Do I really need to learn Java?

John Flynn transpicio at yahoo.com.au
Tue May 15 00:59:01 EDT 2001


"James Logajan" <JamesL at Lugoj.Com> wrote in message
news:3B001662.3971203 at Lugoj.Com...

> [...] Time is finite; in the end you need to learn to
> solve problems and you need a way to express those solutions using a
> notation you are fluent in.

So far so good.

> Only when the notation becomes a noticable
> impediment should you consider searching for another lnaguage.

I agree with you in principle that it's better in the short term to be
highly proficient in one language than to have a passing acquaintance with
several. (I've done it the "wrong" way, so I should know) ;-).

Trouble is, in the longer term you're not likely to notice that your
favourite notation MAY be impeding your progress until you've seen a range
of alternatives. Also, you're not as likely to explore the possibilities of
your favourite notation until you've seen how it can express ideas and
styles that aren't exactly native to it.

Eg. I think the standard C++ library is wonderful, but it's so different
from previous 'native' C++ that it's almost a new language. I've honestly
learned more about it by using Lisp and dabbling in Haskell than I have from
using pre-STL C++. (It really helps to understand the "whys" as well as the
"hows"). Sure, it's possible to learn about functors &c by reading the
documentation, but unless you've seen them used easily and naturally in
other contexts, they're likely to seem like academic curiosities or
stop-gaps, rather than features that open up lots of great possibilities.

BUT, I do agree that it's important to 'settle in' to a few languages and
push them to meet your needs, rather than constantly chasing whatever cool
features your favourite notation doesn't have. (The fact that Linus Torvalds
wrote an OS in plain old C while I was writing "Hello World" in a dozen
different ways puts this into perspective for me).

> Pick one language, it doesn't matter much which, and start using it to
solve
> problems.

In hindsight, I'd say pick four of the best, preferably from different famil
ies. Pick one as your staple diet, and suck the juice out of everything
else.

Cheers,
J.






More information about the Python-list mailing list