[rescue] slow java code...an example of why it might be happening

[Jeffrey Nonken]

On Tue, 11 Feb 2003 14:50:51 -0600, Bill Bradford <mrbill at mrbill.net> wrote:

> On Tue, Feb 11, 2003 at 02:12:44PM -0600, Joshua D. Boyd wrote:
> > > Thank god I got through college before "Object Oriented" became the
> > > OMG WTF MUZT HAVE NOW suit buzzword.  When I went through, they were
> > > teaching COBOL, Pascal, and C.
> > When did you go to college?  I've been hearing about OO since 1990.  And
> > the only reason it wasn't a lot sooner was because I'm didn't have
> > access to outside information sources that weren't aimed at C64 type
> > people from the very early 80s.
>
> I went to school at a small liberal-arts college in Oklahoma;
> usao.edu.  A few years after I got out, universities went nuts
> about Visual Basic and C++, etc.
>
> (this was in the days when the entire campus ran on a VAX 4000-700A,
> with a lab full of Dell 486/20s running Pathworks, doing LAT terminal
>  sessions to the VAX...)

Heh.

I learned FORTRAN (and APL) on an IBM 1130 in high school, then took CS in
college. More FORTRAN/WATFIV on an IBM 360/50 running OS/HASP. Also PL/C and
SNOBOL4 (used for teaching data structures). Took a course in COBOL just for
the heck of it.

Hollerith, baby. And I could whip out a drum card in minutes. (Most of the
other CS students didn't know what a drum card was or why they would want to
use one.)

Pascal? Wasn't he some famous science guy or something?

There was also an old IBM 1620 upstairs that I got hands-on access to. Taught
myself machine language on that baby... it had something like 30 operations,
no user registers, variable word size, and it thought in DECIMAL fer
crissakes. Kingston FORTRAN II was practically FORTRAN IV. Learned to
touch-type 026 codes on an 029 keypunch; I was the only guy who could punch
for either machine on any keypunch in the place. In doing so learned why the
360 languages allowed alternate characters for things like parentheses and
equals signs (so you could submit jobs punched on the 026).

In my second year that came in EXTREMELY handy; with more CS students than
ever, the turnaround on student jobs on the 360 was officially 6 hours.
Minimum. I could write and debug my entire project in a few hours on the
1620,
slap some 360 JCL around it, and pick up the working output the next morning.

Wasn't until a few years ago that I learned OOP. Once I got my head around
the
concepts, I fell in love. I find it a useful paradigm for helping me keep
bits
of code separate that should be kept separate, and keeping track of stuff
(like data types) that I shouldn't have to keep track of by hand. My current
set of PIC assembly code looks amazingly OO-like despite the total lack of
direct OO support. Using lots of macros as accessors and I do some data
hiding
by pretending I can't see the data outside the modules. (In embedded systems,
for efficiency, sometimes you just HAVE to do inline code on globals. But I
can pretend they aren't globals by using macros and wearing blinders. :)

Of course, I came into OOP from the bottom up, so I have an appreciation of
the underlying system architecture. (I'm also one of those perverts who
ENJOYS
writing code in assembly. :) And as an experienced embedded systems guy, when
I'm writing code I tend to be fanatical about conserving resources. Though
I'm
much more fanatical about it when my entire environment consists of a 4MHz
RISC processor, 20 bits of I/O, 4K words of ROM, and 96 bytes of RAM than
when
I'm writing code for a desktop computer. :)


---
Sic gorgiamus allos subjectatos nunc.