See also: IRC log
No regrets heard
Jim, I've done some work on zip/unzip but it's still ongoing.
Norm: Thanks, Jim. Send a pointer when it's ready for review.
Norm reports that Liam has sent the charter to W3C management.
Norm: Remaining actions are continued.
Murray: I've been working on it.
The draft is at the point now where what we need most is to
determine if it's going in the right direction.
... The document as it stands now, the introduction is quite minimal.
... I've left sections 2-5 pretty much intact. I've added comments wrt the standard names for component inventory.
... I've drawn a list of all the specs from XProc 1.0 and dropped them into this document.
... Verified that all the items in 4.5 were covered at least to some extent.
... Then went on to list a bunch of specs that weren't in the requirements list but have been satisified.
... And then for the rest of the sections, there's very little commentary beyond a few editorial notes.
... In Appendix A, I've added references to all the specs that I think might be relevant.
... The collective input that I got out of the wiki and the mailing list and all the other sources, I've been trying to collate.
... Appendixes D through F reflect those changes.
... I added the list of categorized steps that Alex published a while ago.
... Hopefully it's becoming complete and useful on it's own.
Alex: Going through section 5,
there are a number of use cases which I would say we're able to
satisfy if we stretch the definition of "satisfy"
... I can write a pipeline that satisfies the use case with some fabricated extension step.
... 5.10 for example, could be satisified if we had a digital signature step.
... A bunch of them are like that, they require steps that we don't have in our vocabulary.
... In other words, some we can write out and some require steps that we don't have. And some require interpretation of the use case.
Norm: I think if we can satisfy a use case with a new step, we should collect up those steps.
Jim: So you're looking for a
definition of "satisfaction".
... For example, 5.9, run a custom program. We have an optional p:exec step, does that satisfy it? I think it does.
Alex: I think of grouping these
things into three categories: we can write a pipeline that
satisfies the spirit of the use case, we could if we had a new
atomic step, and we can't without making language
... So 5.10, it's not clear how signing fits in, but 5.12, just requires any of a number of "store" steps.
Norm: I think starting with a three-bucket cut makes a lot of sense.
Jim: Would that be full, partial, and not?
Alex: I don't think it's partial.
Norm: I think we can say "use cases we can satisfy now, use cases we could satisfy with a new atomic step, and use cases we're not sure about."
Alex: We've given users the ability to define their own steps, so we might, but don't have to, write it for them.
Murray: So if we find a case like
that, we add a new step to Appendix E and that informs our
... From my point of view, the use cases and requirements is fodder for the back section and we get there later.
Alex: I like the idea of grouping
things into three categories and we can finesse the middle
... Maybe we can just give them labels without worrying about the subtlety of the descriptions.
Jim: I'm happy to review them as well.
Alex: I personally would like to see us write the simple pipeline that demonstrates the solution for categories 1 and 2 at least.
Norm: I think that's consistent with what we said last week.
Alex: That will also help us decide if it's easy or not.
Norm: I think sending the
solutions to the list is a good idea.
... Murray, can you checkin what you've got done by next Tuesday so everyone has a couple of days to review it before the next meeting.
Alex: And I'll try to get it checked in.
Norm: I think the document is going in a good direction. Murray, are there any other questions you'd like answered before you continue?
Murray: Not that I can think
... Feedback on the use cases would be really great.
Norm: Any further comments?