Survey of dsource projects

[Robert Fraser]

David Wilson wrote:
> I don't understand, what is the use case for these stats? The
> sourceforge.net project stats have been in all the years of that site,
> practically useless. Measuring commits, LOC change, downloads forum
> posts, etc., is no good measure of a project's "rating" or
> "progression". I just don't see the point.
> 
> I see no reason why a project that hasn't been touched in 10 years,
> but has a stable, mature code base should be ranked lower than some
> college project that is getting 100 feverish commits a day from some
> OCD victim making whitespace changes.* :)
> 
> +1 for alphabetical order or some sane categorisation scheme
> -1 for anything other than a nice summary page showing recent commits,
> forum topics, etc.

For D, at least for libraries that are intended to be distributed in 
source form (and even many binary ones) if the last commit was more than 
a few months ago, they're likely not to compile on modern compilers. 
Even though D1 is supposedly "stable", there have been a number of 
breaking changes (the version/extern "fix", the .init "fix", etc.). 
Source libraries that won't compile, while they can still be useful, are 
difficult for a new user, and so their inactivity should be noted.

I find stats very useful. While it's true that stats can be messed with, 
in the general case a project that has a high number of commit and LOC 
changes is likely being "actively developed". Does that mean it's more 
useful? Certainly not, but that's something users might like to see, or 
even sort by.

It also helps users find projects that are very actively developed but 
whose developers don't post status updates on d.D.announce very often 
and foreign-language projects whose developers don't use the English 
forums here.