Chandler Westwood Advisory Board -

Chandler Update

overview of last 3-4 months

0.2 released Sep 2003

OSAF reorganized into 4 working groups

0.3 scheduled for Feb 2004

Heikkig Toivonen - repository dev mgr

mimi yin - UI designer

Canoga Planning

- Canoga will have baseline PIM features with a few well chosen 'cool' features

- using "best of extreme programming model" - rapid iteration

- trying to contain risk of product development

- strong bias toward paring down feature list rather than lengthening canoga developments

- focus on four major app areas: email, calendar/tasks, notes, and contacts

- IM will be provided in a supportive role - im capabilitites will be in chandler, but mostly to support the other four areas. No video, no multiple buddy lists. But buddy lists will be integrated with contacts, presence management will be supported, etc.

- Prioritized features found on OSAF wiki topic: ProjectOverviewTable

- better update on canoga timeline when dev process is sorted out.

Four Working Groups

- Design - Chao

- Apps - Katie

- Repository - Heikki

- Community - Mitchell

- Mitch responsioble for high-level intertroup issues

Dev & Planning process

- aim for dot release one per qauarter

- 6 or 7 milestone releases within each dot release

- design team proposes major stories or features for each dot release

- apps and repository teams responsible for fine-grain schedule estimates and implementation

- design works with apps and respository to finalize planned scope and schedule

- rapid iteration and recalibration esp. at eatch milestone. monitor progress with 'demos' at each milestone.

0.3 release

- base architecture frmework in place:

- chandler presentation and interaction architecture (sits on repository, under UI)

- parcel loading - 3rd party providers can provide

- repository enhancements including transactions and multi-threading

- agent and notification framework

- content model

- unit tests framework and real tests(!)

Initial "caterpillar" UI

0.4 release (sometime around Spring CSG)

- goal is to be experimentally usable for a few key end-user tasks

- next rev of UI landscape

- generic information management capabilities

- basic calendaring

- users and groups framework

- elementary sharing (calendar sharing will be probably first demo)

- access controls

- queries and searching

Design Team Roles

- requirements gathering and rationalization

- user interaction and interface design

- product scoping and prioritization

- feature specification

- user testing

Design team: near-term deliverables

- key usage patterns (to elucidate key features that are required)

- UI landscape

- end-user content model - what are the end-user concepts of teh schem

- users and groups design

- sharing design

- security design

- initial major capplets (chandler applets) scoping and high level spec

Apps team - near-term deliverables

- Caterpillar UI - really uses repository, metadata framework, etc.

- Repository Viewer

- RSS reader

- Infrastructure to support UI goals:

Caterpillar UI is starting point for non-siloed UI design - not separate UIs for calendaring, email, etc.

We need a "steady trickle of ever-evolving sales artifacts"

Repository Team Roles

- support app development

- build and release management for chandler (hiring a release engineer and a qa person)

Repository team: 0.3 deliverables

- basic low level repository features completed

- unit, stress and perfomance tests almost complete

- repository documentation and data model

- preliminary querying and indexing

Repository team: 0.4 focus areas

- Queries, indexing, and searching

- access control

- sharing

- authentication

In the midst of having success integrating full-text search into repository - looked at Lucene - issue is that it's written in Java, and they're a Python shop. getting it to run on all three platforms, etc.

Canoga security update - remember, this is *not* Westwood security

some security decisions for canoga

- user authentication - user name and password

- data not sent in the clear over network

- passwords not stored in the clear locally

- rely on OS to encrypt local repository (aprt from passwords)

- access control (details still evolving)

- prevent propagation of malicious code

Each of multiple users on a single machine will have their own repository.

When you create a share you'll be able to decide what's being shared, with whom, and who has administrator rights.

Per-attribute access control will be supported at the repository level - how that bubbles up into the UI is a current topic of discussion.

Mitch- "my reading of the history of Internet based innovations has been that the early generation of instances of things have not focused on security but have focused on behavior - like the wiki - people find counterintuitive - OSAF's wiki has each page globally writable, but that hasn't been a problem. One of our considerations with Canoga is to focus on the sharing affordances and give people lots of flexibility, and then we'll learn a lot from that."

Sharing Sample

- decide to share view

- select contac, permissions,set/genearate password

- send chandler url to contact with password (email, IM, ...)

- contact receives chandler url, clicks it, and logs in with password

- contact can see view, change password and do permitted ops

Calendaring

Goal is to build product around accepted open standards.

CAP has not been ratified - discussion may have stalled in the calsched working group.

Pieter talked to Nathaniel Borenstein about the state of CAP yesterday. He talked with Pat Egen

Dave Thewlis, is co-fouunder with Pate Egen of the Calendar and Scheduling Consirtyuom - www.calconnect.org - a new nonprofit.

Pieter talked with Marten at Oracle - still on record as haveing Oracle Calendar evolving to CAP compliance

Lisa Dusseault, CalDAV proposal author - works for Xythos

There appears to be a good impedance match between webDav as a protocol and the needs of Chandler for repository replication, which makes calDAV of greater interest...