- From: Dave Reynolds <der@hplb.hpl.hp.com>
- Date: Sat, 10 Jun 2006 17:36:13 +0100
- To: Sandro Hawke <sandro@w3.org>
- Cc: public-rif-wg@w3.org
Sandro Hawke wrote: > [mostly for UCR editors] > > UCR is getting rather disorganized. It's kind of unreadable. > > Here's a proposal for how to organize it, motivated in part by a > discussion with David yesterday, after the meeting ended. If people > like this organization, we can start to figure out how to get the text > to look like this. (I think this is all editorial stuff that doesn't > need Working Group approval, except that the WG needs to actually be > able to read & understand the content it approved in the meeting.) Seems good to me with a couple of minor caveats .. > 1. Introduction > > ? what are rules? why standardize? some history? > > 2. Use Cases > > for each use case: > > title > text > links to requirements, maybe CSFs > (later: links to test cases) > (maybe: links to more detailed versions on the wiki for people > really trying to solve a problem like this) > > 3. Requirements > > define our terms ("covers", ...) > > for each requirement: (in alphabetic order by title) > > short title (no more than 40 characters - used for links) > statement (1 paragraph) > links to use cases and CSFs which motivate this requirement > additional comments > either: approved for phase 1 // under consideration for phase 2 > (maybe "under consideration" items don't appear in WD?) > (maybe group by this flag, and then alphabetize within groups) > > 4. Goal Analysis Goals not Goal Analysis. I'd put the goals/CSF before the requirements. To me those are our top level statements of intention (hence all my hassling about a fourth goal :-)) not just a way to group the requirements. > > description of Critical Success Factors process / terminology > > diagram -- maybe a imagemap with links to appropriate > descriptions (maybe even as pop-up on mouse-over if > someone feels motivated) > > for each goal > short title > statement > link to CSFs (implicit in outline form) > > for each CSF > short title > statement > link to goals (implicit in outline-form) > link to requirements, and maybe CSF's > > 5. Coverage (RIFRAF) > > for each discriminator: > short title > explanation, including alternative vocabulary > flag: in phase 1, unresolved whether in phase 1, not in phase 1 > maybe some kind of grouping/clustering/hierarchy > (as in current draft) > > later (WD3?) - for each rule system/rule language, and for > each dialect, how does it match up to the discriminators? > (this would be a large table, or perhaps a set of tables, with > one per dialect). A table against the raw discriminators is likely to be fairly confusing. We have a lot of discriminators already and more to come. I think we'll need to some significant grouping and refinement (some sort of informal human-guided principle components analysis [*]) and should tabulate just against those summary discriminators. Dave [*] And yes, I know that the FCA that Hassan has mentioned a few times is somewhat related to this though without lots of classified instances I think a human guided grouping is likely to be more useful for the document.
Received on Saturday, 10 June 2006 16:36:08 UTC