Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 SMWG Service Management Book Refactoring Report Anthony Crowson Colin Haddow October 2009, ESTEC October 15, 2008.

Similar presentations


Presentation on theme: "1 SMWG Service Management Book Refactoring Report Anthony Crowson Colin Haddow October 2009, ESTEC October 15, 2008."— Presentation transcript:

1 1 SMWG Service Management Book Refactoring Report Anthony Crowson Colin Haddow October 2009, ESTEC October 15, 2008

2 2 SMWG Book Refactoring TM TC Bilateral Core Split current book into Core and Services  Prepare to add new services without changing Core

3 3 SMWG Book Refactoring  How did we change Nicolas’ model? Not much! Moved some details down to Forward / Return books - E.g. Carrier parameters - Which meant add placeholders in Core Rationalised (we think) parallel structures - Space Link Events Started introducing stereotypes to formalise the colour- coding

4 4 SMWG What did we do?  Cut & paste text to one of the new books  Replace diagrams with Magicdraw And a couple of new ones  Minor rewording, not fully consistently Remove > Consistency with new model  “Template” extension book not worked out

5 5 SMWG Examples

6 6 SMWG Service Books Collation  Core as B1 with deletions  Service Books: Table contents moved No new text written

7 7 SMWG Conclusions  Yes, the approach is feasible  No major problems But plenty to discuss  Core remains large Can we reduce repetition?

8 8 SMWG Suggestions  Consider re-arranging chapters e.g. Information Structures: common to Configuration Profile, Service Package, Service Agreement - Carrier Profile - RetrievalTsProfile - Event Sequence - etc Service Specs - Common structures within service e.g. identical / similar between Request and Return - Operations Move Document Exchange Protocol to own book? - Supported by idea of alternative mechanisms - All of it? Or are message patterns still fundamental?

9 9 SMWG Future Discussions  How prescriptive can / should we be about where model can be extended? We have «extensionPoint» Should all others be non-extensible? Should some specific ones be non-extensible?  Decisions on what level things should be at - Retrieval in general – currently only applies to return services, but could apply in future to e.g. ranging Parameters duplicated between R and F services. BUT in some cases constrained by different SA parameters. General approach policy needed.

10 10 SMWG Future discussions  Section numbering: try to preserve 1:1 match?  Requirements Numbering E.g. ASCSPU-FW-1 for forward service book  Namespace / naming convention for extended diagnostic sets Lower level adds diagnostics specific to the service Impact on schema needs to be taken into account.  Multiple inheritance Will probably all go anyway  Modelling tool & version Magicdraw 16.5 used so far 16.6 now out, some useful improvements claimed

11 11 SMWG Book Refactoring: Pages  Blue-1 (Chapter 4, 5, 6, 7) – 322  Core – 268 Only 17% reduction! Fully formatted as B-1  Service Books just the bits cut out of core no new / copied text  Forward – 23  Return – 36  Bilateral – 14  Final ratios will be less extreme!

12 12 SMWG Repetition  BB is a BIG book  After refactoring, Core is still large  Yes, there is lots to say, but… Some of it gets said rather often!  Repetition adds to the perception of complexity  Can we make it more DRY ? DRY Don’t Repeat Yourself

13 13 SMWG Repetition  Example: 3-Phase Operations  NOT an attack on current approach!  Illustration of one aspect to examine

14 14 SMWG Repetition Example  Every 3-Phase Operation says

15 15 SMWG Repetition Example  Every 3-Phase Operation says ? ? ?

16 16 SMWG Repetition Example  Every 3-Phase Operation says

17 17 SMWG Repetition  Other aspects of repetition include UML relationships described in text More generally: does expressing syntax rules as requirements add to perceived complexity of spec? Is there a better way to textualise the UML? +

18 18 SMWG Repetition  For repetition Everything spelt out where it’s used Less need for cross-referencing, looking up  Against repetition  More to read and recognise as “the same as last time”  Repeated edit on change  Opportunities for inconsistency  Size  Can we find clear but less repetitive approach?


Download ppt "1 SMWG Service Management Book Refactoring Report Anthony Crowson Colin Haddow October 2009, ESTEC October 15, 2008."

Similar presentations


Ads by Google