Re: Debugging

[Aaron Turner <aturner@linuxkb.org>]

Jason,

Who writes web apps in C anymore?  EVERYONE is doing ASP, PHP, mod_perl,
Java Serverlets, or some other high-level language.  The reason is that
these tools are making the execution speed difference gap smaller and
smaller, whiles making the time-to-market faster and faster.  No, I
wouldn't write a disk-defragger in Perl.  Nor, would I build a web site
engine like the LKB in C.  

I think you will quickly learn that the speed advantages you'll find in C
are not only less than you expect, but greatly outweighted by the
requirements placed on the coder. As long as I am a part of the LKB, I
will not support a port to C or C++.  It just doesn't make any sense for
us to move in that direction while the rest of the world is moving in the
opposite.  

I feel that a C/C++ port would prevent us from ever gaining a reasonable
developer base, and that is one of my major concerns.  Go back and read
the emails from the developer interest forms.  Notice how few hours people
can spend on our project.  If you can get people interested in working on
the project because it will help them in their professional carrers,
you'll find people are much more interested in helping.  Right now people
who know mod_perl are in VERY HIGH demand and are commanding premium
salaries.  Working on the LKB would be very interesting to them since it
allows them to show off and hone their skills.  I'm sure the same could
be said of JavaServerlets coders as well.  There aren't many web C++
positions from what I can tell.

Anyways, we've hashed this enough I think.  It's obvious we won't agree,
and that's ok.  We'll worry about that when the time comes.  We have more
pressing issues to deal with for now anyways.

Anyways, my email/web/irc access for the next week will be limited at
best.  I'll try to keep in touch as I work on the CmdSchedule script.

-Aaron

On Mon, 6 Mar 2000, Jason Pincin wrote:

> > And franky, I'm not sure OO is this silver bullet you make it out to be.
> > OO programs are slower and OO adds additional complexity and overhead.
> 
> Don't believe me.  Look at the truely LARGE projects goin on... they all have an OO design.  There's
> reasons for that.  OO adds complexity?  Hrm.  Well if we hadn't implimented an object system, we'd be
> SOL right now when small things changed... object systems protect you... but nevermind that...
> 
> > My point isn't that we should choose Perl over Python or C++, or OO vs
> > non-OO.  My point is that you need to get developer buy-in on whatever
> > language we use.  If you want to get out of coding and start looking at
> > more big picture things, then this should be very important to you.  I
> > can't stress this enough. If you want to code forever, then do what you
> > will.  If you want to get out of coding like you always tell me you do,
> > then let the developers decide what to do.
> 
> I'm not looking at this from an LKB perspective when I say C++.  When I re-write the code, it will NOT
> be for the LKB.  It will be for a seperate project entirely... and it will be an application, not a
> service, and it will be C/C++.  My hope is that when that project is usable, the LKB will migrate to
> it... and I don't see any reason why it wouldn't... for I never plan to shrink my involvment in it...
> so...
> 
> Thats why the language isn't open for discussion.  :)  For an actualy application such as what I have
> in mind... a server/daemon... Perl/Python/PHP... nothing can touch the speed of C.  And that's what'll
> matter too me then :)
> 
> -- 
> Jason Pincin
> Linux Knowledge Base Project Leader (http://www.linuxkb.org/)
> http://vodka.linuxkb.org/~chardros/
> 

--
Aaron Turner, Core Developer       http://vodka.linuxkb.org/~aturner/
Linux Knowledge Base Organization  http://www.linuxkb.org/
Because world domination requires quality open documentation.
aka: aturner@vicinity.com, aturner@pobox.com, ion_beam_head@ashtech.net
The difference between `Unstable' and `Usable' is only two characters: NT