Re: [Quote] When should I use Perl?

[Old_Gray_Bear]

When should you use Perl? When should you use any scripting (interpreted) language? Why for the "First System", of course. You know the First System, it's the one that Fred Brooks (The Mythical Man-Month) says you will throw-away.

• The First System is the 'management simulation' for the Powers-That-Be; something they can look at and decide that the title-bar should be puce and not magenta.
• The First System is the 'limited functionality' demo for the End-Users; so they can see how the various widgets interact with each other.
• The First System is the one with the 'abstract data-base interface'; so that the DBA's can try different databases out and find the best one for the User's needs and Management's requirements.
• The First System is the 'We Think It'll Look Like This' system that Operations uses to write the Run-Books and Procedures from.
• The First System is the one that supplies 80% of the needed functionality after a week's worth of work by two people.

And once you have your "First System" up and running, then is the time to decide whether or not it's worth while to re-write something works and something that people are familiar with in C/C++/Java.

I am very fond of building mock-ups of The System in Perl so the various parties can all get their hands on something early on in the design process. Changes are much easier in scripting languages. I can give Them almost instantaneous feed-back ("Here it is in magenta; (click-click) now here is puce"). And if my "mock-up" is good enough, I get my System designed and out the door that much faster.

Yes, it's not as "fast" or as "sexy" as it might be if it were written in C/C++; but it's here, right now, in your hands; not sixteen months (if we are very lucky) away. Yes, The System is written in Perl (or Python or Visual Basic or REXX) -- so? It serves our needs now while the new super-whiz-bang is being developed by the Crack Infrastructure Development Team. Think of it as a bridge between today and the day that the Crack Team finishes....

I have had a lot of experience in bridge-building. And I build them as if I will be using them still in five years time. I may over-engineer my bridge a bit, but I get to sleep better at night because of it.

[blazar]

Good example, and it may have well been given in answer to the original question. All in all seems like a particular instance of optimizing developer time...

[bibliophile]

"Bridge building"... I like it.
It did leave me with an image of a mile-long bridge extending out into the ocean, with nothing on the other side...

"The Crack Infrastructure Development Team needs to push back implementation another three months."
"No problem. We can just add another span to the bridge!"

:-)