Forum OpenACS Q&A: New project CMS, KM etc.

Posted by Ciaran De Buitlear on
    We're early in the design phase of a new project for a large government  body.  We will probably be using the latest version of the OpenACS, Oracle 9i if possible.  The main requirements are as follows, quite roughly.

1) Content management - version control, workflow, content approval, metadata storage,  permissions, workflow.
2) Knowledge management - we're really trying advanced searching (complex searching including concepts etc.  mostly hidden from the casual user), taxonomy of concepts / keywords, Browse type search involving the taxonomy or part of,  user "categories of interest", some kind of "push" of relevant documents to users, experts database - "who knows what" ,  Find "content like this",  My bookmarks / share bookmarks, link content / share links.
3) Multilingualisation  - fairly standard "label multilingualisation " content in one language at a time.
4) Workflow - Not strictly content related.
5) General collaboration:  FAQ, bboard, file storage.
6) Portal type easily customizable interface with news and other feeds for "stickiness".  Some kind of access to e-mail here also would be desirable.

Off the top of my head I imagine we'll be trying to get the Openacs 4.6.1 working with Oracle 9i, and all the obvious packages.  Where I see us extending the toolkit would be in the areas of:

a) Advanced searching - not sure what to do yet exactly but looking into it.
b) Taxonomy - Try to leverage functionality from the following features in Oracle text:  Knowledge base - make visible and extendable to the user by using their browser, use some of it for interest categories etc, Section group, Thesaurus, Lexer, Relevance ranking.
c) Look at interoparabilty between Oracle 9i and openacs - Portals, single sign on / LDAP etc.
d) Portal - we may want to add some functionality here.
e) Experts database.
f) "Push" - a bit like "new stuff".

Anyway most of this is just thinking out loud and it's very early days.  I will keep you posted if you're interested and I'd be grateful for any comments or if you just want to watch this space...

Posted by Don Baccus on
There's already work underway to resurrect LDAP authentication.  It won't use the built-in Oracle hooks because it needs to work for both PostgreSQL and Oracle.  You can e-mail Lars at Collaboraid for more details.

Categorization ... there's some categorization stuff available already in the content repository and there's work underway do do a more general categorization package, check the recent thread in regard to this.  Folks should organize and work together if possible, though I realize client deadlines can make this hard or impossible to do.  There's no reason to duplicate effort if we can avoid it, and general categorization is badly needed.

Posted by Ciaran De Buitlear on
That's great - I'll follow those pointers.  I'm keen to extend the OACS rather than repeating other people's work.  I will keep this thread going with more detail as we get a better idea of what we're at.
Posted by Don Baccus on
Also check on threads here discussing the existing CMS (UI sucks, much of the included functionality does not) and the content repository.  The CMS/CR allows for the attaching of templates to content types or individual template items, automatic form generation and handling (only useful for simple page flow and content input but better than nothing) and other stuff.  Jeff Davis has been working hard to make these work as they were originally designed by Ars Digita, in both Oracle and Postgres.

Also if you're not already familiar with the templating system's form builder package, study it and the ad_form wrapper function we now use to write self-submit form generation and handling scripts.  Makes things much simpler.  The autogenerated forms mentioned above also uses the templating system's form builder package so you can customize the look of input forms site-wide by writing one new template file.

Workflow - we have a new, simplified workflow package courtesy of Lars Pind, who wrote the original acs-workflow while at Ars Digita.  The new one uses a finite-state machine model rather than petri nets - FSMs are sufficient for the simple workflow requirements we've run into and relatively few people are familiar with Petri nets.

Posted by Talli Somekh on
Very cool! Congrats on the project, Quest!


Posted by Danielle Hickie on
Hello Ciaran,

I think it might be nice if the client has a choice of how the content lays out. You mentioned a portal-like layout potential. I think a good content management system would allow an authorised user to associate the inital content to several other types of content.

For example a news item might have three pictures, or none. It might be linked to other news stories. It might be linked to a critical document that has just been published.
A news story with three pictures, just for example,  would imply that instead of just designing one portal like template, you would design a more flexible system that could respond to the type of content.

Don and Lars can tell you about the best functional way to do these things. I've just had alot of experience with what clients want. Our people want alot of flexibility with the design, but thats because we are an extremely reactive bunch, used to changing our entire homepages daily. You might find that being in a Goverment situation, you'll need to change the front pages less often, but you'll get alot less minor hassle if you allow for more flexibility.

I hope this helps - it might seem a bit obvious :)

Posted by Don Baccus on
Portal-based is a great way to go, and if our new portals package was stabilized with a good UI I'd actually suggest you  guys think about taking that approach, Danielle.  Converting Greenpeace's panels and story snippets into portlets would be quite easy because those things are implemented on your site with included templates - which is what the portal system does for you.

I'm planning to dive into cleaning up our portals stuff relatively soon, that's my plan at least ...

Posted by Rocael Hernández Rizzardini on
Hi Ciaran, great news!
Yes, it should be done in a Portal-based form, that's the way to go.
Are you planning to discuss openly with the community about the Content management & Knowledge management?
Both a really important issues for the future of openacs, I strongly suggest that the core things should be widely discussed here.
Posted by Ciaran De Buitlear on
Thanks for the feedback so far.  What I plan to do now is to work on detailing the client's requiremnets a bit better and also get the team to start to look at the features of OACS that I feel will be most useful.  I think we will be almost certainly be looking at expanding functionality in some areas.  I would like to discuss the issues that come up openly with the community, and there does seem to be some interest, so I think that's the way to go.  "Watch this space" as they say...
Posted by Frank Bergmann on
Hi Ciaran,

May I ask you how your project is progressing? We are currently setting up a KM group at La Salle university in Barcelona, and a mulilingual search engine would also be one of our first goals.
However, the second goal would be a "personalized" search engine, using Bayesian networks for selection purposes and a lot of other modules.