Mailing List Archive

Support open source code!


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ++CD-ROM drive



Jonathan Shore wrote:

> Hmm, he did mention *efficient* here - maybe lisp qualifies by another
> standard.  Last time I checked the lisp/scheme groups they were still
> talking about GC algorithms (and that after 10+ years).  Interesting topic
> but it gives you an idea of some of the problems with the language.

Lisp groups have been discussing garbage collection algorithms for around
FIFTY years.

Automobile manufacturers have been discussing efficiency for around ONE
HUNDRED years.

Guess what, it's not going to stop, improvements are always going to be
useful.  Well maybe not while Microsoft is in control, and maybe not in the
environments of most of the employers I've had, but traditionally the idea
of improving a product was considered valuable.

There is another memory management style that is often considered the
opposite of garbage collection, but it isn't really the opposite.  That
style is to let memory leak when programmers make mistakes in their careful
control of memory allocation and deallocation.  Now in fact, I almost favor
this style.  I like efficiency, and I think that programmers should free
memory when they know that they don't need it any more, instead of waiting
for big time-consuming collectors to clean up their messes.

But, I also like error checking.  This also isn't a useful idea when
Microsoft is in control, and not in the environments of most of the
employers I've had, but theoretically it has some uses.  Programmers do make
mistakes.  If programmers were allowed to have their mistakes caught, so
that machines might degrade slowly instead of crashing, and maybe
programmers could be informed of their mistakes so they might have help
correcting them, there is this theoretical possibility.  Garbage collection
could still be employed to catch memory leaks when programmers make
mistakes.

If you look carefully, you'll even find some languages that force all array
indices to start at 0, some that let the programmer specify the starting
index, and some that don't even let the programmer specify multidimensional
arrays in a clean, obvious, readable, reliable manner.  After FIFTY years.

Yup, this gives you an idea of some of the problems with the industry.

Yours sincerely,
Norman Diamond


Home | Main Index | Thread Index

Home Page Mailing List Linux and Japan TLUG Members Links