On Fri, 29 Feb 2008 13:48:52 +1300, "Pete Dashwood"
<dashwood@[EMAIL PROTECTED]
> wrote:
>
>
>"Judson McClendon" <judmc@[EMAIL PROTECTED]
> wrote in message
>news:IPGxj.111608$K27.8711@[EMAIL PROTECTED]
>> "Pete Dashwood" <dashwood@[EMAIL PROTECTED]
> wrote:
>>> "Bill Gunshannon" <billg999@[EMAIL PROTECTED]
> wrote:
>>>> "Judson McClendon" <judmc@[EMAIL PROTECTED]
> writes:
>>>>> "Michael Mattias" <mmattias@[EMAIL PROTECTED]
> wrote:
>>>>>> "Pete Dashwood" <dashwood@[EMAIL PROTECTED]
> wrote:
>>>>>>>
>>>>>>> It has always puzzled me why so many (particularly COBOL) people
>>>>>>> hesitate to make the leap to a different language, when
>>>>>>> "programming ability" is an underlying skill, that really
shouldn't
>>>>>>> be language dependent...
>>>>>>
>>>>>> You are assuming the presence of fundamental programming skills.
>>>>>>
>>>>>> However, many of the modern development tools/environments allow
>>>>>> "developers" to create applications without ever learning those
>>>>>> fundamentals. A few clicks, a few drags, a few drops and presto!
you
>>>>>> can call yourself a programmer.
>>>>>>
>>>>>> With no such tools available, people of our generation HAD to learn
>>>>>> the fundamentals, so for us changing languages or development
>>>>>> environments is pretty straightforward... except when we find
>>>>>> ourselves in one of these newfangled IDEs where fundamentals don't
>>>>>> matter.
>>>>>
>>>>> You're right. And when these new "programmers" face a situation that
>>>>> requires actual programming skills, they're lost. I think a
demarcation
>>>>> between the two different skills would be useful. Perhaps something
>>>>> like "application assembler" rather than "programmer" would be a
>>>>> better description for such people. It always gets me when people
>>>>> who can only write HTML (for example) claim to be "programmers."
>>>>> It's like a typist claiming to be an "author."
>>>>>
>>>>> I've mentioned this here before, but a cousin of mine who is the
same
>>>>> age as me, and has a MS in CS, and has taught CS at university
level,
>>>>> has a daughter who just finished her BS in CS. He made sure she did
>>>>> learn good programming skills, but said he was dismayed that the
>>>>> university CS department put so little emphasis on it.
>>>>
>>>> If you are disappointed now, take a look at "Angel" the "new" way
>>>> to teach programming at the University level. Things do not bode
>>>> well for our industry in the not too distant future.
>>>
>>> On the contrary, I think the future is bright. (Mind you, I'm an
optimist
>>> and I ALWAYS think the future is bright :-))
>>>
>>> The "programming" you are bewailing is no longer relevant. As I may
have
>>> mentioned here before, it was a phenomenon of the latter half of the
20th
>>> century.
>>>
>>> Judson's cousin is simply dismayed by change. The University have it
>>> right.
>>>
>>> If you did a Masters in English, would you expect to learn Sanskrit?
>>>
>>> Do you need more than a passing knowledge of Greek, German, and Latin
to
>>> study English Literature?
>>
>> Pete, I think you're missing the point. I agree that what they're
teaching
>> now is sufficient for most applications programming. But for the
>> foreseeable
>> future, we will still need skilled programmers to build the tools.
>
>Nope. You are woefully out of touch. The tools we have now can be used to
>build better tools. The days of telling computers what to do, line by
line,
>are definitely numbered. We don't need to and it is far too expensive to
do
>so. Who codes in assembler any more? No need to. Why don't we? It's the
same
>argument.
>
>The kids today ARE skilled programmers. It's just that their skill sets
are
>not the same as yours. (Neither do they need to be...)
>
>Furthermore, your "foreseeable future" is very different from mine...
>
>
>>You aren't
>> going to build .NET, for example, using .NET or other drag and drop
tools.
>
>Don't need to; it has been built. It can be easily extended without
having
>to resort to low level tools. The magic phrase is "encapsulation of
>classes".
>
>I delivered a desktop application recently that did a fair bit of stuff.
It
>was written in C# and used Legacy COBOL components also. I expected it
would
>be a very large installable. It wasn't. Around 5 MBs and 4 MB of that was
>COBOL run time... Then I realised... it was using all kinds of classes in
>the FCL which don't have to be delivered... years and years of effort by
>programmers unknown, to provide encapsulated functionality that simply
>works. Why re-invent the wheel?
>
>> If we aren't teaching those skills in our Computer Science curriculums,
>> where are future programmers going to learn them?
>
>They won't need them, any more than my English Lit. Major needs Sanskrit.
>
>Your argument is postulated on current understanding of how computers
work.
>(The Von Neumann model).
>
>That model is already being challenged in a number of laboratories, on a
>number of fronts. Given that it gets replaced (and I believe that will
>happen within 15 years) there will be no point in people learning
>"programming" as you and I understand the term.
>
>What has most opened my eyes to this has been my relatively recent
dipping
>into Query (LINQ) and Lambda expressions and functional programming. For
>decades we have believed that data can only be manipulated in the ways we
do
>it; SQL is King. Now it turns out it isn't, there are much more powerful
>ways to manipulate data, and storage technologies and devices will only
be
>limited by the software we use to drive them. Running SQL on a
>multiprocessor, content addressable, nanobased storage device is like
>printing invoices in Roman numbers.
>
>We need approaches that are completely hardware independent so that
multiple
>concurrent processes can be initiated and assigned automatically by the
>system where possible, and our interface to it facilitates that. (LINQ
query
>expressions are a perfect example of this...)
>
>Our current approaches derived from sequential hardware, which was all we
>had; we are due for an upgrade.
>
>It is going to be hard for many here to accept that what they spent
decades
>on learning and applying is being overtaken by events.
>
>Whether they accept it or not, that is the reality.
I think Linus Torvalds would give you a big argument on that. He will
NOT accept C++ code into the Linux kernel. It has to be C.
>
>Pete.
|
