perl-diddler has asked for the wisdom of the Perl Monks concerning the following question:

I was wondering what was wanted in the way of scripts for CPAN. I write a fair number and occasionally I wonder if they'd be of any use to others -- though usually I think they aren't polished enough to even make the attempt.

The script I had problems with the other day, I've gotten to a point that might make it worth a '1.0', but it's so specific to my needs I don't know how much I'm deluding myself about it being something that others might be interested in. OTOH, I've seen a wide range of code on CPAN from stuff way over my head (but useful), to stuff that seems very 'arcane' and has me scratching my head thinking "well, I suppose there are some people who might find it useful, and probably if there is even 1, it was good to have posted it..."

I don't know if anyone remembers the vague content, but the script 'iomon' is linux specific, and requires read access to the /proc/diskstats file.

Other requirements:
reads dir /dev/mapper.
wants 'mount' and 'stty' to be in the path for info requirements.

Is this the type of thing useful for the scripts section?

Replies are listed 'Best First'.
Re: requiremens for CPAN scripts?
by JavaFan (Canon) on Apr 07, 2011 at 09:14 UTC
    Anyone can upload anything on CPAN. If you have written something and you want to make it available, upload it. Don't question yourself whether it will useful enough. If it isn't, people won't use it, and it'll rest on some disks, taking only a fraction of the space of an picture.

    Don't worry. Just upload.

Re: requiremens for CPAN scripts?
by anonymized user 468275 (Curate) on Apr 07, 2011 at 11:13 UTC
    Although scripts might potentially be useful to people as-is, it is far more likely that a generic version of the key algorithm(s) you are offering in the form of a module will help other people.

    For example, a script that takes two arrays of numbers as I/O input and returns the correlation co-efficient is not as useful as a (callable) module to do the same. This is because, unlike a script, such a module can be reused hundreds of times in sequence or parallel without resorting to a relatively slow I/O interface nor the repeated spawning of a pair of shell and perl processes.

    Furthermore, the use of an object-oriented interface permits the use of inheritance so that people can customise outside the box, without having to edit your script. In this way your contribution, if "sexy" enough, would grow in its influence.

    One world, one people

      I'm not sure how one would make a 'generic' module out of a program like 'top' or 'xosview', (for example).

      They are designed to be run and left running for some time, enabling you to monitor what's going on.

      How would you make that into a module useable by other programs?

        I couldn't find a CPAN script called top - but hypothetically if one exists and does broadly what unix top does: I would say: a script that produces data on top-consuming processes is not as valuable in the longer run as a module that can be inherited and locally augmented to monitor additional information about the processes or to filter the processes on different criteria, e.g. I/O instead of CPU, than what top does - a script has a limited application, a module has no limit.

        One world, one people

Re: requirements for CPAN scripts?
by toolic (Bishop) on Apr 07, 2011 at 19:16 UTC
place for getting comments?
by perl-diddler (Chaplain) on Apr 07, 2011 at 19:12 UTC
    Do people post code for 'comments' here, or just for seeking answers to problems. I.e. If I wanted to get feedback, ideas for improvements, or have a casual review of code, is 'seekers of perl wisdom' a place for that, or is there another forum?

    I wouldn't want to offend anyone by assuming anything.

      Meditations:
      it is generally considered appropriate to prefix your node title with "RFC:" (for "request for comments").
        Well, that screw that idea sounds like... ;-)

        glad I asked first.