[racket-dev] ohloh page

15 minutes ago, Neil Van Dyke wrote:
> 
> Below is very opinion-heavy seat-of-pants reaction.  I'm not
> familiar with Ohloh, and I could be off the mark...
> 
> The Ohloh site seems oriented towards PHBs who use 'analysis' like 
> "decreasing year-over-year development activity" without wondering 
> whether it's at all accurate and meaningful.

They're pretty good, I think, about showing the caveats.  For example,
in the code history (and project cost etc) there is a comment that
says that often they see only part of the history, which you can see
with an initial spike of code.


> (Yes, that's what they say about Racket, with a yellow warning
> graphic.

That's because they count only source code -- and that includes only
known languages.  (Which is actually a good decision.)  This explains
the two big drops in code: one was renaming *.ss file (which they
recognize as Scheme code) to *.rkt (which they don't know about).  And
since we're left with very little code that counts, the second drop is
probably Matthew's revision of the GUI -- eliminating a ton of C/++
code which was a significant part of the little code that they see.

Hopefully they'll get John's patch in which would make things look
closer to the reality.


> And only 2 people on that site claim to use Racket, which just looks
> funny.)

Yeah -- but that's just people who bothered creating an account there
and declare that they use it...  Note that the "plt-scheme" project
has 21 users, which should move over once my request for marking it as
a duplicate of "racket" goes through.


> A quick poll of open source developers I bumped into, including one
> loosely connected to Ohloh, said nothing positive about it.
> Personally, I would invest a little effort in making Racket look
> good on Ohloh now that we're there, since that might serendipitously
> come in handy for a PHB someday.  But I don't know that I would link
> to Ohloh or mention them outside of their own site right now.  Too
> much like those tacky bogus "Top x% Web Site" awards in the late
> '90s.

Yes -- they started in a promising way, then it got neglected and
bit-rotted as well as internet-rotted away to near extinction.
Recently they started working on it again (changed owners or something
like that), so maybe it'll catch on.  (Both then and now I got there
after I've seen a bunch of random mentions.)

Another thing that might catch on is masterbranch -- coming from
stackexchange: https://www.masterbranch.com/racket-project/497230
But it too is unaware of Racket -- it says:

  Mainly written in C, make and HTML. A total of 16 developers made
  931 commits

(I didn't see any way to add a language there though.)

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!