hi!
from time to time, i try to install and run freenet. today was such a day,
and i do not think that the participants in this project will like what i
have to say.
first of all, i do not want to talk about the features or capabilities of
the network and service design. i want to talk about the project,
the progress made so far and what conclusions i draw from that.
now. every few months, i try to set up a local transient freenet node, and
would like to try out basic operations. result: i get some obscure
java services running, with no documentation whatsoever, and data retrieval
or insertions utterly fails, partly because the network setup might be wrong,
node references missing, or known nodes being swamped, partly because of
bugs in the program itself, partly due to missing information.
this basic setup did not change at all over the last year or so.
with such a scenario, missing documentation, no test environment and no
working and verifyable demo cases, i as a software developer am too much
pissed off to get involved into any further development, because in my life
i have had enough hours of coding work being wasted on projects set up
this badly, which will lead to complete and utter failure caused by
starvation in about 120% of the cases.
what i see is that you are trying to create the network itself and the
server software at the same time, which will not work because there are
too many potential error sources in this environment to sort out any
existing problems.
you do not make any progress, because as good as nobody is able to help with
the development for lack of documentation and useable test environments.
you do not get powerful storage nodes because nobody lets this buggy java
software swamp their server for long.
people trying the software get quickly annoyed because it simple is not
useable.
my conclusion from this is:
you can't build the real net while the server software is that bad.
to get enough help in improving the software, you need to provide good
user, administration and development documentation and help potential
developers with test cases that can be verified locally,
be it on single machine or small local network environments.
with the current speed of development, you might come near to the
HURD experience.
so i would opt for an emergency break. revise the project organization,
try to prepare a sound development environment and write enough initial
documentation to get people on the train. also think about the prototype
implementation platform.
maybe i am wrong. but this would mean that i got it all wrong, which
again would prove that i am right at least in large parts, wouldn't it?
i would really like to support the project, but it simply does not make any
sense to me the way it goes right now.
--
peter koellner [EMAIL PROTECTED]
___
support mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support
