🚀 go-pugleaf

mgreenaw@lionel.st.nepean.uws.edu.au (Mark Greenaway) writes:

>jmartin@cs.ucla.edu (Jay Martin) writes:

>>What language is used for first class is not as important as 
>>which languages are used in the core classes and which languages
>>are used in research.  If you take your list and sort by 
>>the top 40 US CS departments, you get:

>I disagree. The basic fundamentals formed in an introductory course 
>affect the student's view of computing for all courses afterwards.

I agree somewhat, but alot of the time teaching languages will be
mentally dismissed later by the student, especially if it is
impractical teaching language.

>>First, Pascal is a ancient joke, you can't really do anything real in
>>it today and in my opinion those schools that teach it are too
>>bleeping lazy to change CS1 to a practical language.  (My
>>undergraduate institution UCI said they were going to convert to Ada,
>>that was 10+ years ago) Pascal's antiquated design makes wonderful
>>fodder for C programmers to ridicule pascal style languages ({} vs
>>begin/end, etc).  Its a temporary language which quickly gets switched
>>to the real "core" language.  This is most likely to be C/C++.

>Practicality is not the measure by which you should judge a teaching 
>language. However, I agree that Pascal's design is somewhat antiquated. 
>This does not diminish it's value as a teaching language, at least 
>initially. And the switch from Pascal to Modula-2 etc. is easily 
>accomplished, when necessary. The same can not always be said for C, or 
>for C++.

Pascal syntax is now old and cruddy and its missing many modern
concepts.  There is no point in wasting time on it when the students
must quickly go on to (hopefully) Modula*, Eiffel, Ada,... anyway.

>>Scheme is also a theoretic joke.  Its for those computer scientists
>>who think programming is non-scientific vomit and want to shoe-horn
>>all the pathetic programming "CONCEPTS OF COMPUTER SCIENCE" into a
>>single quarter so they won't have to teach it very often.  (They don't
>>care that they have 4-5 years to teach programming/software
>>engineering...).  Thus, Scheme is also a temporary concepts language
>>(which like Pascal through its worthlessness promotes the conversion
>>of students into C biggots).  Depts like Berkeley and UCLA quickly
>>switch to C in the second class (CS2) and basically use C for the
>>rest of the way (UCLA has now killed Scheme for the first class and
>>teaches only C++).

>Again, I could not disagree more. Scheme embodies the mathematical 
>underpinnings of programming, and challenges many fundamental design 
>assumptions in modern programming languages. Please read "Scheme And The 
>Art Of Programming" from MIT Press (one of the top CS institutions in the 
>world), and see whether you still feel this way.

The mathematical underpinnings of programming are pretty useless if
you ask me.  Maybe I should read the above book, but I have already
read Abelson, which I see as an incredibly bad introduction to
programming (They should rename it "Fun Lisp Hacking while trying to
look Scientific").  That Computer Scientists see the book as great is
just another indication that academic Computer Science neither
understands nor respects basic Software Engineering and practical
large scale programming.

>>So basically for most universities in the above list, you can just change
>>Pascal => Pascal/C and Scheme => Scheme/C.   C is thus around 100%.

>Not really. Your argument is illogical at best.

Its an observation, not an argument.  And I observe (graduate 
student programmers):

C C C C    C C C C  Cy-C Cy-C... (Sung to the Monty Python's SPAM song)

>>From what I have seen in research, Unix/C/C++ almost completely
>>dominatates most software oriented parts of computer science like
>>operating systems, etc.

>Yes, C is an excellent systems programming language. UNIX is also quite 
>nice. However, I am not foolish enough to think there is one true 
>language which is perfect for every task I may wish to accomplish with 
>it. Nor am I foolish enough to believe that the field of play won't 
>suddenly change, forcing a total rethink of your philosophy. 

C and "excellent" and UNIX and "quite nice" do not, in my opinion,
belong together in the same sentences.

>The fundamentals of Computer Science bear more importance than 
>comparitively minor issues like which language we use for the time being. 
>Languages come and go, as do Operating Systems. The Church-Turing thesis, 
>on the other hand, is liable to be with us for a long time yet.

I am sure the many mathematicians masquerading as Computer Scientists
agree.  I say that Computer Science should concentrate on programming
language design, software and operating systems,etc (instead of just
dismissing them as being temporarary) and CS should not primarily be
some mathematical playground for ivory tower types.

Jay