Conference bulova::decw_jan-89_to_nov-90

    I don't think there's an easy answer to this question: quite a few
    parameters need to be taken in account:

    �	What are your career development goals?

    �	Will you be asked to teach DECwindows for some RISC hardware?

    �	You teach DECwindows; let's face it: about 90% of the code written
	nowadays in the world for DECwindows, Motif, MS/Windows, News,
	Presentation Manager, Open Look is written in C. If you would like
	your course to evolve and grow or be refined, you'll have to become
	a proficient C reader if not a proficient C programmer.

    �	You said your management is happy, and so are the customers you
	teach DECwindows to; I'm pretty sure you'll start to see more and
	more customers asking for a course geared at C programmers. Your
	management will then come back to you and say 'Well, how about...'.

    You probably aren't REQUIRED to learn C. Whether you should learn C is
    up to you, given some of the parameters outlined above and others. Why
    don't you start learning C as a midnite project? I'd be surprised if
    you ever regret it. IMHO, it's a real pity that a proficient FORTRAN
    programmer will have to pick up C as the next programming language to
    learn, given that some other languages really are at least one
    generation ahead of C (Modula, ADA...); it's a variation of the 'Common
    Denominator Law', but that's a rat hole which belongs to SOAPBOX ;-)

    Regards,
    Denis.

    We have plenty of requests from independent software developers for
    expertise in DECwindows programming in Fortran.  It's hard to find
    folks who are proficient in both.
 
    Mark
    

On this side there's a lot of DWT and X happening in Fortran, ADA, and 
Pascal and few if any examples.  Now C is a prerequsite for the class 
and often students don't know what a structure is.  We've really lost 
by not being able to teach the VMS bindings. While learning C may be
helpful to you what our customers need is more of the other languages.
This trend will only grow as the mystique wears off. 

    Well, I'm taking a class in DECwindows Programming in Bedford this
    week; the only prerequisite was 'experience with Structured
    Programming'.  Well, I feel I've fulfilled that requirement in the past
    with many years of FORTRAN, Pascal, etc... ** all but C.
                                                  
    Now I'm really wishing that I'd taken the time to learn C before
    attending the class...I probably would benefit more from the
    course.  Instead, I find myself trying to translate the C code into
    something I understand before I can even BEGIN to make sense out of the
    XLIB and DECwindows Toolkit calls.
    
    I guess my point is...C should be a prereq for teaching as well as
    learning DECwindows.  I know how I feel, and since this is primarily a
    customer course, I can imagine how they feel.  Feel free to comment.
    
    Catherine
                                                   

    Knowing C ought not to be a prerequisite to doing DECwindows 
    programming.  That most people think it is just shows that we failed
    to provide a proper programming interface.
    
    Still, accepting the reality that the interface was designed with
    blinders on, we could do a MUCH better job of making it easier to
    use languages which are much more widely known and used than C
    (namely FORTRAN) and those whose importance we cannot ignore 
    (namely Ada).  By pretending these languages don't exist or aren't
    important, we alienate many customers who don't want to use C,
    and also serve to tear down the competitive advantage we have built
    up on VAX over the years with the common language environment.
    
    What we need to do is to devote more resources to making the
    multilanguage bindings WORK!  And then to teach them and show
    how to use them.  These other languages aren't going to go away
    just by wishing it so.
    
    		Steve

    While we're on the subject of prerequisites for teaching DECwindows,
    I'll hop on my soapbox.
    
    I believe Digital has done a great disservice to its customers and
    marketing partners, not to mention its employees, by assuming that
    professional instructors can be given three weeks to "learn" X and
    DECwindows, and then presume to be teachers.
    
    When I agreed to teach the ISV DECwindows course about a year and
    a half ago, one DEC manager asked me how much time a prospective
    instructor should be given to learn DECwindows before being required
    to teach the subject to an audience of programmers. I told him that,
    given the person was experienced in programming under VMS and Ultrix,
    in C, and had a reasonable knowledge ographics, that 6 months of
    study, equipped with all available documentation, a color workstation,
    access to NOTES, and several serious programming projects would be
    a sufficient preparation. Of course this was considered laughable.
    
    I believe a teacher should have command of his or her subject,
    from beginning to end, inside-out, pressed down, shaken together,
    and running over! And if I were a customer paying $1500 or more
    for a DECwindows programming course, and found that it was taught
    by a novice who never heard of DECwindows a month ago, I would be
    a ve angry customer.  And if I were the instructor being put
    in such a compromising position by my management, I would be just
    as angry.
    
    I'd better get off my soapbox before I get too worked up. The reality
    is that Ed. Services philosophy of course development and training
    doesn't include gaining mastery of the subjects involved, and
    Digital is the poorer for it.
    
    Cheers,
    Dave
    

    >The reality
    >is that Ed. Services philosophy of course development and training
    >doesn't include gaining mastery of the subjects involved,..

    I'm afraid you are right.
    
    I joined Ed Services in April or so, to teach A1 programming and
    some "standard" VMS courses (Utilities,SM1,SM2, etc..). When I have
    been asked about whishing to teach a DECW programming course,
    I said yes (because I know "a bit" DECwindows *use* and I am very
    interested in the subject), so I have been sent to Paris to get
    the DECW prog training (5 days). And here I am.
    
    Experience in C: 0 days (but in Fortran 10 years) :-)
    
    The main concern of Ed Services is making money providing courses at
    the lowest cost of training for the instructors. The quality that our
    customers get is *only* depending on the professionalism of the ones who
    prepare their courses, not on the time mgt leaves us to do so. 
    
    (is it really the right place to go on?)
    

    DECwindows (Xlib and the toolkit) is just a set of procedure
    specifications.  It shouldn't be necessary to know C in order to teach
    or to learn DECwindows, any more than you have to know Macro to teach
    the VMS run-time-library.  It _is_ necessary to know something about the
    data-structures used by the toolkit (in particular that some (but not
    all) of the strings in the language-independent bindings are
    null-terminated, and addressed by reference rather than descriptor, and
    that just about everything that can be is passed by value rather than
    reference), and it should be emphasized that unlike most VMS-provided
    procedures (I'm assuming you're from a VMS background), the toolkit
    does no (or hardly any) error-checking of parameters. However, provided
    you remember these peculiarities, it's not necessary to know C.
    
    The problems mentioned in .-2 (I think), about C appearing to be a
    pre-requisite for learning DECwindows, emphasize a problem with the
    documentation and course design, rather than the toolkit itself, and
    this is a problem that should be fixed.  It's not too bad provided you
    stick to the toolkit (and contrary to some recent notes in this
    conference, it is possible to build complex applications with
    reasonable performance using toolkit widgets), but any documentation
    on Xlib or widget-writing seems to be exclusively C-oriented.
    
    Given that a goal (probably the major goal) of the toolkit was to have
    an implementation that ported between Ultrix and VMS, Digital can
    really only blame ourselves for not having provided a better portable
    programming-language than C on both platforms.  If Ada (for example, or
    a DEC-supplied Modula-2 or -3) had been available on both platforms,
    perhaps more attention would have been paid to producing truly
    language-independent bindings.