[Ifile-discuss] Re: Effect of widely differing volumes on ifile classification

[clemens fischer]

Jason Rennie <address@hidden>:

[on skewed data]
> address@hidden said:
>>  btw, is there an URL to that paper you mentioned?
>
> It's been submitted to a conference, so I'm not making it available
> on the web yet, but I can e-mail you a copy.  Do you prefer
> postscript or pdf?

the format has to satisfy the following contraints:

1.  produce (compressed) files less than 100kB,
2.  be viewable by xpdf(1).

oh.  too bad:  i just checked if xpdf(1) displays postscript, but it
doesn't.  if somebody knows of a free postscript viewer working on
unix that is less awkward to use than ghostview, i'd really
appreciate it!

the space requirement is an unfortunate result on my experimenting
with anti-spam techniques.  consequently i set up and use my own
email system, which is well known to spammers and hackers.  i've had
revengeful people sending me megabytes of noise using up the full
bandwith, so i had to narrow that window of vulnerability.

  clemens