Hackathon 2017 Website

Notes and timekeeping by Sergey. Presentation by Lea.

Scope: contents of website, it's appearance, how it is developed.

High level:

• website should target both new users and advanced users (developers will find a link to github (where the developer docs will be))
• landing page should be illustrated directory and have:

• gallery + api cookbook
• quicklinks to getting started/ install
• show activity: current development, news, testimonials
• compact contacts

Details:

Section 1a: Shogun title + mini description with buttons for Get started/ Install/ Develop

Section 1b: Gallery

Section 2: Examples

Section 3: On-goings

Section 4: Contacts

Notes from Discussion What is good in scikit-learn website?

• ML details right at front page
• Testimonials. We should collect at least some.

What is good in current website?

• Examples are good.

Use-cases of the website:

• Just looking
• Using it
• Want to contribute (esp. GSoC)

We need:

• A (mind)map of supported machine learning algorithms
• Show it is active and developed
• Much more compact contact details (email, irc, stackoverflow, ...). "Contact Us" list of people should go away.
• Simple website?
• A gallery of examples
• Three-pane alignment of News / Users / ...
• Main buttons just below the "Shogun" label: download, github.
• Navigating stuff (where am I right now?)
• Have a "Get started" page and a link just in front of you (like tensorflow)
• Have a "Programmer's guide" (like tensorflow)
• Single page site with sections. The more you scroll the more information you get. Links lead to something deeper.
• Choose better/brighter color.

Would be nice to have:

• Show some prediction about user and their intention. Prediction should be trained using Shogun.