On 19/01/11 10:52 AM, Alex Huntley wrote: > I don't regret that GHDL is written in Ada at all - it is a brilliant > language with some real strengths, plus it's close similarity with VHDL > makes it much easier for VHDL engineers to pick up than C/C++. GCC/Gnat > can still compile a mix of Ada and C/C++ anyway so using either one does > not stop you from interfacing to the other.
Gee, let's re-write it in VHDL! (Just kidding, there'd be a lot of library development). > Just because C/C++ are popular does not make them the best language to > use for a given job. Amen. (I am reminded of the the president of the Federation's speech at Camp Khitomer, though. "Let us redefine progress. Just because we can do a thing, does not necessarily mean we must follow that thing." Star Trek VI: The Undiscovered Country.) > > There is an Ada port of GTK if graphical front-ends are required. > Oh, vey. There's also existing IDE's which are likely written in C++. > As far as System C and System Verilog go, I think it better to have a > really good VHDL simulator (like GHDL) than a mediocre simulator having > a go at implementing bits of equivalent languages. As an amusing side > note, a few years ago I heard a former chief technologist of Mentor > Graphics say that System Verilog gave Verilog all the things that VHDL > had from the start. Would adding system C / System verilog capability > make GHDL that much more useful to that many people. Why Python > scripting - most commercial simulators use TCL - this would be a far > better choice if GHDL was to incorporate a scripting interface. Adding additional languages parsers might be easier in C++. There are several open source C++ implementations using AIRE. It might be worth encouraging development on grt separately from the gcc front end. Any separation still needs to have a common view of libraries. The amount of code you introduce might be sufficient to encourage it to be re-targeted, separately. Were the interested community grow large enough some of that might spill over into analyzer land. > Anyway, I'm not trying to fan the flames of a language war with that > anecdote - just to say that I think it's better to concentrate > development of GHDL as a really good VHDL simulator than try to make it > the jack of all trades and master of none - if it's going to branch out > then it can do regardless of the language it's written in - Ada gives a > lot of benefits and saves Tristan a lot of bug fixing due to buffer > overruns, bad type conversions and all the congenital problems that come > with C/C++. You could note that military use of Ada has dropped off dramatically (so has non-FPGA use of VHDL for that matter). What happens when the Ada community starts noticeably downsizing? We could start having tool issues. The question should be is Ada sustainable (followed not too closely by asking about VHDL). _______________________________________________ Ghdl-discuss mailing list [email protected] https://mail.gna.org/listinfo/ghdl-discuss
