From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: *** X-Spam-Status: No, score=3.2 required=5.0 tests=BAYES_00,RATWARE_MS_HASH, RATWARE_OUTLOOK_NONAME autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,a1530a5c9e0c4594,start X-Google-Attributes: gid103376,public From: "Joachim Schroeer" Subject: An example how to sabotage Ada! Date: 1997/09/01 Message-ID: <01bcb6ff$538d0dc0$5ed653c1@sim01.amst.co.at> X-Deja-AN: 269482065 Organization: Customer of EUnet/PING Austria Newsgroups: comp.lang.ada Date: 1997-09-01T00:00:00+00:00 List-Id: Sorry, I sent this post a few days ago, but maybe it was too hidden in the deep tree of replies to "Differences between Ada and C++". I want to see, if I get some more comments to my points and post it again. Especially look at the last point below. Joachim --------------------------------------------------------------- Mike Young wrote: > > In counter-point, though, high quality class libraries are every bit as > important as a good compiler, perhaps even more so. For example, > Microsoft's much maligned MFC allows even a moderately skilled > programmer to quickly put together a usable skeleton for an application. > Its pervasiveness lends some guarantee that this core portion works > reliably. > > Regardless of its other attributes, Ada lags far behind in this > important regard. Until supporting libraries reach the same level of > maturity as those accepted as commodity in C++, GNAT will remain an > educational oddity only in Win32 and UNIX environments. > Use bindings ============ Libraries written in C++ are no reason, why C++ should make software development more easy than Ada. C++ libraries may be used by Ada developers as good as by C++ developers. Ada95 can be combined with C and GNAT can bind C++. The Ada community should look, what libraries are around in C++ and simply use them via bindings. If software companies can afford the luxury to make complex software waterproof in C/C++, let them do it, we use it. -------------------------------------------------------- I got this answer: > Sorry, many libraries consist of templates, which you cannot bind to > Ada; you have to rewrite them as Ada generics. This will be a Good Thing > when it gets done, but it is not easy. > and asked this question on reply: Ok, but is it not possible to bind the instantiations? And what is with the SGI tool?, They built a binding to the inventor library with it. Joachim -------------------------------------------------------- Give binding generators to the Ada community ============================================ Is it so difficult to write a binding generator? There exists one for C (c2ada), but that thing is hard to install. You have to gather the pieces from 3 different locations in the net and it's written in 2 languages (C and Python). In the articel "Interfacing to other Languages using GNAT and Ada95" by Cyrille Comar and Robert Dewar I read, that SGI has a binding generator for C++. Why do they not advertise that program and sell it? What I would want are (working) binary distributions for those programs, or sources, if the program is written completely in Ada. I always have difficulties to install programs written in C from sources, not so with Ada: copy, gnatmake, ready! There exists the gcc for C and C++. I thing for a student, studying compiler construction, it is not very difficult, to use the frontends for a binding generator. Pay people for writing Ada libraries ==================================== The development of the booch components by David Weller is a light in the tunnel. But why has he to do that in his free time, why are people not payed for such efforts? Make it possible to use tools together with Ada =============================================== There are a lot of application domains (GUI builders, modelling, simulation tools ...), where you get tons of documentation how to use those tools with C / C++. Then you phone the company and ask: 'What's with Ada?' 'Oh yes our tool works with Ada!' ... and then you get one sheet of paper where that marvellous tool is described and I guarantee: it's 10 years old, generates/accepts only Ada83 and if you are very lucky, it works. But it only imitates C in Ada. How to hinder Ada ================= I always have the impression, that Ada's sucess is hindered massively. In february 97 in Munich took place the OOP97-conference with numerous speakers and over 1000 participants. You could hear dozens of presentations dealing with C++, Java, Smalltalk and so on. The german organization "Ada Deutschland" managed it, that a few Ada presentations are accepted by the organisation "SIGS conferences". But what happened? All other presentations were held in the big main presentation rooms, where hundreds of people could listen, but the Ada group was banned to a small room in the neigbour hotel. Also the Ada presentations were not mentioned in the time tables. The organisor of "Ada Deutschland" (an Aonix employee) ansered to my question, why this happens, that "SIGS conferences" in USA did'nt want Ada pesentations in the main courses. Especially they do not want that such a specialist like Tucker Taft tells something about differences of Ada95 and Java or Ada95 and C++ in front of hundreds of potential Java / C++ customers. And Tucker knows exaktly the weaknesses of those languages, as the Ada95 designers evaluated existing technology and Intermetrics combines Ada95 and Java by the "Java vitual machine". So only 20 people (who already knew most of what he tells) could enjoy the deep technical knowledge and the entertaining presentation style of Tucker. Joachim