Locked-down vs Free-form folder structure

I have encountered this question on every project I have been involved in.  On one hand, having a locked down folder structure promotes standard navigation and storage.  On the other hand, coming up with a folder structure template that everyone will understand and use is quite difficult.

On one project I was on, we convened a “governance committee” of about 10-15 people from the different business units to come up with a standard folder structure for the entire company.  It took us about four 2 hr sessions to get to create a template structure.  Each of the reps then presented the template to structure to his/her business unit.  Let me say that, this was not the end of it.  We had to have another 2-3 meetings to resolve naming schemes and whether we should add some additional sub-layers.

In the end, the governance committee agree that we would only impose structure at tier 1-3 levels.  Anything below Tier 3, would be free form.  A good software design usually requires input from all types of users and ultimately a solution will contain compromises between different group.

Advertisements

7 responses to “Locked-down vs Free-form folder structure

  1. Based on your experience, what is the best way to implement a lock-down of a folder structure? I ask because you’ve probably seen and field tested various methods.

    I would imagine that the most out of the box approach would be using the built in security model to disallow writes in lockdown areas. On the other end of the spectrum, enforcing the business rules in Webtop actions would leave the door wide open for any other application (or at the very least force duplication customization which has costs).

  2. I agree – using the native security on folders (ie give only READ permissions) is the easiest way (no development required) to lock down folder structure. If customizations are required and multiple application support is a concern (ie both webtop and desktop apps are deployed), you can create TBO (Type-based Business Objects) to enforce business rules. TBOs sits on top of DFC and can be deployed to both webtop (app server) and individual DFC applications (eg desktop).

  3. Have you seen google mail? No folders at all, and works great!

    My idea is to have a dual structure – business unit based for creation and modification purpose and taxonomies hierarchy for all who access the information.

  4. Google mail works great because the user controls how/when labels are added. This is not conducive when sharing documents with other people.

    Microsoft Sharepoint used to have a non-folder structure. However, a lot of users complained and a lot MS Solution Partners created customizations for folder creation that eventually Microsoft included folder paradigm in the latest Sharepoint version.

    I agree that taxomony hierarchy would make more sense. In order for this to work, you have to have a taxonomy to begin with. Webpublisher uses a taxomony hierarchy for web content storage, but this requires creating folder mappings for different content attributes.

    In the end though, a standardized folder structure can be thought of as a rudimentary taxonomy.

  5. I completely agree that people see it “strange” when there’s no folder structure, but personally I think that this is rather a matter of habits then theirs real need: a user must say WHAT is the document he/she is creating, but not WHERE it should be kept.

    The way the taxonomies work is quite different from the folder: I must navigate to the leaf folder to see the document kept there, but I’m cutting the set of documents I can see when navigating through the taxonomies.

    What I see is that Documentum is not optimized to work with such a structures because you need to make DESCEND queries on the taxonomies(i.e. folders) hierarchy.

  6. Good point. People do have bad habits that are very difficult to overcome at times.

    I agree with you on the implementation of categories in Documentum. To me, they should of designed using some sort of tagging hierarchy. This way you can search directly on the tag vs traversing folder tree.

  7. I agree with both, and I think both have a place. When you are dealing with millions of documents or a system that can grow to that size it is a good idea to recognize and have designed a system that would locate by metadata in other words a search paradigm rather than a folder paradigm, the sear volume located in a folder(s) would make it unlikely to effective located content.

    Metadata allows you that if the paradigm or window into the system is based on the metadata provided.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s