[vlc-devel] Future of the update mechanism

[jpd at videolan.org]

On Thu, Jul 30, 2009 at 04:30:30PM +0200, Georg Seifert wrote:
> I do not completely get the excitement about wether use Sparkle or not.

I think I've explicitly said before that I'm not interested in bashing
any single product, and conversely I am not at all interested in hearing
how great any single product is.

But let me spell it out once more: I want to hear about alternatives.

I don't want a ``use this or not'' discussion. I want a ``this is the
problem, these are the requirements, these are possible solutions''
discussion.

Without something to compare it to we cannot evaluate any single
system properly. Without disagreement we cannot discern where the
differences are, nevermind resolving them, if necessairy.

What you're doing is singing the gospel so as to drown out the dissenters.
It's an old technique and works often enough. ``Look! Shiny!''

I'm not interested in ``Shiny!''.

Because it does not get us a better solution. It merely gets us stuck
with the latest hype (``over 500 software projects cannot be wrong!
jump!''), until the next hype comes along. In the meantime, it looks
like the people that managed to mount an actual argument are leaning
against the idea of automated updating entirely.

So, where are the alternatives? How do we know that this particular
thing is the best currently available, that it meets our needs and that
we want to solve our problem that way? If so, we can move on to deciding
whether its drawbacks are worth it. But have we defined our problem
already? We started with a solution looking for one of those.

So far, I'm inclined to think not so. Complete failure to discuss even
the `minimalist' and `none' options are key in that thinking, as is lack
of other options and this thing's inherent limitations.

I don't want another salespitch; I want a discussion.