Perl and JSP

[Saran.j.jegan]

Hello,

we have developed an web based application for monitoring
and controlling some machines using Java & JSP in apache tomcat, now
we have planned to replace some java programs such as serial port
communications and socket level communication with CGI scripts using
perl & c , Hope it will be faster than old java codes,Am in need to
know inserting CGI is advisable and whether it will be faster &
reliable, can any suggest your experience in similar cases

 

[xhoster]

Hello,

we have developed an web based application for monitoring
and controlling some machines using Java & JSP in apache tomcat, now
we have planned to replace some java programs such as serial port
communications and socket level communication with CGI scripts using
perl & c , Hope it will be faster than old java codes,

I see no reason to think that a *good* Perl implementation would be faster
than a *good* Java implementation. And for sockets, I would expect that
the bottleneck is at the system level anyway (or at least, I assume that if
the bottleneck wasn't at the system level, it wouldn't be related to the
sockets in the first place, but at what you are using the sockets to do,
and since you specifically mention sockets and not what you are using the
sockets to do....), and so even C probably wouldn't make much of a
difference.

Am in need to
know inserting CGI is advisable and whether it will be faster &
reliable, can any suggest your experience in similar cases

You can write slow and unreliable code in any language.

My Perl-CGI tends to be more reliable than the JSP I've seen, and when
speed was considered an important design criteria, my CGI were generally
faster than the equivalent JSP I've seen. I suspect that this is largely
because I'm a better programmer than the people writing the JSPs are. Some
of it may be that with Perl's flexibility and speed-of-programming, I could
try various techniques and choose the best one, while that is more
difficult with JSP (at least for me).

For some kinds of things, achieving high performance will require mod_perl
rather than Perl CGI, so if you are unwilling to use mod_perl or its
equivalent than that is one check for JSP. Also, if you are using Windows
servers, I think that would be another check for JSP.

But mostly it comes down to asking yourself what language are your best
programmers best at programming?


Xho

 

[Saran.j.jegan]

I see no reason to think that a *good* Perl implementation would be faster
than a *good* Java implementation. And for sockets, I would expect that
the bottleneck is at the system level anyway (or at least, I assume that if
the bottleneck wasn't at the system level, it wouldn't be related to the
sockets in the first place, but at what you are using the sockets to do,
and since you specifically mention sockets and not what you are using the
sockets to do....), and so even C probably wouldn't make much of a
difference.


You can write slow and unreliable code in any language.

My Perl-CGI tends to be more reliable than the JSP I've seen, and when
speed was considered an important design criteria, my CGI were generally
faster than the equivalent JSP I've seen. I suspect that this is largely
because I'm a better programmer than the people writing the JSPs are. Some
of it may be that with Perl's flexibility and speed-of-programming, I could
try various techniques and choose the best one, while that is more
difficult with JSP (at least for me).

For some kinds of things, achieving high performance will require mod_perl
rather than Perl CGI, so if you are unwilling to use mod_perl or its
equivalent than that is one check for JSP. Also, if you are using Windows
servers, I think that would be another check for JSP.

But mostly it comes down to asking yourself what language are your best
programmers best at programming?

 

[Saran.j.jegan]

Ofcourse Your points are more obvious ,we are in plan to make
the code to run in both windows and linux,
i think it will be a good experimental one ,certainly its going to be
implemented on a industrial PC which will be more compact one with XP
embedded and linux , so removing java fundas will give some space and
perl's flexibility will help a lot, as you said it also depends on the
programmers hands with the language

 

[J. Gleixner]

[...]
i think it will be a good experimental one ,certainly its going to be
implemented on a industrial PC which will be more compact one with XP
embedded and linux , so removing java fundas will give some space and
perl's flexibility will help a lot, as you said it also depends on the
programmers hands with the language

The word boondoggle comes to mind.

You have something that works on a reliable platform and now you
want to change everything to see if it will run a little faster? It
seems like a complete waste of time and money, to me. IMHO,
your effort would be better spent in figuring out ways to speed
up your current system.

Don't get me wrong, I use perl for everything, but when your
company's expertise is in Java/JSP, switching to another
language and system could be a disaster and put you further
behind in figuring out the actual problem.