I need your opinions and suggestions about storing documents in a DMS system. I think the DMS should save its own copies of the documents, not their original path on the disk. So, the DMS should have its own space to write on.

But what about the way they're stored? Do they have their own extension different from the original one? What about the algorithm that stores them? What about the algorithm that retrieves them? What do you suggest??

  • What do you consider a DMS to be? Is this comparable to a CMS or respository? First thought that comes to my mind when you say DMS is something like sharepoint. – Chris Apr 9 '11 at 11:28
  • 1
    I assume by DMS you mean a "Document Management System"? It would probably help if you could be more specific about what type of "space" you envision your DMS using for document storage. – vjones Apr 9 '11 at 20:46
  • What kind of documents? A software system for storing "design documents for software"? If so, you might want to look for "software requirements management system" instead. – rwong May 6 '11 at 6:17
  • Who will be the everyday users, maintainers (IT) and stakeholders in the system? – rwong May 6 '11 at 6:18

I'd give every document a unique file name, perhaps based on a sequential number and store it in a hierarchy of numbered folders. Location of any actual file in the structure is based purely on the time that it was added to the DMS. No extensions on the names. No meaning to the file names.

All of the metadata would go into the DMS database. Owner, group, mime type, file name and extension, date created, categories and keywords -- everything in the database. Oh yeah, and a reference to the disk location with the actual file goes into the database too.


After you stored the document in the DMS, do you need to know where it originally came from?

You can just assign a guid or some other id to the document and make it retrievable through it.

If you want to know the original location this can be in some metadata field for the document.

Just using the filename as an identifier would be unsafe, you might run into collisions unless you mimic the original filesystems directory hierarchy. Using the full path plus filename would work.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.