Would it be so hard to write a short sentence for modules you like. You could even have a list of canned phrases that you would cut'n paste.

Suggestions:

• I use it all the times, it just works,
• it was a little hard to understand how to use the module but it is so powerful that it was well worth the investment,
• a pain to install but a pleasure to use,
• very elegant interface, hides very well the complexity of the underlying code,
• the author is really responsive to bug reports and feature requests,
• offers both a procedural and an OO interface, I usually don't bother with the OO one,
• very nice module, However, I'm not sure I like the name of the module,
• ...

I took my own advice at submit-cpan-ratings - upload ratings to CPAN for stuff you've used.

C:\Documents and Settings\diotalevi\My Documents> submit-cpan-ratings 
+Devel-Foobar\

Modules used:
   DBI-1.21
   POE-0.21
*  WWW::Mechanize-0.12 (already rated)

Comment on DBI-1.21 (enter two blank lines to finish your comment):


Comment on POE-0.21 (enter two blank lines to finish your comment):
I love it. I get the benefits of IPC and/or threads all without having
+ to use threads or forking!


Use this comment? (Y/n):

Rate POE-0.21
  Documentation (0 - 5)? 4
  Interface (0 - 5)? 3
  Ease Of Use (0 - 5)? 4
  Overall (0 - 5)? 5

Thank you! Now uploading your ratings to http://ratings.cpan.org.
[download]

What do you know, and how do you know that you know it? touches on the value of reviews (and how I process them). Let me say something more specific now.

Knowing that someone liked a module tells me very little. Knowing why John Doe liked it tells me a lot more. Knowing that merlyn liked a module tells me more still. Knowing what merlyn liked about that module is really helpful.

The aggregate opinions of strangers whose level of information I don't know has some value, but only if I have no knowledge of my own to bring to the table. If I know Perl reasonably well, just reading the documentation tells me more of value. Since most of us can assume some level of Perl expertise, for many of us a random rating is useless. Good, bad, indifferent. I don't care if I can't tell whether I should care about what you think.

If I bring some knowledge to the table, a review is much more useful. From a review I can judge something about how much this reviewer likely knows, what the module proved useful for, and whether or not it works as the author promises. You can also read that as a list of what I would like you to establish in a good review. Note that it does not take a lot of space to establish this information.

Of course if my aditional knowledge includes knowing about who cares about the module, that opinion has a lot more weight than the most carefully written review can.

And finally, if I know that a person is knowledgable, I'd really like to know why that person liked a given module. There are two reasons for this. One is that it is feedback for me on how to effectively decide whether a given module is good. Secondly it really helps me decide whether this module is going to fit what I need.

In the end, of course, what I really want to know is not how great a module is, but whether it solves my problem. I can live with a module I dislike that just happens to do exactly what I need acceptably well. (I've used Mail::BulkMailer and would again.) I have no use for the best module in the world if it solves a problem that I don't have right now. (Sorry Text::Autoformat.)

So quality is only one factor in deciding whether I'll choose to use a module. An important one, but only one.

Honestly, not only are ratings without reviews harmfull, but when I'm comparing modules and all the review says is "It rocks!" or "It's the best!" and then gives it five stars, that doesn't really help me that much either. I understand you want to give feedback to the authors, but remember that us users try to use these reviews to help evaluate the code.

How about saying a little of what it does, what it does well, what it doesn't do well, and what it needs? I'm not recommending you ramble on for ten or so paragraphs like I do, but a little more thouroughness would really be appreciated.

My intent is not to give the module authors feedback--there are already mechanisms to do that--but rather to give the programmer looking for which CPAN module, of the 7 that fit their purpose, an indication of which is most widely used. In this case, I would seriously examine a module that had a rating of 4 with 75 rankings/reviews over a module with a 5 rating but only 2 reviews.

If there are already 2 or 3 reviews, how much more can be said about the module, and how much more will the searcher actually read? How many people read ALL of the reviews on sites like Amazon? After a point, it seems more useful to me to have a statistically relevant sample size than a more detailed critique. 2 ratings, after all, must be assumed to be the module author and his mom. ;p

--Solo

--
You said you wanted to be around when I made a mistake; well, this could be it, sweetheart.

Right, but what if the module with 75 ratings is eight years old, is well known, and published in three different books, and the other module with five ratings is only a year old but happens to be a vast improvment over the older one?

I agree that you shouldn't repeat what's already been said, but I've seen modules with, say, seven rantings/reviews where not one reviewer ever bothered to say why it was so good, just that it was, and that doesn't help me that much when I need to compare it with something else.

Good module!!!! A++++++++++!!!!!!!!!!! Will use again!!!!!!!!

I believe the correct spelling for that is "!!!!!!!!!!!11".

What bitching?

--Solo

--
You said you wanted to be around when I made a mistake; well, this could be it, sweetheart.

Bad reviews exist on all such systems. Should Amazon ditch their review system because of a few bad apples? Brainstorming over a ratings system that shouldn't be taken that seriously seems silly.