KlausRuedigerHase
commented
11 years ago Merlin and Sylvain, I agree with Merlin's concern and after having talked to ERA just yesterday (July 25, '13) they have found quite a couple issues in SS026 which are related to interactions between the wayside and vehicle side functionality and especially under real-time conditions when response times of either side are not sufficiently defined. That means that a modelling of only one side (here the on-board as defined by the SSRS) is not sufficient for analyzing these real-time problems. Actually what we may have to do, is modelling Sub-System-Requirement-Specifications against each other: A "SSRS-Vehicle" and a "SSRS-Track-side", which is combined the SRS again. That brings me to the question: Why not first modelling the entire SRS and once the SRS has been decomposed into very basic functions and data exchange between those basic functions have been defined (classical functional analysis), then drawing a line around all those functions, which belongs to the "SSRS-Vehicle" while separating these from the functions related to the SSRS-Track-side? This is obviously not only an RBC issue. Many more way-side components are involved. Another importent issue then becomes the interaction between way side and vehicle side during transition into and out of the class B systems. Again, only modelling the vehicle side SSRS will not sufficiently cover real-time problems during the transitional period. So how do we want to proceed?
Dear Sylvain, I hope you can still fix this issue it in your document or in another document. So my concern is that: In Chapter 6 of your document, you told about what the SSRS is and what should be implemented in this document, but when reading your document again, I miss an explanation about: why we shall in the scope of the openETCS project not model the subset026 in its full extent, but shall model the SSRS.