> I had a big argument over something Dave Winer said a few years ago and
> he captured my rant and posted it at
> http://static.userland.com/misc/timBrayOnSdk.html
>
> For what it's worth, I think that at the architecture level, the choice
> between demanding to see the bits-on-the-wire and being satisfied with
> an API is probably the deepest one there is. -Tim
>
Tim,
Interesting read, and I agree and disagree at the same time.
An API can be elegant and hide some of the underlying complexity, (not the
window, or xwindows API's for sure).
But in most cases, programming in a API you give up some control of how
somethings function to how the designer of the API decides. In some areas
this is a good thing, Giving an API to simplify the handleing of messageing
in windows for example, instead of handleing every message in the
program itself.
It comes down to a simple thing actually. At what level do we allow programmers
in. Should every programmer have access to the lowest levels of the system
and control the bits. Or should some programmers be locked out of that level
by an API, to offer a more Secure and Stable platform to program on (not that an
API always offers that)
Personally I remember when windows 3.1 came out and the programming module for
it.
I came from programming DOS programs, where I basically had control of the
system at
all times when I wrote a program. Not that windows 3.1 was a true multitasking
enviroment, you still had to give control to the underlying system. And believe
me,
compared to the programming module for then and what we have in 9x and NT. it is
much better. But we are also dealing with a mostly closed source product in
that
enviroment.
In the Linux enviroment. anyone and everyone can program even down to the lowest
level of
the kernal, if they so choose. But most programmers don't have that skill to be
at that level.
Some don't even have that skill to be beyond the api level of xwindows.
While it is nice to be at that level and have that type of control. At times it
isn't warrented.
my bit level programming days were when I built a 4 bit computer from discreet
componets
and and designed the machine language for it. Even wrote a simple assembly
langauge for
it to make it easier to program. Or programming a remote system for some early
robot controls.
While those controls are good, for most people I would probably design something
at a higher
abstraction layer. Not to many people program these days at the assembly
language level.
depends on the needs and the business module. As you said about Java in the
article.
Douglas
