[SciPy-User] Main Site Design Questions

[Ralf Gommers]

On Mon, May 20, 2019 at 4:13 AM Christina Lee <chrissie.c.l at gmail.com>
wrote:

> As both people are realizing scipy.org needs a design overhaul and
> discussing doing so, I'd like to start a discussion clarifying aspects of
> SciPy that would drive design choices.
>
> These might seem like annoying and weird questions to someone used to a
> different type of thinking, but this is still problem-solving. The code
> itself gets more useful if there is a lower barrier to entry.
>

Hi Christina, thanks for the questions - they're not weird and annoying (at
least to me), I can see how answering them well can lead to a better design
and information layout. You already triggered me to think about changing
the scope of the site (last question).


> * How should the typical user* feel* about SciPy?  What gut-level
> impression should the website leave?  How do you* feel* about SciPy?
>          For example: reliable, dependable, cutting-edge, fast, novel,
> exciting, ... These will influence aspects like colors, fonts, images,
> spacings, etc.
>

Reliably, dependable, fast: yes. Exciting: for parts of SciPy yes, I hope.
Cutting-edge, novel: no. When new features are proposed for inclusion in
SciPy, one of the things we look at is whether the feature will be widely
applicable (>1 science domain) and that it's "proven" to some extent (by
use, citations, etc.). If it's cutting-edge research, we encourage creating
a new package for it or finding some other more suitable package.


>
> * What is the most important thing for a user to do?
>        Fairly certain the answer to this question is "Install".
>

Install, although that's a one-time (or few-times) action. Then use. Unlike
NumPy, there's not a couple of key concepts to grasp. What would be good is
for the user to understand what is in SciPy and how the package is laid out.


> * What is the most important thing for someone to learn about SciPy?
>       Assume they know absolutely nothing and you have 10 seconds to catch
> their interest.
>

That's always hard:) I'd say that it's a "standard library" of high-quality
fundamental numerical algorithms - whether they need statistics, linear
algebra, special functions or optimizers, SciPy is the first thing they
should look at.


> * You have more than 10 seconds: Anything really cool about SciPy?  Done
> anything amazing with it? Any cool plots?
>

The recent black hole image and the discovery of gravitational waves would
be high on my list. Would be cool to get some high-impact examples from a
broader range of fields, like biology, economics, engineering, etc.


>
> * Do we have information about both the typical SciPy user, and the
> typical person visiting the website? If most people have barely touched
> Python the first time the come across SciPy.org, that changes presentation
> and wording.
>

I think slightly more advanced than "barely touched Python". NumPy, Jupyter
and Matplotlib are probably things beginners would look at before SciPy.
There's indeed still mostly scientists and engineers, and students in those
fields. People who do not consider themselves programmers, and are just
needing to get their task or experiment done.

Regarding the website, there's of course the complication that scipy.org is
about the "SciPy ecosystem" and scipy.org/scipylib about "SciPy the
library". And then there's SciPy the conference as well.

I actually think that today it would be an improvement to make the focus
narrower. Disambiguate the name, point people to other places for the
conference and ecosystem, and then focus on the library.

Cheers,
Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-user/attachments/20190520/1a8670f8/attachment-0001.html>