Selective Download of folders/files

[careck]

Hi,

as I am trying to test collanos on a project with a 'bandwith challenged' person, I was thinking how useful it would be if a user could select certain folders inside a workspace to be 'updated manually'.

E.g. if I have a workspace like this:

Projects/
             P1/
             P2/
             P3/

Someone might only be interested in files under P2/ and could then chose to sync P1/ and P3/ manually when the need arises.

A prerequisite to this would be a kind of preview function which would present the following prior to actually performing the synchronisation of a workspace:

* which files will be created as new
* which files will be updated
* how much data will be transferred in the sync
* how long would it take to complete the sync (based on earlier measurements)

If a user could then go through a folder preview and pick and choose which folder to sync.

I see we are now moving into SCM territory

Cheers,
Carsten.

[Franco Dal Molin]

Hello Carsten, thanks for your input and suggestions!

Generally, yours is a desirable feature from the affected individual's point of view, but needs some careful considerations. If this is implemented to rigidly, it could have important downsides! Let me explain: Let's assume we would have a pure "on-demand" synchronization. Only content creators would store the content, all others would wait to "fetch it later". Now, what would happen if the original creator wasn't online (e.g. left for a long vacation :-)) just when somebody else needed/wanted the content? It won't work! Ok, if only few members within larger groups would *exceptionally* switch this option on, combined with clear warnings about the consequences, then it might work.

On a related note, we plan to further improve the replication system in various ways in the coming releases. Currently content is "fetched from any" member peer, but without any optimization. If fact, a peer could fetch content from a far-away low-bandwidth peer, while another member peer sits idle in the same LAN with very high bandwidth. In the near future, we will begin to measure the bandwidth between any two peer connections, and fetch it always from the fastest "source".

Another way to address this issue, is to introduce synchronization priorities and temporary pauses. Imagine you have several workspaces, and somebody just loaded many GB's worth of pictures into a Holiday space, while you are waiting for an important document in an urgent project space. You would want to mark the project space as "High Prio" and the Holiday space as "Low prio", or even pause the replication for a period of time. My point here is: Often it is not about "not wanting" some content, but rather about "not wanting it right now". This solution would not have the downside outlined in the pure "on demand" scenario.

Finally, users want transparency in general, as discussed here: http://forum.collanos.com/index.php?topic=1372

So, long story short, I see the following elements that need to be addressed - more or less in the following priorities:

1. Transparency about pending content, synchronization volume, and estimated time.
2. Priorities and settings, to better control the "when" and "how" of replication traffic
3. Optional limitations to set, blocking content based on some criteria

One final thought: If you have really heavy documents, and/or very many of them, and/or need them only occasionally, you could consider combinations of solutions: You could put the heavy stuff on some online-document-repository, and share only references (e.g. link objects) into the workspaces.

Best regards,