Re: [Chicken-users] Rewriting Svnwiki in Java or PHP

[Graham Fawcett]

On 4/1/07, Nelson Castillo <address@hidden> wrote:
>   This way, even junior programmers with little programming knowledge could help
>   us grow the number of lines of code that Svnwiki currently has.

With respect, Nelson, making it easier for junior PHP programmers to
"grow the LOC" of a project sounds like a recipe for disaster.

It can be healthy to have a second implementation of anything.
Rewriting it in PHP isn't going to magically attract a horde of
developers, though.

I just read Alejo's blog post:
http://wiki.freaks-unidos.net/weblogs/azul/rewriting-svnwiki
where he raises these points:
  1. it has been proved to scale to a great scale,
  2. we will very soon have a free software implementation,
  3. it is a very well-known language (ie. a lot of people know how
to code on it),
  4. it tends to make the code very easy to maintain and
  5. it will probably run significantly faster than PHP.

I'd like to respond, here among friends:
#1 has more to do with the JVM than Java; and other things have been
proven to scale (even PHP). There are other JVM languages if you are
looking for choices.
#2 isn't a reason to choose Java, merely one less reason not to choose Java.
#3 is definitely true.
#4 is extremely arguable. Most serious Java projects I've seen have
degraded into an unreadable pile of abstraction layers. There is such
a thing as too many interfaces. And the code is not succinct at all;
lots of maintenance code needed (e.g. for exception handling). IMO,
Python and Ruby have a good sweet-spot for short, generally-readable
code.
#5 is a reason not to pick PHP, but it's also a reason not to leave Scheme.

Pick whatever language you want guys, and best of luck: it's the
developers not the language that makes the difference in the end. (Me,
I think I'd write a Scheme DSL for writing subversion-based wiki
applications, and show my junior PHP developers what a Real Language
can do. ;-)

Good luck,

Graham