[jdev] The State of Our Code-bases

maqi at jabberstudio.org maqi at jabberstudio.org
Sat Aug 28 14:40:25 CDT 2004

On Sat, 28 Aug 2004, JD Conley wrote:

>>> With all of these pieces laying around half complete, why aren't we
>>> picking them up and giving them a nice and polished finish?
>> Because there's no one able to pick them up?
>> When these Jabber projects die, they tend to just stay dead unless the
>> original maintainer resurfaces.  Getting the code up to par is not a
>> matter of focusing on what we have (because I think we do this already),
>> it's a matter of resources.

I think a major problem is that much Jabber software is written by small
teams. For example, AFAIK only one person is working on jabberd2 (as is
the case with jabberd14, too). Therefore, much know how stays implicit,
preventing interested people from joining development. For example, code
documentation is almost absent in most projects (one has to say that the
situation improved a lot in the past though as for example both jabberds
have at least some code documentation now as well as rather good
user/admin documentation).

> See http://www.saint-andre.com/blog/2004-08.html#2004-08-16T15:49  Your
> opinion is shared by many...
[Saint-Andre mentioning Python]
Some time ago, I unwillingly started a programming language bashing thread
in jdev when I said that the C implementations are a pain and waste of
time. Just look at the transports which are a nightmare from both a
security as well as a stability point of view (but again, one has to say
that the situation improved a bit in the last years). That's why I'm
especially happy with James Bunton's new transport implementation in
Python. Just compare the readability of this code with the old MSN-t one
:-). One can easily learn Python *and* start hacking on PyMSN-t in the
time required just to get an idea on how the old C transports work :-).
And don't forget that with Python (or, in fact, any other decent framework
or high-level language) you do not have problems with exceptions or
memory leaks, two things that make most C programs unusable.


More information about the JDev mailing list