Paul M Foster wrote:
> On Wed, Feb 17, 2010 at 01:52:08AM -0200, Manuel Lemos wrote:
> <snip>
>>> Having to register to download classes from is a
>>> nuisance. Manuel says this is up to the individual developer. This may
>>> be technically true, but Manuel *offers* this as an option. Contrast
>>> Source Forge, which performs a similar function but does not require any
>>> registration to download anything. I imagine that the registration
>>> allows Manuel to tightly monitor site usage in a variety of ways.
>> Actually it is much more than that. If you download a package, the site
>> keeps track of that and next time the package is updated, it send you an
>> e-mail alert so you can get the latest version, unless you do not want
>> to be notified of course.

per project rss feeds?

>> Also the site counts how many distinct users downloaded each package and
>> builds top download rankings . If you download a package more than once,
>> it only counts once, so the top download charts are accurate, making it
>> fair for everybody.

no more accurate than storing the ip (or a hash of it); people forget
details sign up again and so forth - can never guarantee accuracy here

>> For users this may not be very important, but for authors it is very
>> motivating. Authors are happy that the site lets their users know about
>> updates of their classes. Authors also like to see the progress of their
>> packages in terms of users that have downloaded it.

can be done with the aforementioned, no need for logins

>> The site also provide a blog for each package, so when the author wants
>> to post something new about the class or ask for feedback, a message is
>> sent to the users that downloaded the package.

good; but again rss & offering an option to subscribe by email.

>> There are other compelling reasons but this is basically why more than
>> 2600 authors submitted over 5000 packages. The site gets them attention.
>> Other sites like Sourceforge cannot provide this level of attention
>> precisely because they do not require users to download even if the
>> authors wanted that.

I'd debate this; but trying to stick to positives - sf do pretty well
with their stats is all I'll say

> This type of question has been asked many times on this list,
> particularly for "voting" type projects: How do I ensure that a person
> can only vote once (etc.)? No answer I've ever seen, besides insisting
> on a registration/login, has ever been satisfactory. The above is a
> real-world example of this in action. And as Manuel details, it has
> some definite benefits to users and developers.

ack; fact is if you wanted to skew results you'd just create lots of
accounts - distinct ip and cookies can cover it just as well; there is
no "perfect" solution.

> Again, having to register/login is a pain. But ads are a pain, too. It's
> a trade-off.

always need to figure out which has more benefits though; more downloads
and exposure (better conversion ratio) vs better stats and less exposure
(low conversion ratio) + account for things like people downloading
because they can see the source ala google code - i for one always check
the svn browser on google code before downloading, sure many others do too..


PHP General Mailing List (
To unsubscribe, visit:

Reply via email to