Overview

This page lists a set of high-level features we should consider dropping from Kibble. It lists pros and cons to try to make the best decision.

Mitch has argued that in our dog food focus, we need to really focus on just the basic features for Canoga. We need to ask fundamental or even outrageous questions to make sure we are planning on just the ideal minimal set of features.

This page is about dropping big chunky features, or at least turning big chunky features to small implementable ones. We have a separate effort to thin down features at a much more fine-grain level. These are not addressed here.

Candidates

Chandler as a full-service Contacts Management application

Proposal Drop contacts management as a mini-application within Chandler i.e. no separate helper window of contacts, no contact support for phone numbers, addresses, birthdays, etc. Instead remember relevant email addresses and have good polished UI to provide email address auto-completion where appropriate.

Beyond Kibble, we focus on integrating native OS address books and LDAP with Chandler.

• Pros
  • Contacts are increasingly becoming integrated or embedded as an OS functionality. Users don't want to have a different address book for each application. They want an integrated address that is common to all applications. Apple Address Book serves that functionality. Microsoft has promised similar integration in Longhorn (with a built-in myContacts folder), although there is no de-factor address book in Windows XP. JuergenBotz suggests LDAP integration for Linux (which is a functionality we need to provide for Westwood anyway).
  • In our design explorations, we have discovered that contacts are not regular processing items (such as messages and events) and are more like long-lived but less voluminous resources. They are thus not directly contained in Item Collections and less central to Chandler.
• Cons
  • It is often useful to identify the person rather than just the contact address (e.g. email address). For example, you may want to search all email from "John Smith" who has three known email addresses.
  • We may need to associate a person's WebDAV account and her email address
  • It is useful to associate a nickname to people who appear frequently

Instant Messaging

Proposal

Instant Messaging (or technically XMPP-like technology) have :

• Synchronous direct communication with one or more parties
• Presence management for identifying whether a user is online and integrating this knowledge to contacts management

We propose dropping these two features for Kibble. This does not preclude using XMPP for infrastructure services such as event notification

• Pros
  • There are many IM clients in existence, many with features and polish that will exceed what we can provide in Canoga
  • We have not found compelling use cases for integrating IM with Chandler, except for features we're considering dropping here (e.g. presence info with Contacts and interleaving with Item Conversation workflow)
• Cons
  • We have promised this feature

User Modeling

Proposal Keep track of items sent to and from the user. Drop all remaining more complicated user modeling such as organizer and invitee for Events and requestor, requestee relationships for Tasks.

• Pros
  • Simplifies user mental model and required accompanying user interface considerably
• Cons
  • Advanced user workflows such as collaborative task management will not be directly supported in Canoga

Generic Viewer

Proposal This was intended to be a mixed viewer capable of displaying PIM items and third-party items co-mingled. Proposal is to drop this functionality.

• Pros
  • We need to test and explore several third party capplets and parcels before we can really design a decent generic viewer
  • We have not put in any design work, the issues involved can be pretty challenging
• Cons
  • It is hard to validate Chandler as an integrated platform without this viewer.

Collections/Attributes Manager

TBD

Item Browser

Proposal Drop item browsing features. Rely on sentence-based rules and search instead.

• Pros
  • Item browsing is a new and untested feature, catering to mainly very advanced users
• Cons
  • Our sidebar doesn't allow for nested collections to provide for user hierarchy. Item browsing was intended to fill this need.

Calendar Month View

Proposal Drop calendar month view. Beef up mini-calendar to cater to use cases that would require the month view. We should discuss this in a separate Calendar Design context.

Sharing Circle and Certificates

Proposal Drop concept of sharing circle and certificate-based peer trust relationships. We should still use certificates for secure TLS transport

• Pros
  • Our more server-centric approach (and move away from pure P2P) to sharing lessens the utility of sharing circle.
  • WebDAV server based approach suggest using WebDAV-native authentication methods as dog food.

Sharing with non-account holders

Proposal Do not allow users to share collections with users who do no have an account on the same WebDAV server

• Pros
  • This feature requires significant modifications to existing WebDAV server
  • Our Canoga hosting plans allow for all users to be on the same Canoga hosted server
  • Users can rely on single item sharing to collaborate with non-account holders
• Cons
  • This is a compelling core feature for our Canoga target audience: small workgroups with low reliance on sys admins

Item Conversations

Proposal Rather than the rich IM-like item conversation feature, we simply have a shared note field, for sharers of an item to comment on the item itself.

Documents as First Class Items

Proposal Support email attachments, but do not allow them to be promoted as first class Chandler items. We will at least get the infrastructure and architecture for this feature build into Kibble under advisement from CSG. Looking into options to get community leverage to demonstrate this capability

Help

Proposal Provide very limited online help. Not a focus in Kibble

Browser-like Navigation

Proposal Drop browser-like navigation affordances in Kibble

Help

Proposal Provide very limited online help. Not a focus in Kibble

Internationalization

Proposal No proposal but one sorely needed. At the least we need a set of explicit i18n goals for Kibble.

At a minimum, we should continue to provide good software development hygiene around i18n issues. But should we build tools and provide features to facilitate internationalization?

Items removed from 0.6 and 0.7

• Email Features
  • Email centric-search - to and from fields
  • POP support
  • HTML rendering
  • Spell-checker
  • Email thread aggregation

Outrageous Questions

These are questions probably worth asking but they are outrageous because they go pretty deeply against how we've been thinking about Chandler and how we're going to develop it.

• Should we do cross-platform parallel development rather than focus on one platform first?
• Is attempting to be both a platform and an application at the same time too difficult?

-- ChaoLam - 13 Aug 2004