Thu Jun 2 21:58:48 CDT 2011
On Thu, Jun 2, 2011 at 2:57 AM, Thomas Kluyver <firstname.lastname@example.org> wrote:
> I think in fact we're saying the same thing, but from two different angles.
> I'm not suggesting we write reams of content in straight HTML. Sphinx/RST is
> the right tool for the bulk of the stuff, augmented with a wiki for the more
> changeable stuff. But, as you say, a couple of key pages - certainly the
> landing page, and perhaps also the download page - need to be visually
> I think this is going to be easier to achieve by designing them in HTML than
> by trying to hammer Sphinx's output into the form we want. Sphinx can be
> themed, but the examples I've seen all still fit within a certain Sphinx
> model, which doesn't feel ideal for a homepage.
We're saying mostly the same thing, though I'm still saying that
sphinx *can* be the right tool even for what the part you have in
- sphinx makes it very easy to have specific pages in pure html (ideal
for heavily manually designed pages)
- while your statement above is true of most sphinx sites out there, I
don't think it's a necessary outcome, just the fact that most people
seem to do only minimal tweaking of the theme templates. But even the
web ignoramus that I am was able to change my site (http://fperez.org)
to look somewhat different. You can still recognize the right sidebar
easily, but I think it doesn't feel quite as doc-like as others, and
this was literally just a tiny bit of changing the default theme
template, with *zero* hand-coded html.
So I still think that a small amount of fiddling with shphinx, with
the deliberate intent of making a theme that is deliberately designed
*not* for documentation. I think mine is a decent starting point,
since I already nuked many of the doc-driven elements out. It just
needs a litlte more visual flair.
> I'll try building Komal's repository a bit later on.
> In terms of the landing page, have you had a look at the file I sent above?
> I'm by no means a designer, but that's how I envisage the basic shape of it
> - a little content, with the focus on a set of key links to downloads, docs,
> mailing lists, etc.
I just did now, and I think that's on the right track.
Back to Brian's comment and the question of resources, while we hope
to have later this year some extra resources for this kind of thing, I
honestly think that doing something now is not a bad idea at all, as
long as it doesn't become a major project for anyone. It's not
*certain* we'll have spare resources for this, and our current
situation is so sad (an old moin wiki and a mishmash of github stuff)
that an improvement would be very welcome, in my view. As long as
it's not a huge sink of effort, I don't see any harm in improving this
direction now, if you (or anyone else) has the energy for it. That's
why I proposed it as one of the India projects, which Komal picked up.
More information about the IPython-dev