Hi Niels,
Please allow me to join your interesting discussion, which my son brought to
my attention in forwarding a digest of it.
I've been struggling for a long time with similar thoughts on how to handle
the daily flood of information on my KDE desktop. There is one idea in my
mind for some time now, I like to share with you.
I'm thinking about a common working cache to keep all data items (objects).
>From the user perspective, such a system could be something like a 'Document
folder' which is accessed via a kio-slave. The storrage backend could be any
presistent system, such as: A file system, a Relational Database or even an
Object Database.
Associated with each object, the following minimal set of meta data
should be
stored:
- a unique identifier, automatically generated - like a hash
key (md5)?
- its mime type
- its URL of origin
- time stamp
- valid until (VAT)
- user definable attributes like a file name, message ID or
annotation -
(search criteria)
- to be processed flag (TBP)
- priority ID (PID)
- derivation flag (DF)
- link ID to its predecessor... (for versioning)
So for example, I select an image within my browser to save / modify.
- It is already in my cache.
- so all I need to dom is activate it for my TODO list by changing the
flags
'TBP', 'VAT', and PID.
At a later stage, I could change to my TODO list, which is a just a "View"
into my object cache based on VAT,TBP and PID. - By selecting the picture and
activating the associated application tool to change it, a new object would
be created within the cache system,
- using a local URL
- with an activated DF
- a URL link to its origin - which would either point to the
original cached
object or the URL of origin, in case the cached version has expired...
Another example could be a web page, currently presented in my browser. This
is a container object with all its different parts (objects) already stored
within my cache. - Now, I imagine I could do the following tasks:
- displaying the HTML source
- creating a local archive version
- forwarding a copy to another viewer/editor either direct or using my
TODO
list
- forwarding a copy via email
All these tasks could be initiated without creating interim files or reloading
the objects from its point of origin.
The same concept could be applied to any object, either locally created or
retrieved via any URI path.
I believe that such a generic data cache could not only lead to more intuitive
experience for the end user, but open up new ways to archive, replicate or
encrypt user data.
Is that something worth to follow up?
Regards,
Rainer
_______________________________________________
Appeal mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/appeal
