Re^5: Can I please have *simple* modules?

Its not surprising to me that you have found your code suitable to your programming style and thought processes. The real question is, how many other authors find your method factories suitable to their programming style and thought processes. And if every one of us with a custom method factory that we prefer uploaded it to CPAN there would be hundreds of such modules, all slightly different, and all the more useless as the difficulty of figuring out which is The One To Use would become all the more intractable.

Anyway, I should add that releasing such a factory as part of the Rose framework is to my mind totally fine. You arent advertsing it as a generic method factory, and you aren't suggesting that people outside of yourself should use it for anything not Rose. Which seems to me on an abstract level to be about the same as DDS where i have about 10 lines of code that do method generation. I didn't bother to stuff it into a seperate module because DDS isn't a framework. But to me the two are both just utility code used to make our own lives easier. They arent intended to make anybody elses life easier, which is good, because they probably wouldnt.

---
$world=~s/war/peace/g

Comment on Re^5: Can I please have simple modules?

Replies are listed 'Best First'.
Re^6: Can I please have simple modules? by siracusa (Friar) on Nov 25, 2005 at 15:08 UTC
Anyway, I should add that releasing such a factory as part of the Rose framework is to my mind totally fine. You arent advertsing it as a generic method factory, and you aren't suggesting that people outside of yourself should use it for anything not Rose. Actually, I think it's a perfectly acceptable generic method maker. Both the framework and the specific method types bundled with it are suitable for a wide range of purposes. What I meant when I said I'd used them elsewhere is that I've used them outside the context of Rose. It's association with a particular framework may keep people away (or, more like, keep people totally unaware of it), and that's fine. But that's also a separate issue from it's actual utility or merit. As for CPAN being crowded if everyone did what I did, I say the more the merrier. I believe it's a self-adjusting system in the long run. Don't forget, there's also a danger in having too few choices. A not-so-good module may become popular simply because it's the only one of its kind. Better to have the other problem: too much choice instead of too little. Arguably, CPAN crossed that line long ago. Holding modules back at this point is not going to reverse the situation. It's better to try to maximize the value of the alternate scenario. Of course, the usual standards for CPAN submission apply: don't submit a module unless you intend to support it, include thorough documentation, have a good test suite, etc. Those things should always apply. But if you meet those requirements, I say upload away. You'll already be head and shoulders above the average CPAN module :)	[reply]

Replies are listed 'Best First'.

Re^6: Can I please have *simple* modules?
by siracusa (Friar) on Nov 25, 2005 at 15:08 UTC

Anyway, I should add that releasing such a factory as part of the Rose framework is to my mind totally fine. You arent advertsing it as a generic method factory, and you aren't suggesting that people outside of yourself should use it for anything not Rose.

Actually, I think it's a perfectly acceptable generic method maker. Both the framework and the specific method types bundled with it are suitable for a wide range of purposes. What I meant when I said I'd used them elsewhere is that I've used them outside the context of Rose. It's association with a particular framework may keep people away (or, more like, keep people totally unaware of it), and that's fine. But that's also a separate issue from it's actual utility or merit.

As for CPAN being crowded if everyone did what I did, I say the more the merrier. I believe it's a self-adjusting system in the long run. Don't forget, there's also a danger in having too few choices. A not-so-good module may become popular simply because it's the only one of its kind. Better to have the other problem: too much choice instead of too little. Arguably, CPAN crossed that line long ago. Holding modules back at this point is not going to reverse the situation. It's better to try to maximize the value of the alternate scenario.

Of course, the usual standards for CPAN submission apply: don't submit a module unless you intend to support it, include thorough documentation, have a good test suite, etc. Those things should always apply. But if you meet those requirements, I say upload away. You'll already be head and shoulders above the average CPAN module :)

[reply]