At 10:01 AM 8/23/2006, Frank Brickle wrote: >On Wed, 2006-08-23 at 08:06 -0700, Jim Lux wrote: > > > But is this in PowerSDR (which is my frame of reference/context) or in the > > Linux fork? > >Strange as this may seem, it's the Windows version that's the fork. It's >intended that there will be no difference in the future. > > > And, I assume this effort include a port/compile/translation of the Erlang > > kernel into the Windows environment? > >Erlang code is cross-platform by design. > > > Is there any preliminary description of this development? > >Among the developers, yes.
Would this description be suitable to see the light of day, or will we wait for it to be revealed at some time in the future? Would this conversion of the dsp core of the SDR1000 further imply that PowerSDR will not use the C version of dttsp in future versions? Is there a schedule for when this might occur? That is, should someone contemplating writing DSP code for integration with PowerSDR struggle with the existing C dttsp and the C# PowerSDR wrapper (since they are quite intertwined), since if they were successful, there's an immediate path to distribution within the flex community, OR, should they hold off some (hopefully short) time, during which they could come up to speed on Erlang, so that their contribution could be integrated into the post-Erlang-conversion of dttsp. Philosophically, I like the idea of using a development language that enforces good practice/facilitates development of these things. However, the more non-standard things are, the less likely you are to get contributors, because now a would-be contributor must jump up another step to start contributing. It's a lot less intimidating to start trying something by modifying an existing C code base than in some language that is brand new and little used. > > And is this "shown" in some tangible form (i.e. a document that describes > > the process model)? I googled for "concurrency dttsp" and found > > nothing. Or (which is perfectly acceptable) have you just talked among > > yourselves and drawn on the blackboard and assured yourself that you've > got > > it right? > >No, actually, I wrote it down over two years ago, and as the concrete >implementation solidified among the developers, everybody was convinced, >eventually :-/ > >It's not difficult to do. There are a few papers by Per Brinch Hansen >which lay down the informal terms in which it can be described. I can >dig them up if you're interested, or they're on the web. P.B. Hansen described the structure of dttsp? Or a formalism for proving behavior of multithreaded applications in general? I'm familiar with many approaches to validation of multithreaded applications (or, communicating sequential processes, to use another formalism). What I was interested in is the specific details as applied to PowerSDR (those few pages from a few years ago, that is...) Jim _______________________________________________ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/ FlexRadio Homepage: http://www.flex-radio.com
