FAQ
From FezWiki
[edit] Feel free to add your own questions about Fez and we will try to answer them as best we can!
[edit] How is Fez development funded?
Fez development is currently being funding by a grant from the Australian Department of Education, Science and Training (DEST) via the Australian Partnership for Sustainable Repositories (APSR) project (see the UQ eScholarship page for details). The funding was granted for the period 2005-2006, but has been recently extended until the end of 2007.
Fez Shibboleth integration has been funded by a MAMS minigrant. It will be released as an optional authentication feature in Fez 1.3. It was released in Fez 1.2 with partial functionality.
Near to the end of the APSR grant other funding opportunities will be assessed. By this time we hope the Fez open source community to have grown to a more mature level. As the software is being used by UQ some degree of development will continue in any case.
[edit] What are some live production sites using Fez?
- UQ eSpace - UQ's production Fez site. We have now migrated from ePrints and will migrate from ADT (for theses) and other systems early in 2007.
- eSpacio at UNED, Spain (Fez translated into Spanish)
- Portal for the Norwegian Ministry of Education and Research (Fez translated into Norwegian)
- Colorado Alliance of Research Libraries
- NSDL Materials Science Digital Library
See Sites Using Fez for a complete list.
[edit] What Skill set would I need to implement Fez?
(This was written by Caroline Ayers on the RUBRIC project in reply to a question from one of our partners)
This person would need at least basic knowledge of the following:
- MySQL
- Fedora
- Fez - written in PHP, but also CSS, html (Smarty html templating)
- Apache
- (related other pre-req software)
- Understanding how it all fits together
This would also involve that person (obviously) having programmer level access to the server on which Fez ran (something to be considered, depending on your IT department's policies).
[edit] What hardware/operating system do I need to host Fez?
You can put Fez on Linux, Windows or Apple OSX. All the required software can go on the one server including Fedora, Apache, MySQL, PHP etc. Or you could put each service on a different server, eg one server for Fedora, one for Fez, one for MySQL - it is up to you and you can configure the setup to your specific server environment.
UQ is running our production Fez (UQ eSpace) all on a Dell Poweredge 2650 server with Dual Xeon 3.4GHz and 4Gigs of RAM with Centos 4 (Linux). The hard disk space required by the software is minimal, the main disk usage will be in the size and type of collections you wish to put into Fez.
Fedora is runs on its own Apache Tomcat so its always helpful to give Tomcat lots of memory. Preferably 2GB although some Fedora sites have dedicated Fedora servers with 8 to 16 GB of RAM. CPU power is also important, but unfortunately there hasn't been much benchmarking stats made available to show performance across different platforms and specs. As always - the more the better and faster Fedora and Fez will run.
[edit] Does/will Fez have full-text indexes?
It is a planned feature for the next release (1.3) to full text index PDF and possibly MS Word documents.
[edit] Does Fez extract preservation metadata using JHove?
Yes Fez extracts technical preservation metadata using JHOVE as one of the automatic workflows when a user adds or updates an object with file attachments. Fez has recently been audited against the PREMIS preservation model and most of the requirements were met. The main areas that need more attention are 1) better human understandable object history logging 2) GUI administered version control. Both of these features were already planned for in the Fez 1.3 roadmap before the audit.
[edit] Can changes to configuration be version controlled?
Problem: Fez has a web GUI for changing a lot of configuration. This has the advantage that a lot of configuration can be done my non technical staff. But how can repository administrators keep track of changes, roll them back, and move configuration that might be developed in a test system into production?
Solution: (Taken from a feature request for 1.3 in the Wiki proposed by Matt) ->
It will be essential to have an upgrade path for customised XSDs and workflows. At the moment the new database when it is upgraded will overwrite any changes that users have made to the XSDs and workflows. I propose a system that lets you export and import workflows and XSDs (including all the matching fields stuff) to XML files. When we send out version 1.3, the upgrade should use these XML files rather than overwriting existing tables.
So when developing this upgrade functionality for Fez 1.3 we will take into consideration configuration management. For the current version of Fez in a test and production rollout environment we suggest making the changes to the test Fez site first, then migrating the
related configuration MySQL tables:
- fez_xsd
- fez_xsd_display
- fez_xsd_display_matchfields
- fez_xsd_display_mf_option
- fez_xsd_loop_subelement
- fez_xsd_relationship
So basically all the xsd tables. One way to version control this might be 'mysqldumping' all these tables to Fez and wrapping a version controll/rollback management panel around this process.
[edit] How does Fez integrate with 3rd party software?
This depends of what type of software and how you want Fez to integrate with them. For example Fez does use 3rd party software as part of the workflow system for gathering preservation metadata (with JHOVE) and converting images (ImageMagick).
For additional tools like these (command line executables) you could add a workflow web-service wrapper by writing a simple PHP script based on those above two.
Fez can already integrate with database feeds from organisational database systems for academic author lists, organisational structure and the relationship between your authors and those org units. In Fez 1.3 there will be admin panels and some automation for these, but in Fez 1.2 you need to manually copy these database feeds into the Fez org tables with something like SQLYog.
Integration with other systems like your organisational central authentication/authorisation source (eg Microsoft Active Directory or an LDAP server) is easy to setup in the Fez configuration so your users can login with their institutions username and password. Fez also supports Shibboleth federated authentication and authorisation with eduPerson attributes, integrating with the Shibboleth service provider software.
Fez has a batch import system that could be extended to cope with feeds from a Library catalog system (eg MARC record import). It already handles METS, ePrints and file import.
It really depends on what interface the 3rd party software offers, but usually they are application specific so will need some specific tailoring.
[edit] How does Fez deal with virus checking?
Fez currently doesn't have a virus check workflow on ingest of files, but we are considering it for a future release with something like ClamAV. This would act in the same way as the JHOVE and ImageMagick workflow webservices.
After a quick google it looks like DSpace might do something like this in the future as well: (Link to a DSpace Sourceforge forum entry on ingest virus scanning)
[edit] Does Fez support alerts for new content (either through email or RSS)?
Not yet, but its something we have on our wish list for things to do before the end of the year.
[edit] How does Fez deal with federation and harvesting?
In Fez 1.2 you have the option of configuring the Fedora OAI provider to provide access to your content. We are aiming to release a Fez OAI harvester in 1.3 that uses FezACML security and OAI sets to provide better control of the OAI provider.
Also in 1.3 will be a provider for Picture Australia and Australian Digital Theses for their harvesters.
