paul@mtnmath.UUCP (Paul Budnik) (11/20/89)
The operator overloading and object oriented features of C++ form an excellent framework for providing high level application oriented language extensions via class libraries. Such extensions will be of greatest value if they can play together. For example I am developing a C++ based tool for digital signal processing and will want to add image processing and matrix operations. The code for the DSP operations is proprietary. However, the language notation ideally will conform to (or become) a public domain industry standard. Is there any group or organization coordinating such C++ application oriented class libraries or interest in starting such a group? A formal standards process would be premature, but having some focus for collecting information and airing issues would be valuable. Paul Budnik
mhorne@ka7axd.WV.TEK.COM (Michael T. Horne) (11/21/89)
(directed primarily at comp.dsp, but others in comp.lang.c++ may wish to comment) In a recent article by Paul Budnik: > > The operator overloading and object oriented features of C++ form an > excellent framework for providing high level application oriented > language extensions via class libraries. Such extensions will be of > greatest value if they can play together. > > For example I am developing a C++ based tool for digital signal > processing and will want to add image processing and matrix > operations. The code for the DSP operations is proprietary. > However, the language notation ideally will conform to (or become) > a public domain industry standard. For DSP applications, just a few simple extensions can make an incredible difference in algorithm prototyping time. I've already implemented my own set of extensions (e.g. shift registers, fixed-point data types with corresponding math operations, etc.) and have used them extensively for (relatively) rapid prototyping of algorithms that eventually are implemented in hardware. A standardized extension (formal or informal) would be very useful and would greatly reduce duplicate work. > A formal standards process would be premature, but having some focus for > collecting information and airing issues would be valuable. Agreed. An informal discussion group (for DSP applications, if no global coordinating group exists) may be in order. Assuming this hasn't already started in one form or another, is there sufficient interest in an ad hoc group on C++ extensions for DSP use? Mike mhorne@ka7axd.wv.tek.com ...!tektronix!mhorne%ka7axd.wv.tek.com
apa@warwick.UUCP (Andrew Patterson) (11/21/89)
In article <1012@mtnmath.UUCP> paul@mtnmath.UUCP (Paul Budnik) writes: >For example I am developing a C++ based tool for digital signal >processing and will want to add image processing and matrix >operations. The code for the DSP operations is proprietary. >However, the language notation ideally will conform to (or become) >a public domain industry standard. > >Paul Budnik (Paraphrased) Wow - I'm developing a C++ based tool for DSP as my final year project here at University of Warwick. I'm hoping to produce a complete CAD system under Sunview. Unfortunately I can't e-mail to the US, so I'm having to write to you this way... Basically - do you want to share some ideas concerning desirable facilities such a system should have and whether it's worth bothering to try and use Sunview from C++, or should I stick to C to write the front end? Oh - any other contributions welcomed, of course. Pat -X- @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @ Andrew Patterson - CS Undergrad and general layabout... @ @ Local: apa@emerald Janet: apa@uk.ac.warwick.cs @ @ Arpa: apa@cs.warwick.ac.uk Post: c/o CS Dept. Univ. of Warwick, England @ @----------------------------------------------------------------------------@ @ "They keep calling me!" - jOY dIVISION (RIP) @ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
rex@aussie.UUCP (Rex Jaeschke) (11/24/89)
> For example I am developing a C++ based tool for digital signal > processing and will want to add image processing and matrix > operations. The code for the DSP operations is proprietary. > > Is there any group or organization coordinating such C++ application > oriented class libraries or interest in starting such > a group? The Numerical C Extensions Group was formed in May '89 to look at extensions to ANSI C. One active member of the group is Kevin Leary from Analog Devices, a DSP shop. They have their own extensions too. Other NCEG members (such as Cray) also want to add matrix support directly in the language. As the convener of NCEG I'd like to be kept informed of your progress and to get some documentation as "prior art." Rex ---------------------------------------------------------------------------- Rex Jaeschke | Journal of C Language Translation | C Users Journal (703) 860-0091 | 2051 Swans Neck Way | DEC PROFESSIONAL uunet!aussie!rex | Reston, Virginia 22091, USA | Programmers Journal ---------------------------------------------------------------------------- Convener of the Numerical C Extensions Group (NCEG) ----------------------------------------------------------------------------