[meta-sig] Switchover to mailman

Greg Stein (Exchange) gstein@exchange.microsoft.com
Fri, 27 Feb 1998 03:46:52 -0800

Can we concentrate on *usability* ? If somebody wants your address, then
they'll get it simply by trolling the archives. Let's just face that
fact and make the site as helpful and as useful as possible.

For example, that autogen GIF idea? Bleh. As a user, how do I copy an
email address from my browser into my mail client? Can't. Gotta retype
it. Forget that... just about the same as not having it in the first
place, IMO.

Also, why just list it in plain text? Let's just go ahead and create
mailto: links. That will be very handy.

Please realize that a spider is not going to stop at a CGI script.
That's looks like just any other page. They'll traverse it just as
quickly as a .html page. And yes... the magnet site was crawled. Just
search for your name in altavista and you'll see pages on magnet.

I would recommend that we allow the per-person privacy setting. Most
would be on by default and the list would have an emails page by
default. There have been a number of times when I'd like to see the list
of people and I want to simply hit a web page and find it.

Let's make the system easy for *us*.

-----Original Message-----
From:	Ken Manheimer [SMTP:klm@cnri.reston.va.us]
Sent:	Thursday, February 26, 1998 11:22 AM
To:	Andrew Kuchling
Cc:	meta-sig@python.org
Subject:	Re: [meta-sig] Switchover to mailman

On Thu, 26 Feb 1998, Andrew Kuchling wrote:

> Guido van Rossum writes:
> >I suppose the SIG page would eventually have a link to that page (in
> >fact I wonder why it doesn't yet) which would be a dead giveaway.
> 	This raises a new question for me; I'm working on resurrecting
> the Web archives of the SIG lists, and currently e-mail addresses are
> turned into mailto: URLs, which could easily get crawled.  Should this
> policy be changed?
> 	To some degree the damage is already done; the archives were
> live at Magnet for a while, so they may already have been crawled by
> spiders looking for e-mail addresses.  But it's still worth fixing
> this if people think it necessary.  We'd have to write a mailto CGI,
> and put some simple database underneath it.

It seems to me that this level of indirection would do the trick, and it
seems like a good idea.  I have some reluctance, though, about adding
yet another system to the installation.  I'm currently thinking it's a
good exercise to keep a duplicate copy of a public system going, not
just for the sake of an offline place to experiment with changes, but
also as a measure of how hard it is to manage the system in the first
place - and i'm noticing that the number of separate but interdependent
sub components makes an enourmous difference in the difficulty of
replicating the system.  I guess there are *some* benefits to either
monolithic systems or independent sub components...


Meta-sig - SIG on python.org SIGs and mailing lists