[SATLUG] thoughts on languages

[Bruce Dubbs]

Travis H. wrote:
> On Tue, Dec 05, 2006 at 12:11:10PM -0600, Bruce Dubbs wrote:
>> I can go on and on detailing the problems with SLOC as a measure of
>> productivity.  However, it *is* a better measure than nothing.  I just
>> have not heard of anything better.

> Certainly there's a lot more PHP code out there, no doubt about that.
> It's very easy, but it's hard to use securely, much less properly,
> and the bar is set _very_ low.  For that matter, there's more HTML
> than PHP probably... I'm not sure why that should matter.
> 
> I'd have to agree with Justizin, lines of code is something you
> should try to minimize, if you were to hold the functionality the
> same.

If you want to mimimize SLOC, have you ever looked at Forth?  :)

There have been all sorts of studies that agree that the more you can do
with a line of code, the more "productive" you are.  There are other
issues too.  You don't write a kernel driver in perl or as a bash
script.  You don't write a word processor in assembly.

The problem with misusing PHP as a web application is not, IMO, due to
PHP.  It is more that the users don't understand HTTP (and possibly
mysql) and how it can be misused.  This is especially true because
relatively intelligent people with no training can do "something" with
PHP.  Most people can fly an airplane for a while without much training.
 Its just easy to get into trouble (Takeoff is optional, landing is
mandatory.)

What are the options?  You can increase safety by removing
functionality, but experts then will find it difficult to do what they
want.  That goes pretty much against the Unix philosophy.  The idea here
is the responsibility should be placed on the programmer/user, not the
language developer.

Now, as root, on your system, do:

rm -rf /*

Bash (or rm) certainly lets you do that.  Should this be disallowed?

  -- Bruce