Jackie Dietz counted messages to EdStat in the week of March 26
and found that of the 74 valid messages, 32 were sent originally
to EdStat and 42 were sent originally to sci.stat.edu.  Thus if
the link between EdStat and sci.stat.edu is disabled, users of
these two services will see roughly only half the messages they
are currently seeing.  This implies that allowing the link to be
disabled is counterproductive for people interested in statisti-
cal education -- we lose the synergy of the larger forum, and
each forum becomes somewhat narrower and perhaps more "provin-
cial".

For example, as Jackie noted, allowing the link to be disabled
means that the readers of EdStat will no longer receive the in-
formative messages from Rich Ulrich, Herman Rubin, Radford Neal,
Duncan Murdoch, Paige Miller, Robert Lundqvist, Art Kendall,
Aleks Jakulin, Jerry Dallal, Eric Bohlman, Glen Barnett, Thom
Baguley, and others.

In addition, allowing the link to be disabled means that readers
of sci.stat.edu will no longer receive the informative messages
from Paul Bernhardt, Don Burrill, Robert Dawson, and others.

Some writers have pointed out disadvantages of keeping the link.
I discuss disadvantages that I'm aware of in the appendix.

If we wish to keep the link, we have a problem:  The company that
maintains the link has not responded to Jackie Dietz's e-mails.
This may be because the company (which is an Internet company
with no interest in statistics education) sees little commercial
benefit in maintaining the link.  Thus when the EdStat server
moves to its new e-mail address it may not be possible to ask the
company to reconfigure the link to use the new address.  Thus if
we wish to keep the link, it may be necessary to build another
link.

(We might escape building a new link if NCSU could forward the
messages from the existing link to the new EdStat address and if
the existing link would accept e-mail messages from the new Ed-
Stat server.)

If we wish to build a new link, this will require programming and
testing.  Help in building or maintaining the link might come
from a professional organization with an interest in statistical
education (e.g., ASA, the Statistical Education Section of ASA,
ISI, IASE, IMS, or a statistics department at a university).
Help might also come from a statistical software company, which
would presumably have the necessary programming ability and which
might also be interested in helping the statistical education
community.

In view of the loss of synergy if the link between EdStat and
sci.stat.edu is disabled, I very much hope that we can keep the
link.  Thus I hope that someone can provide a reasonable solu-
tion.

If the solution requires funding (e.g., to pay a programming firm
to build a reliable long-term link), MatStat may be able to help.

Do you know someone who might help?  Can you bring them into the
discussion?

Don

------------------------------------------------------- 
Donald B. Macnaughton   MatStat Research Consulting Inc
[EMAIL PROTECTED]      Toronto, Canada
------------------------------------------------------- 


APPENDIX: DISADVANTAGES OF KEEPING THE LINK

Here are my thoughts about disadvantages of keeping the link be-
tween EdStat and sci.stat.edu:

- EdStat introduces the string "[edstat]" at the beginning of the
  subject line of each message.  This can interfere with the
  proper "threading" of messages in some newsreader software.
  However, this isn't a problem because this feature can almost
  certainly be turned off by the new listowner if enough readers
  request it.

- If, as occasionally happens, EdStat fails to send a message
  that was submitted to sci.stat.edu, some writers (e.g., me)
  send the message a second time to ensure that EdStat subscrib-
  ers receive the message.  This results in a duplicate message
  in the newsgroup.  (This might also happen in reverse.)  How-
  ever, this isn't a serious problem because these failures are
  rare and it is very easy to ignore duplicate messages.

- One writer claimed that newsgroups are "dead".  This is incon-
  sistent with the fact that (assuming Jackie's sample is repre-
  sentative) the majority of valid messages sent by EdStat COME
  from the sci.stat.edu newsgroup.

- One writer suggested that spam sent to EdStat comes from the
  newsgroup.  Although I don't have statistics about this, my own
  experience reading sci.stat.edu is that very little spam occurs
  in the newsgroup.  (However, the amount of spam depends some-
  what on the Internet Service Provider [ISP] one is using.  Some
  ISP's may do a better job of filtering newsgroup spam than oth-
  ers.)


- Some person or organization must be found to maintain the link.
  However, many people and organizations are interested in stat-
  istical education, so finding a maintainer might be easy.

- From indirect comments, I suspect that some readers see other
  disadvantages of linking EdStat and sci.stat.edu.  I hope these
  readers will state these disadvantages so that we can see if
  it's possible to correct them.


.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to