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=-0.3 required=5.0 tests=BAYES_00,LOTS_OF_MONEY, REPLYTO_WITHOUT_TO_CC,XPRIO autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,984e922902f4f4ee,start X-Google-Attributes: gid103376,public From: safetran Subject: Can Ada by popularized faster ? Date: 1997/10/09 Message-ID: <343D1DD8.B60A668A@kaiwan.com> X-Deja-AN: 279041846 X-Priority: 3 (Normal) Reply-To: safetran@kaiwan.com Cache-Post-Path: web.kaiwan.com!unknown@198.147.107.129 Organization: Safetran Systems Corporation Newsgroups: comp.lang.ada Date: 1997-10-09T00:00:00+00:00 List-Id: Yes, I do think that we can do more than we are doing. Maybe I am just not aware of all the efforts going on and will be happy to be enlightened. I come across uninformed criticisms of Ada quite often and find it irritating. Which is what prompted this post. Other than Ada I also use C/C++ and so am quite aware of the benefits that Ada provides. However discussions with collegues and others makes me realise that Ada suffers from a very strong market perception as being cumbersome, restrictive, slow, large, bulky and "that" defense thing. Few "amusing" encounters: I have been asked "It must be really slow to program in Ada." When I asked why, the reason given was that in Ada we have to type a lot of words (syntax is wordy) where-as in C/C++ most of the syntax is cryptic (and so it takes longer to program in Ada !!). Another interesting example I found in the Slackware LINUX Unleashed book. This book has a section on the GNAT compiler and in it the author states that Ada was so named as to honour one of the developers of the language ! Check it out next time you go to the bookshop (I think it was page 560). For a new project I am starting I need an in-circuit emulator. So yesterday I called a few vendors. One of the vendors told me that they do not support Ada as Ada code is very slow and their engineers think that Ada is not appropriate for embedded systems. BTW, this is one of the big suppliers of in-circuit emulators. I think the Ada community needs to do something urgently to change this market perception. I read someplace that a DOD colonel was proposing spending $1 million to do an advertising campaign. Is that the right way to do this ? I don't know. I guess we need to go talk to some savvy marketing types. Look at what Sun has done with Java in less than 2 years. Hardly a day goes by without us hearing or reading something about Java. People say that its all hype and marketing - does it really matter ? Its working and Java is becoming popular. Programmers and magazine articles refer to Java's elegance and strong type checking or that it is a safer language than C/C++. Ada has had these since 1983. But few outside the Ada community know that. I have talked to C/C++ programmers who have criticised Ada because it has strong type checking and it does not let you do "hacky" things. But then they go ahead and praise Java for its strong type checking !! The level of ignorance that exists in the world about Ada, even among experienced developers, is really quite amazing. One of the problems I perceive in the Ada market is that Ada tools vendors have been quite complascent. How many vendors have products that compare with Microsoft VC++ or Borland C++ and at a comparable price ? I am told that Ada compilers/tools should be more expensive because it takes a lot more effort to develop an Ada compiler (than a C++ one) and Ada compilers catch a lot more bugs thus reducing development effort and time. While I agree with all this, the bottom-line is that if the tools cost a lot more or are not as user friendly then the average company/programmer are not going to buy them. In addition, if the Ada tool vendors think that they have a captive market in the defense and aerospace companies then they are not going to do a lot of marketing for Ada. However, as we have seen it does not really pay to be complascent. Another related problem is that there is very little press coverage of Ada. Most of the coverage I see is in defense oriented magazines. It is not common to see articles on Ada and its benefits in magazines like Embedded Systems Programming, EE Times etc. C/C++ articles, and now Java ones, dominate most pages. A few things that we could do are: - Improve Market Perception: Ada has a very poor market perception. We need to do a lot more advertising of Ada's benefits as a programming environment and language. Specially benefits as applying to the commercial world e.g. improving time to market or lower development costs etc. The fact that Ada can be used to do telecommunication systems, databases, systems (OS) programming or device drivers should be advertised. Few people know that Ada can be used for these kinds of applications. A _LOT_ of people think that it is a language used only for defense systems or only for the embedded world. There needs to be a lot more articles in the common and technical press to increase awareness. Another example is that very few people know that you can convert Ada programs into Java byte codes. - Improve Market Visibility: Send out newsletters or mailers to developers (like Sun does for Java) advertising the benefits of Ada and successful Ada projects. People say negative things about Ada and then this becomes self feeding with more people criticizing the language without even knowing anything about it. And while the original criticism may have been valid its context is completely lost along the way. Mailers and magazine articles can be used to try and dispel some of the negative perceptions. - Improved Tools: Have better Ada tools from a lot more vendors - these should be more user-friendly, conform to the current state of the art for C++ tools and be easier to use/learn. eg the VxWorks environment for C/C++ which integrates compilers, linkers, debuggers, in-circuit emulators, design tools etc. The tools should also be cheaper. - Improved Processor Support: We need ports of GNAT as a cross-compiler to common platforms (eg X86, Motorola 68K and PowerPC). Maybe we could provide a graphical, user friendly interface. Ports of GNAT to run on (say) VxWorks, pSOS+ and other commonly used commercial real-time OSes would also help. - Maybe we should consider some way that Ada can be more easily accomodated on lower end micro-controllers. There is a huge market that uses lower end 8 bit controllers. And contrary to popular perception this market is not becoming smaller as 32 bit processors become more pervasive. Is there something we can do to target the 68HC11's, 68HC08's etc that are out there ? Sometimes these processors are used in safety critical systems; with C or C++ of course. Is ACT or someone else in the Ada community attempting to tackle these issues ? -- Rakesh