[Spambayes] SpamBayes for Eudora

[Tony Meyer]

> Sorry to keep bothering you but I've followed the
> directions you referred me to as far as I'm able:
> 1) installed Python
> 2) downloaded SpamBayes
> I'm stuck trying to run "python setup.py install"
> using Command Prompt.  If I run it from the python
> directory, it can't find the setup.py file.

Kenny was right here (I missed that you were in the wrong directory), and
you need to run it from the spambayes directory.

> If I try running it from the spambayes directory it
> can't find python.

So you need to tell it where Python is.  Either explicitly for that one
command: "c:\python23\python.exe setup.py install", or put it in your PATH:
"set PATH=%PATH%;c:\python23" "python.exe setup.py install".  (Obviously, if
you installed Python elsewhere, you'll need to modify those).

> More generally, I don't know if you guys are really 
> interested in having a large set of users, but if you are,
> it seems to me that you need to have MUCH more explicit,
> detailed directions or a Windows install program.

1.  This is pre-release software.  I think that if you looked at alpha
software in general, you would find that documentation and setup were the
weakest points of the software.

2.  It's expected that anyone that isn't capable of running a python script
will want to use a Windows installer (i.e. *nix and Mac users are probably
capable enough of doing that), and so there is one of those for them.  The
one that includes more than the Outlook plug-in (i.e. sb_server et al) is
only an experimental build at the moment, but that's likely to change very
soon, and IAC, the whole project is still pre-release.

3.  The most successful open-source software (IMO) is that which somehow
encourages people to submit improvements, including to the documentation.
If the instructions are confusing, then what we need is for people that find
them confusing to tell us what needs to be changed - explicitly - not just
"they aren't detailed enough".  We're the developers - we know how it works,
and it's difficult to know exactly what a non-developer will have trouble
with.

4.  If you compare the documentation to the first 1.0a1 release, you'll see
that it has improved, and will continue doing so.  We are well aware that
this is one of the biggest weaknesses - we explicitly say so all over the
place, including in the documentation itself!  Hopefully, by the time we get
to a 1.0 release, it'll be adequate, if not better.

> I'm supposed to have been a professional programmer
> from 1950 to 1991, although I did precious little
> actual programming.  But I did lead the group that
> produced Fortran in the 50s and another group developing
> the functional programming language FP in the 80s.
> And I've been using PCs since they existed.  So it seems
> like a bad omen for wide usage of SpamBayes if I have to
> struggle for hours trying unsuccessfully to follow your
> directions

I'm sorry, but it seems an incredibly bad omen that you therefore don't know
that to run a program it has to be on the PATH, or an absolute PATH given.
Sure, many people won't, but those people would be expected to use an
installer program, not run commands in a console environment.

=Tony Meyer

---
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes. This
way, you get everyone's help, and avoid a lack of replies when I'm busy.