Hello everybody, we are Roberto Negro, an Information Technology teacher in an High School in Cuneo (near Turin, Italy), and Piero Giuseppe Goletto, a software engineer and Linux expert. We are the project leaders of "GRANDASCHOOLNET". This project has the main purpose to leverage on some existing capabilities in the open source software field by which to create a network of all Schools in the Cuneo Province, which is one of the largest in Italy and where there is a demand for some advanced services both for the students and their families. 1. THE STORY SO FAR Initially we set up our portal following a "sure" and "consolidated path": Apache as a web server, adding PHP and Postgresql in order to deliver some dynamic content. You can see our portal at http://www.grandaschoolnet.it. The server is Linux-based, has had record uptimes, and we are glad to say that all in this project is made exclusively with open source software also by students. We added phpgroupware as a groupware solution in order to allow teacher and students to work together on projects. We presented the portal in a convention on March 14th and 15th 2003, here in Cuneo and that convention created broader interest especially in the content management area. So, we devised the need to have a real content management and content delivery solution in order to allow both teachers and students to create and publish content in a secure and user-friendly way. We chose Zope/CMF/Plone, on which we are migrating, and you can see the result of our efforts so far at http://www.grandaschoolnet.it:81/granda. We beg your pardon since the portal is only in Italian at the moment. Zope/CMF/Plone complies perfectly with our requirements, and since it is, in our opinion, an excellent product it formes the reference platform for our project. We will soon integrate in our Plone portal both Silva (Infrae's document management solution which is also open source) and FLE3 (e-learning application). Both are Zope based so this integration is a non-issue but we feel that it's interesting for you to know. 2. WHAT WE NEED TO DO First of all we need to consolidate and power-up our portal. This requires a tight integration between various components: Apache web server, Squid proxy server, Zope/CMF/Plone, AwStats as a log analyzer, and, more, consider that we have devised the opportunity for a Plone/OpenGroupWare integration which we will discuss in detail later. This is the reason why we crossposted this message on various mailing lists, since we feel it involves suggestions and contributions from various parts. We beg your pardon about that and we will make (and publish) an excerpt of all what we learn in this process on our portal. Our intention is to create an infrastructure similar to the one that Matt Hardy described in this document (see http://matthardy.us/mywiki/HighPerformanceZope), which he says he is working at. Although we won't implement at this stage a ZEO cluster like Matt will do, we think it is an opportunity we need to explore; we expect to get to serve users in the thousands in 1-2 years. Our idea, and we are kindly asking you for confirmation on feasibility and correctness of these details, is to activate the Zope/CMF/Plone and Apache (with mod_php, mod_kerberos, mod_ssl, mod_rewrite), FastCGI, Squid (using this as a proxy) and Postgresql on the SAME server in a way such that Apache/Squid will work as a cache and as an authentication mechanism based on Kerberos or on OpenLDAP. We could estimate a need for at least 256 Mbytes RAM on this particular solution (the server is an assembled Pentium IV standard machine with 40 GBytes Hard disk; the kernel is Linux 2.4.20 and distribution is a standard RedHat 8.0 in which Zope has been installed from tarballs). At this very moment neither Apache nor Squid nor FastCGI are installed on this server. 3. THE GROUPWARE COMPONENT We read an interesting article from Reuven Lerner, published by Linux Journal, from we quote: "The latest example of a killer app in the Open Source world is Plone [...]. Plone has become an increasingly prominent piece of software in recent months and has brought many people into the Zope community. And [...] Plone may well be the killer app for Zope." In our opinion Plone is the killer application for Zope but we feel that Plone's success would be much easier if a complete groupware solution is available for it. We have recently (in fact, last week) read about the OpenGroupWare project and we feel it very promising, and we think that Open Groupware is an easy path is to create simply the Plone interfaces to OpenGroupWare and get, that way, the specialized solution integrated with Plone. 3.1 How would we do this? We feel that such a project, released as open source as a Plone app, would be a no-nonsense contribution both for Plone (and by the way for CMF and Zope) and for OpenGroupWare. There would be a single Plone "product" for every OpenGroupWare functionality, which would interface to OpenGroupWare as an XML-RPC client. The user interface would be integrated in the Plone Look and Feel. Of course all this will be released with the same license as Plone. As we ask your opinion on this subject, we feel that it makes sense to adopt the "next generation" releases. We read that the beta of release 2.7 is out, so is it better for us to wait for the completion of next fall's Zope 2 and Zope 3X releases, since, we read, next fall's Zope 2 will be the basis for the migration from Zope 2 to Zope 3, instead of starting with Zope 2.7? Again, should we stick to CMF 2.0 (which relies on Zope 2.7, is it correct)? What are the exact relationship between Plone 1.1 and the Zope and CMF releases? We thank you all for your kindness and patience and for the explainations of our mistakes you certainly will give us, and hope to read from you soon. Best regards Roberto Negro (roberto.negro@multiwire.net) Piero Giuseppe Goletto (piero.goletto@tin.it)