Re: lynx-dev anonymous usage of Lynx (was Arizona)

[brian j. pardy]

Philip Webb wrote:
> 981009 Brian Pardy wrote: 
> > I think asking for a dedicated machine is often too much to ask for when
> > dealing with the sort of sites that are allowing anonymous access to users
> > for whatever reason. Said sites are probably (I'm assuming, having no real
> > experience) operated on little or no budget, and anything we can do to
> > make Lynx A) more secure in general, or B) make it a more usable program
> > for people providing this kind of service is a Good Thing to me.
> 
> hardware's quite cheap these days, lots of it second-hand.
> of course, if someone gives them free programming support
> they won't see any need to improve their hardware.

Free programming support is what volunteering is, isn't it?

Anyone can ask for a feature/bugfix, it'll either get done or not. If it
would take up too much developer time and there's no one willing to devote
that much time, it won't get done. I really don't think it's valid to say
that time spent fixing the problems of one group of users (anon users)
versus time spent fixing another group's problems (single-user installs)
is somehow less valid.

Part of the reason I run Lynx (and Linux, and ...) is precisely because I
don't need to plonk down money for more powerful hardware or rely on
donations. I think that's one of our strengths.

> > People running sites allowing anonymous access to the net are a wonderful
> > thing, and I'd hate to see them be left out in the cold[1]. 
> > To me, Lynx is a way that I can attempt to give back to the community that
> > has provided so many things for me (the net, so much great free software,
> > friendships). I think the people that help others get on the net at their
> > own personal expense (time/money/etc) are doing quite possibly one of the
> > best things the net can have done for it. And anything I can do to help
> > these people out makes me feel good.
> 
> i certainly can't quarrel with your fine words
> & you undoubtedly put your own time & skills into the effort,
> but you may be overestimating how genuinely deserving
> typical users of freenets are (eg someone today mentions winter vacationers).
> it may have been true when Lynx started, but i doubt if many N Americans today
> have to RELY on freenets for access to the Internet.
> if all anonymous Lynx is really doing is saving residents taxes
> they would otherwise have to pay for better library service,
> i can't see any point in it.

As long as someone sees a point, they'll code. 

I also don't think saying there aren't "many" people where such is their
only access is that valid either. 

> > Philip Webb wrote:
> >> Lynx started out on a campus freenet, but has long outgrown those origins
> >> & today tends to be installed -- a/a used -- mainly by individuals
> >> running Linux, Windows or DOS or shell-account users on UNIX (like me).
> >> there appears to be no-one currently developing Lynx
> >> who also runs an anonymous site & so is motivated to be security guru,
> >> so i'ld say we have to ask managers of sites with anonymous users
> >> to contribute their own time & skills to the effort insofar as they can:
> >> otherwise, they should not expect security vs anonymous miscreants
> >> to have any kind of priority in future Lynx development efforts.
> > That seems a curious thought given how much coaxing was required
> > to get you to attempt to debug the rcfile problem you were having.
> > I don't think any bug is an acceptable bug.
> 
> you shouldn't be directing these words at ME!!
> I (emphatic) had to persist in reporting the bug
> in face of retorts that "we can't see it here" etc!
> i've done everything to track it down very promptly,

Everyone accepted that the bug existed, but given an inability to
reproduce it, you just can't debug it.

(I remember about a year ago I couldn't compile ncurses using glibc. I
asked Tom about it, he said "it works just fine for me". Nobody else,
anywhere was having any problems. I had to figure out that I had some
stupid header file sitting around from somewhere that was mucking my
glibc compile, and anything that compiled with it. It was the weirdest 
thing I had ever seen. execve() didn't work.)

> but the advice i have been given has been  c 75 %  correct on average.

I didn't see any blatantly wrong advice, and it's rather difficult to get
information that is totally correct to someone on a Unix that you don't
have access to.

> in the end, the only solution to date is my own crude hack at the code.

Give it time. That's why it's a pre-release version.

-- 
E's Corollary to Schachter's Hypothesis: 
"...and porn.  And a list of ska bands."