My name is Inigo Montoya. You killed my folder. Prepare to die.
Inspired by @piewords latest post, this is my contribution to the banter around whether or not folders should be voted off the island. You’ll find background in posts by Sean Hederman , Chris Riley and a cast of thousands in the comments. I thought this was a little long for a comment but I suggest you check them out. After reading it all though my take away was “Folders are familiar,fat,old and slow. Search is better but you probably need both to keep the old people happy.”
Folders Aren’t Born Bad. They Get That Way From Misuse.
I’ve said many times – buying a product is buying a collection of decisions. So much of this debate is rooted in biases for and against the CMS products and design decisions baked into them. Consequently one products manner of implementing the construct shades all of our opinion one way or another. No one is objective here – that said…
Somewhere buried in these thoughtful comments is the idea that folders are really just representations of a class of metadata applied to an object that permit collective action on that set. The collective action whined about (I mean rationally discussed) here is navigation but it could logically be applied to any persisted container model. The problem arises when this or that product requires the manifestation of this container model, creates a user interface dependency on it, and some designer somewhere is offended because it prevents him from animating the breadcrumb or some other random programmatic feat of daring do.
My observations are simply that
1) you can provide content management capabilities without implementing the folder containment/navigation paradigm.
2) real people understand folders – even if you consider them as lame as having a aol email address
3) content creators may want to navigate content using a completely different paradigm from the people consuming it. Don’t assume both communities agree on the usefulness of folders.
4) the fact that a products implementation requires folder metadata does not necessarily mean the user experience is bound to it.
5) all content management system MUST have some manner to persist containment/collections else things like security become impossibly complex.If not folders then something and you probably won’t like that either.
For whatever reason some ECM implementations have folders baked into the design. It was a choice. Don’t like it? Pick a different one but realize that what comes with it are all of the other decisions that went along with it. In the end, working around a given ECM’s use of folders is no different that learning their API or internal table structure or security model and may be less work than compensating for the other differences. But be warned – do it poorly and the implementation will fail but that doesn’t mean its the folders fault.
People love to containerize…until they forget which container they put something in and then they search for it. The Interweb is one big collection of containers…a search for anything is an instant container. ‘nuf said.
It seems that this only became an issue because of certain vendors rationalizing their lack of a natural folder/container implementation. I’ve had the misfortune of managing content in an environment that did not have a decent container mechanism. The vendor proudly claimed that folders were outdated and “prescriptive”, whereas their alternative (tags and searches) were modern and “social”. But as Lee points out, the system could not scale up to manage things that become critical for lots of content and lots of contributors, like security/ownership, notifications, and retention management. IMHO, Documentum gets it right. A folder view can be easily implemented in the UI to make management tasks that rely on a strong container model easier, and to provide navigation for those users that are comfortable with that form of content organization. But in the underlying object model “containership” is just a relationship type and there are many other ways to navigate to the content, in addition to searches.
The question of whether or not folders are relevant to Content Management is an interesting one. In our house, we have cabinets and shelves that used to be neat and orderly when we first moved in. While we were living in there, they became disorganized and unmanageable, so we got containers and labels to sort the things in the cabinets and keep them accessible for our use. We kept tagging and searching to make it sensible and manageable and reorganized the stuff in the existing or in the new containers. That is like a database. It uses indexes to make the contents searchable and it reorganizes constantly in order to speed up the performance. Without indexes, the database would be pretty slow and might be unusable all together. We earthlings need some sort of containers to make them sensible and functional whether they are “Folders”, “Briefcases”, “Cabinets”, or whatsoever. Without the measurable containers in place, any Content Management System will be unmanageable eventually and won’t be scalable.