Cosmo Vision Brainstorming

• Led by Ted Leung
• Notes taken by Sheila Mooney

Goals

• Capture what BCM and Jared are thinking as far as vision.
• Don't restrict stuff to target users.
• Talk about stuff beyond Beta.

Notes

• BCM - Dual nature of Cosmo goals. There are the sharing target end users and the data hub target users. These goals are not always aligned.

Server - Requirements/Features (Jared)

• Backup (Beyond storage layer tools)
• Restore (Beyond storage layer tools)
• DB Indirection
• Logging - more levels
• Admin tools
• Web session clustering
• User quotas and space usage reports
• Routine metrics - accessible via JMX
• Load balancer integration
• Branding updates
• PROPFIND optimization and other perf = desktop sync perf
• Mode to disable user access
• Disable a single user
• File type upload restrictions (configurable)
• 2 accounts with same email address
• Configurable password policy
• Server based throttle of sync (requires desktop co-op)
• Separate user profile db.
• Run Cosmo UI against non-OSAF CALDAV server
• Support for scripting up reports
• Account activation
• Catch up and implement CALDAV final draft stuff
• Postgres/Oracle support - SQL Lite
• CMP client side libs
• Deprecate CMP
• Data preservation across updates
• Scheduling draft spec - CALDAV
• Alert framework for alarms
• PDF's of calendars
• SychML^?/Mobile
• Supporting users who don't want to install MySQL^? Postgres - Filesystem DB/HSQL
• Rebranding a Cosmo
• Ads
• Hackable UI
• Mobile UI
• Don't crash
• XSS Analysis
• Don't hang
• Suck less

BCM's List for Cosmo Server

• There are a number of features people externally are asking about - when we will support this

• External directory integration (LDAP, Kerberos (Apple has these))
  • Auth only, profile in Cosmo
  • Profile in the directory
• Integration with single sign-on/digital ID
  • Shibboleth/SAML/Library
  • OpenID^?
• The 2 item above imply some kind of framework in place to do this

• ACL based security (Apple has this)
  • Would also need a UI for managing these
  • WEBDAV ACL extensions
  • A UI which is part of Cosmo
  • WEBDAV principle support - Kervin - use Cosmo as a poor man's address book
  • Handling groups - implies existence of group calendars
• CardDAV^?
• XMPP support
• Import existing data
• Run Cosmo/Snarf as a windows service
• Portlet/portal integration
• Drag and drop microformat data onto the Cosmo UI
• Live clipboard/microformat
• More MyData^?.org integration
• Google calendar support
• GData/Atom Pub support (BCM thinks this is more important than most items on the list)
• Federated Free/Busy - early stage
• Storage layer research
• Integrating admin interface and file system browser into now Cosmo UI - AJAXy
• Remote calendar subscription
• "Virtual Cosmo" - partitioning a single instance
• WEBDAV DASL
• More efficient synching - diffs, less HTTP transactions
• EIM/morse code

Mikeal's feature requests

• Ability to delete all the data - clear repository action

Target Users

• Who do we think will be running Cosmo and hosting a service?
• Is this is a big group? Small org?
• What do these people need?
• What are the different types of users?

• Service Integrator
  • Need to build the infrastructure for these people
  • Personal information portal
  • There are a set of users who might install Cosmo on their own but not really be developing
  • Somebody might write some small clients for Cosmo. Our target users would in turn use these others pieces.
  • Data on Cosmo can show up in other types of clients
  • This person could be described as an API consumer
  • Example: public API for Flickr - write another UI for publishing pictures (Cosmo would be like Flickr in that way)
  • We can also think of this on the web client side ie: like make it easy for people to write firefox extension
  • Mimi: Why would you put stuff on Cosmo and just not on the other server?
    • It's a central place to put stuff
    • We can share with others using the ecosystem
    • Hosted service is available 24x7
    • Means we can support other interfaces for getting at data even if they aren't there - plausible promise
    • We can have data that lives in other places and put it into Cosmo
    • Make Cosmo the central data source
    • They might be using Chandler or may just be using websites
  • Mimi: Doesn't think the primary usage of Flickr is to go and look at them somewhere else
  • Give people a way to take the data from Cosmo and look at it somewhere else
  • Cosmo would be a repository for all kinds of data and we can look at it from lots of different UIs
  • Bear: if we supported voicemail - we could push this data to people's Cosmo accounts and they can access it from anywhere. Target small business, vertical integration
  • Different types within this group - univesity vs sm business
  • Not clear how the GData stuff translates to a specific target user
  • We acknowledge there are target users within this group
  • Mimi: Need to think about what we need to do on the desktop side to encourage people to even put this kind of data on the server

Jared's priorities for what should be in Beta

• Jared: What happens in Chandler if something fails and we restore a backup on the server. Does Chandler just sync and you have a version from ie: last week.
• Backup/Restore should be a priority
• DB indirection
  • Might need this if we were a certain size
  • We should probably have this
  • Shouldn't be doing this between beta and 1.0
  • For Beta we are preserving data between releases
• Web session clustering
• Load balancer integration
• Data preservation across updates
• PROPFIND optimization and other perf = desktop sync perf
• Mode to disable user access
• Disable a single user
• User quotas and space usage reports
• Don't crash
• XSS Analysis
• Don't hang
• Suck less
• Account activation (support emailing to collections)
• CMP client side libs - Python
• Deprecate CMP

BCM's take on priorities

• WEBDAV principle support - Kervin - use Cosmo as a poor man's address book
• Import existing data
• Google calendar support
• GData/Atom Pub support (BCM thinks this is more important than most items on the list)
• Run Cosmo/Snarf as a windows service
• Storage layer research
• Integrating admin interface and file system browser into now Cosmo UI - AJAXy
• EIM/morse code
• CardDAV^?
• CALDAV final draft
• Account activation

• Potential Help Us Items
  • External directory integration (LDAP, Kerberos (Apple has these))
  • Integration with single sign-on/digital ID
  • ACL based security (Apple has this)

• BCM - for 1.0
  • Live clipboard/microformat
  • MoveMyData^?.org integration
  • Alert framework for alarms, incl XMPP
  • Mobile UI in some constrained way
  • Runtime metrics
  • CALDAV scheduling
  • Remote calendar subscription
  • Groups

Comments

• Mimi: PPD needs to have a better understanding of who and how these features will be used
  • Understanding why people would want to do things this way
• Mikeal: Why do we need Snarf as a windows service
  • Good citizen of the windows community
• John T: Many things on Jared's list that depend on the number of users. How big is the hosted service.
• BCM: Service will be organically evolving system and this will change over time.
• Mimi: Question about framing. Are we assuming that the vision of Cosmo as a server product ties back to the original OSAF mission? Relating this back to small workgroup collaboration that doesn't have dedicated IT. Feels that many of the features on the list are more suitable for larger group scenarios. When we figure out what we want to target for 1.0, what is the bar we want to measure this on.
• BCM: Some of the stuff is not small workgroup stuff ie: GDATA. We have an opportunity to innovate here and it's really important we do this. It makes us competitive.
• Ted: Don't really have a concise statement for the Cosmo mission. The OSAF mission statement doesn't only include small workgroup stuff.