michaelmarth said...

My understanding of the spec is that it is geared towards classic document management systems and their requirements. In fact, the spec explicitly rules out *web* content management as a use case. Therefore, it is understandable that abstractions like e.g. "node" in JCR are not found in CMIS - they make no sense for a DMS.

OTOH I don't get why this DMS focus of CMIS is not discussed more.

Cheers
Michael

I was reading this comment for Michael Marth this morning and I could not agree more.

CMIS (at least the v1.0) looks like focusing more and more only on binary file management. In the latest spec draft "policies" (one of the four cornerstones of the first published spec draft) look like being replaced by a notion of ACL and all prototypes we see here or there are mainly focused on file exchange, file listing, cross-repository file searching, file directories; etc....There are not a lot of prototypes or disscussions out there about dealing with more complex object types in between two content repositories.

The fact is that there is already currently a clear business need to ease binary file management among repositories. Old File Systems or the well-known shared network drives (NFS- CIFS/SMB - ...) are being rapidly replaced by more advanced "File Content Repositories" (FCR - A new paradigm for the ECM industry?) which are offering a few couple of additional value added compared to the previous shared file systems (e.g: file versioning support, check-in/check-out, Properties on files; Relationships;...). All these features are looking like becoming rapidly a commodity and will be probably soon preconfigured and preinstalled with the Operating Systems themselves without having to buy another thrid party ECM solution.

So Simple Shared Network Drives are dead; Welcome to the new File Content Repository driven Drives!

Now will CMIS only act as a kind of standardized vendor-neutral replacement for CIFS/SMB (or WebDAV/DeltaV which never really became a strong standard perhaps due to the lack of active MS support behind it and the delivery of a really buggy WebFolder implementation) ? Or will it evolve in the future releases to manage more complex content objects? Shoud we rather call version 1.0 FMIS (File Management Interoperability Standard)?

In all the cases, even if v1.0 of CMIS will only or mainly deal with binary files, it is already something really positiv for the IT industry in general as too much binary files are currently locked in proprietary file repositories.