A simple request for a server... - Wikitech-l

3 Oct 2005


      The request for a server is simply a request for a central repository of
Wikimedia scripts that are used in the maintainance of any of the
Wikimedia servers, such as Wikipedia.  As it current stands, it is
contributors who run their own copy of their scripts located somewhere in
the world.  Should that person ever disappear from the project, it is
likely that his/her useful source code that was written disappears
forever.  It is immensely difficult to classify this as a SourceForge
project, especially when it is many different pieces of code running for
different purposes.  Additionally, there is a greater concern to not
release some of the scripts to the general public for fear that they may
be used for vandalism purposes.  However, it should be noted that the code
is still useful to someone else who wishs the time and effort to
continuously update it, especially with the changing MediaWiki software.
Furthermore, there is no easy collaborative way to update source code
between contributors.
For example, I have several scripts that run a specific purpose.  If
wished for someone else to take the task of improving my script, I would
have to ask them personally and they would have to send the changes to me.
 There is no easy way to merge changes, like a CVS or SVN service.  We
have many programmers who contribute to the Wikipedia, who write in
various languages, in order to perform a certain maintainance task.  Their
skills may not be up to par in order to become a developer for the
MediaWiki software, but they, as do I, want to use their programming
skills in some manner to help one of the Wikimedia projects.  It would be
immensively useful for a single Wikimedia server to be used as a place for
trusted Wikimedia individuals to perform collaborative programming
projects in the assistance of maintaining various Wikimedia pages.
One of the things about source code that is so great is that it is
universal and can be read by anyone who understands the programming
language.  Case in point, if I wrote a simple bot script that reset the
Sandbox on the English Wikipedia, and someone from the Chinese Wikipedia
wanted to adapt it for the Chinese Wikipedia, he or she, as long they were
a long standing trusted Chinese Wikipedian user, could take my code, and
readapt it for their purposes.
Case in point, English Wikipedian User Kevin Rector wrote KevinBot to do
some bot work, that is explicitly written in C++ (I think).  But his code
is now out of date due to the changes in MediaWiki code causing his bot
code to break.  Someone asked me to update his bot and send me the
executable, but that doesn't help me since I can't make any changes to the
source code to correct his bot.
Part of the frustration of bot code writing is that the MediaWiki code
changes to the point that bots break.  And bot programmers are people who
struggle to maintain their own scripts... they don't have a group of
people assisting them in the effort.  If bot scripts can be updated with a
group of people just like the MediaWiki software, it would be less
frustrating and the bot code would be updated faster.
As it stands, bots are scripts and/or programs that run on individual
computers and servers maintained by Wikipedians.  If anything happens to
them, that small piece of knowledge in which the script is written in
disappears.  And if, for any reason, their computer breaks or their server
breaks... their bot is inactive for an unknown period of time.  There is
no fallback.  Furthermore, their server specs may not be something that a
university or this organization can get.  The power, speed, and stable
connectivity is something that individuals can not achieve on their own.
Granted, if someone leaves, someone can rewrite the work that they did.
But the point is that they are reinventing the wheel just for that to
work.
I don't know where else to ask this, so I figured to email it to this list.
----
Jason Y. Lee
AKA AllyUnion (English Wikipedia Administrator)