Am 04.02.2013 um 10:25 schrieb EBo: > On Feb 4 2013 1:45 AM, Michael Haberler wrote: >> Am 04.02.2013 um 08:25 schrieb Chris Morley: >>>> What about different motion modules optimized different classes of >>>> applications? >>>> >>> >>> >>> Yes Peter I think your on the best track. >>> >>> Unfortunately that's a big chunk of code to try and understand to >>> change it. >>> I wish there was some way to write down the flow of linuxcnc and >>> keep it current. >>> It's like walking in the dark with a plain sheet of paper as a map >>> :) >> >> I fully agree. >> >> A good start would be for instance to make the trajectory planner >> (tp/tc) a module like kinematics. The upside would be an easier path >> to exploring different trajectory planners, for instance Yishin's; >> that is kind of hard to integrate into the current monolithic blob, >> without duplicating the code or using conditional compilation or >> execution, which would just add to the 'huge blob' problem. >> >> this requires abstracting, or removing two RT/non-RT layer >> transgressions in tp/tc where it hooks into emcstatus for CSS and >> spindle-synced moves, but given the original base code is very well >> designed and encapsulated, it is a contained effort. >> >> I would think something similar could be done for spindle handling. >> If it is possible to modularize a linux kernel such that it is >> basically a module-loading skeleton with minimal supporting code, >> then >> this should be possible for motion too. >> >> -- >> >> that is structural rework with no immediate upside, and I have the >> impression such jobs with longer-term impact are fairly off-the-radar >> as opposed to the eternally popular shortest possible fix. At least I >> dont see much of it. > > LinuxCNC-3.0 then as a design philosophy?
I would rather call it sheer necessity in some cases, although individual perceptions of urgency obviously differ. For instance, I dont think broadening the base of RT operating systems, or removing the 'everything on one CPU' limitation requires a lot of philosophical thinking; it is rather 'what is the minimum to be done asap from preventing this project from keeling over'. That is why I postponed from working on LinuxCNC3 grand visions until the RT OS stuff is done. Of course there's always an alternative, the bits dont go away - keep scavenging hardware which somehow, in some way fits the current operations model. That is a dwindling option afaict, nevermind some of the more bizarre limitations attached. - Michael > > EBo -- > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_d2d_jan > _______________________________________________ > Emc-developers mailing list > Emc-developers@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-developers ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_jan _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
