Date: Thu, 15 Jun 2000 06:45:30 -0000
|From:||Yee Su Ling"|
|Subject:||OHS 6/8/2000 minutes|
(Drafted by Su-Ling, and once again Eugene makes me look good with his additions and editing :))
OHS Meeting 6/8/2000, SRI Engineering Bldg, EK255, 4-6 PM
Email him terms, corrections, suggestions, at email@example.com.
Pat says pieces from all contributions thus far on the list can be used to craft this.
Eugene especially liked the first few sentences of Warren's proposal, because it set the context of the project. Pat agreed, and made the following points:
Pat said that in the "first steps" paragraph, we should say what we expect to accomplish over the next three years. We need to include the advantages of solving a serious problem. The OHS and DKR should be mentioned in this paragraph, along with parallel projects that we see growing out of these.
Pat and Lee explain that we are pitching a DARPA program, not a proposal. Programs provide funding for a multitude of projects. Proposals are pitched under the auspice of a program. Once a program is created, DARPA creates a BAA (Broad Agency Announcement) soliciting proposals. In essence, we are helping DARPA create a BAA with our proposal.
Marcello noted that the process of creating a program is a long one, typically around 18 months. Pat added that there are ways to shortcircuit the process.
Other notes on the "first steps" paragraph: We should include the maximum amount of interesting, useful info that will fit in the paragraph. Referencing standards (eg. XML, etc.) that we can build on top of is important. Also important to include scheduled breakthroughs.
Pat said that he has enough from the suggested summary paragraphs to piece together a paragraph of his own, and said that he will send it out to the mailing list for comments.
Storyboard it out. For example, it depends on the style people do their programming. Doug says it's important to have different options for views.
Brainstorming follows on what would be useful views to have:
Lee asked what this means.
Doug noted that this is useful depending on a programmer's style of commenting.
Eugene noted that while this is almost a universal feature, there are still some creative things we can do. For example, he suggested bolding or graying out code based on profiling information -- the more often code is called, the bolder it looks.
John proposes building IP protection into the architecture and discusses 3rd party escrow. Lee says verifiable journaling meets this need.
A very important thing to start with is email. Hypermail, adding linkability and citability.
It is agreed that next week we will go through use cases so we can come up with targets for prototyping.
John talks about WebMinds as a use case.
Doug brought in his personal computer so that he could demonstrate Augment. He showed its command set, context-completion, help, linking, multiple views, journal, editing capabilities, extensibility, etc.
It was very cool.
Yee Su Ling