OSAF Status Update # 1 --February 26 2003

0.1 Release

OSAF staff has been focusing on defining our initial release of code, which we've named the 0.1 release. We've been working to figure out what the 0.1 release ought include, what code is required to make this happen, what APIs ought to be described, which areas will not be addressed in this release (such as security), and the steps we need to take to make this release happen. The 0.1 release will be just what its name suggests -- a very early, partial implementation of parts of Chandler. It's intended to give the flavor of Chandler and get valuable feedback on architecture and APIs. It is not intended to capture more than a tiny fraction of the product we have planned. I've added an "0.1 Release" section to the Chandler wiki. This section now contains a description of our purpose and goals for the 0.1 release, and a brief outline of the components we hope to include in this release.

We hope to release 0.1 before the end of April. Our initial target date of end-of-2002 was clearly erroneous, in part because it was picked before an adequate assessment of tasks and timeframes could be made. So we're "0 for 1" in our track record. We've done a lot more planning now, and hope to improve our track record to "1 for 2."

The developers are focused on writing code. It's still rough and in pieces, and probably not ready for review until the 0.1 release. But OSAF staff members are roughing out the framework for critical pieces of Chandler. Katie is working on the calendar parcel and Andy is working on the contacts parcel. John and Jed are working on the "parcel framework." All four are working on integrating the parcels with the parcel framework. More details can be found in the 0.1 components document. Morgen is getting the build system together.

Chandler Architecture

In preparation for the 0.1 release, there has also been a significant focus on architectural concerns. We had an initial, very general understanding represented by the block drawing created months ago. This was obviously inadequate for understanding or creating a product with Chandler's scope. OSAF staff spent many hours working through questions to develop a more robust understanding of Chandler's likely architecture. These discussions were also influenced by the desire to allow Chandler operate well in both centralized and decentralized environments, as discussed in Mitch's weblog posting of 2 February.

I sat in on the first couple of meetings, and could not figure out a way to effectively capture the discussions. But we are closer to producing an architectural block diagram with explanatory text that should be far more useful in understanding Chandler's design than what is currently available. These materials should be available for the 0.1 release.

Database issues have also generated a great deal of discussion. Rys has been working on documenting the database requirements that ought be met in Chandler. The requirements are likely to include feature requirements, developers' usability requirements, performance-related requirements, data model-related requirements, and perhaps some uncategorized requirements. Once this set of requirements is in reasonable form we'll post it.

Lou is trying to figure out BeepCore^? and thinking about a protocol for accessing the Chandler repository (Repository Access Protocol). We hope that an initial definition of RAP and perhaps some stub code will go into the initial release, but we still need to decide where RAP will reside with respect to the database architecture. The current thinking on the RAP protocol has been posted to the Chandler wiki.

Meanwhile, Aleks is spending a lot of time thinking about security and hoping our opening for a security developer is filled soon so he can stop thinking so much about security. He's also trying to learn everything about full-text indexing: open-source projects that might be available, details about implementations, relevant caveats. If you got info, please let Aleks know through the "dev" mailing list. Aleks also volunteered to give a presentation about OSAF at PyCon 2003 this March in Washington D. C.

I'm sure this summary will generate many questions and it's certainly not a substitute for detailed discussions with the developers. We all want to see code, so I won't ask the developers to spend a great deal of time responding to discussion this Update may generate. I'll answer general questions if I can. But even if I can't respond to questions directly, I'll try to address feedback in future Updates and other communications. In the meantime, I hope the Updates give at least a rough idea what the organization is working on.

-- MitchellBaker - 25 Feb 2003