
Hi group, as you know, the requirement document (RD) is a huge missing piece in our milestones, and also is something that nobody happily wants to do (Gabrielle and I volontieered once, but never really did it :-( ). OTOH, we already have a strawman candidate, and in that sense we are ahead of schedule. Now, that would open a new approach for the RD, but I am not sure if thats a good one. So, here it is, please give feedback. The RD is supposed to evaluate the Use Cases (UC), and to derive requirements to the SAGA API, both functional and in terms of design. Since we have the Strawman, we could do through the use cases, and decide: - UC1, file read: ok in strawman - UC2, replica find: missing in the strawman - UC3, task graphs: 50 % in srawman (tasks, but no dependencies) - UC4, thread safety: not specified i strawman - etc etc That is fairly easy to do, and would give a RC which is very useful to the strawman evolution. OTOH, its not exactly the approach one should go for a RC. Any feelings about this? Would it be stupid for some reasons? Cheers, Andre.
Hi Tom, Shantenu,
I pondered a while about the requirement doc (RD). To rephrase what we know: the RD is supposed to exatrct a list of exact requirements to the SAGA API spec and implemenation from the set of use cases (UC) we have, and from general design constraints (80:20 etc).
Now, we already ha a strawman. One approach for the RC would be, to map the strawman to the UCs, and define missing pieces. That is the whole process backwards, I know - but I think its still a valid approach.
Also, I think that would be much easier than to start from scratch, and would yield results much easier to use.
What do you think? If you agree, I would put that proposal on the list.
Cheers, Andre.
----- End forwarded message ----- -- +-----------------------------------------------------------------+ | Andre Merzky | phon: +31 - 20 - 598 - 7759 | | Vrije Universiteit Amsterdam (VU) | fax : +31 - 20 - 598 - 7653 | | Dept. of Computer Science | mail: merzky@cs.vu.nl | | De Boelelaan 1083a | www: http://www.merzky.net | | 1081 HV Amsterdam, Netherlands | | +-----------------------------------------------------------------+
participants (1)
-
Andre Merzky