorry for that) mail. But now a more subtle question, bear with me one moment: - if I want to use servlets, theorectically no problem, I can still install tomcat or whatever, use servlets, do what I want and answer back somehow to zope. That's fair enough - for more advanced things however I have to get crazy or rely on XML-RPC, which is a good thing, but after all is not 100% java compliant (am I wrong?) In any case I have to follow up with the effort of single willing programmers, and not sun or dc. Now I am asking myself: what if zope would be able to expose its interface also as java classes? Here are some questions: Another thing to consider (depending on you need for coupling) is to use some emerging "Federation" web server/portal standards like the XML-based WDDX for portal content exchange that the Alliare is pushing. I have also used sessions in Zope to store a servlet session id so that a PTK-now-CMF portal login also does a servlet login and session. The Plumtree Corporate Portal product method of communicating metadata between the portal web server and gadget web servers is another way to approach this. The Plumtree server acts just like a web browser to gadgets servers it uses to generate those portal content boxes. It uses http 1.1 to communicate meta information as http headers and cookies between the portal server and gadget server. The metadata that are communicated are things like user preferences, user id, gadget id, and url of the parent page. Regards, Albert Boulanger aboulanger@vpatch.com