Re: lynx-dev Re: LinuxApps: Lynx

[Klaus Weide]

On Sun, 15 Aug 1999, David Woolley wrote:

> >    typical linux machines that would use those packages have a compiler
> >       installed anyway and users can compile from source without problems
> >       and with a better chance to configure it as they want;
> 
> Unfortunately, the typical Linux user these days is only marginally
> more technically competent than the typical Windows user (although the
> latter still argue that Linux is too technical).  They do not know and
> do not want to know how to build from source.
> 
> Unfortunately, they have also been taught, by Microsoft, to want the 
> latest beta quality software, and don't realise that there is also
> alpha quality software out there in the freeware arena.
> 
> A significant element of the Linux community itself, particularly those
> with anti-Microsoft leanings (and companies that sell Linux commercially),
> are positively encouraging Linux to move into this market.

Nice rant.  But I think you are exaggerating, especially when you
claim that "the typical Linux user ... do[es] not _want to know_ how
to build from source." (emphasis added)

> Moreover, there seem to be quite a few people out there who want to make
> a name for themselves by being the first to bring out an RPM, so there is
> competition to create the first RPM of each alpha version (from what I've
> heard, Debian is rather more disciplined in this respect, than Red Hat).
> 
> Although I haven't visited freshmeat; my impression is that it is under
> a commercial imperative to get as many hits as possible, which means that
> it has to be driven by the market demand for the very latest versions.
> Unfortunately, the GPL means you can't embargo such distribution.

Speculation about freshmeat's motivation aside, yes the GPL means we
cannot pick and choose who is allowed to distribute the program.  I
would not call that unfortunate.

> Therefore, although what you say is true in an ideal world, it is no
> longer a realistic expectation in the real world.

I say it is still a realistic expectation for those Linux users that
come to lynx-dev or to sol.slcc.edu for advice in getting lynx.  The
others - what you would call the "typical" Linux users - get lynx as
part of a prepackaged distribution, and then perhaps upgrade it by
using a binary package found somewhere.  They already do that, whether
"we" (lynx-dev hackers or web page maintainers) like it or not, and
don't seem to have any difficulty "finding" lynx.  So they don't need
help from lynx-dev, or from some page linked from <http://lynx.browser.org>,
in order to locate a lynx binary package.

Look here for a bewildering choice of lynx binaries in RPM format
alone:

  <http://rufus.w3.org/linux/RPM/LByName.html>
  (Warning, LONG - it is a list of all packages starting with 'L')

Any attempt from the lynx-dev side to point to one or several of
those, to the exclusion of others, would amount to making a
recommendation.  I don't see that anybody is volunteering to review
all those binary packages and recommend one or more as the "upstream
authors' choice", and keep that recommendation constantly updated.  So
let the anarchy of distribution take care of itself.

There's no obligation (of course) for lynx-dev web page maintainers to
point to binaries for Linux, and I argue it wouldn't even be helpful.
Things might be different if lynx packages were hard to find, or if
lynx were not such a common program included in nearly all Linux
distributions (even small ones meant to fit on one or a few floppies).

Linux lynx users who chose to get lynx through some middleman to get
lynx, and then come to lynx-dev with some problem, can be advised to
either go back and talk to their middleman or compile from source, if
nobody here happens to know a solution.  It shouldn't be lynx-dev page
maintainers' responsibility to tell those folks where to find the "best"
lynx binary package, that way lies madness.

   Klaus