Also invisible from hasgeek.com

Should maybe have:

• Code of Conduct
• Access maps to common venues
• latest newsletter
• newsletter signup
• latest blog posts
• featured blog posts (childcare, for example)

Reported here: https://twitter.com/anandology/status/887978805507346432

@anandology: the time scale on the schedule page is very confusing. 10 mins is 3x longer than 35 mins.

Related: hasgeek/android#3

Possible solution

Calculate the greatest common factor between all the session and break durations, use that as the time for each rowspan, then assign every session item a multiple of the rowspan.

Best case: you have all the sessions and break equally spaced out (so GCM is the same as the duration) so it'll become a simple table with no item needing a rowspan of more than 1.

Worst case: you have 11 hours (9am to 7pm) all 5 minutes apart - which is Funnel admin's granularity. Which gives you a GCM of 5, so 132 rows in the table and items with varying rowspans.

v0.4.3 of Bulma released on July 5th, 2017.

It adds support for a dropdown nav, like bootstrap, as well as an improved breadcrumbs component.

Once the day of an event begins, the event is not available on the website.

Move tests from hasgeek/events:tests/after_build

Currently, our event websites and funnel use a datelocation field to indicate the date and location of an event. This is a freeform text field that's shown as-is next to the event information

The reason it is a freeform field, and not constructed from the date and venue data is so it's flexible for cases where the event might not be so straightforward.

This, however, is a double-edged sword as the lack of a format has made it wildly inconsistent based on who has entered the data, like on talkfunnel.com

On hasgeek.com, however, the datelocation field in the UI is generated from the start_time, end_time, and city information, with little tweaks like ignoring the time component when the event is more than a single day. (Relevant code)

I favor the hasgeek.com approach, and to generate the field from fixed data. If there any exceptions, it'll force us to model a solution to that problem, like multi-day events (For example, every Sunday for 4 weeks) - which we haven't really discussed yet.

This is a better way to cross-sell/promote events together automatically.

Currently, HasGeek does a wide variety of events - Conference, Workshops, Meetups, etc.

We need to flag every event we do as a specific type, this will help keep events of the same type consistent while still allow each type to have its own special requirements.

Workshops use instructor while conferences generally use speaker, and meetups may use register or rsvp instead of tickets, etc.

We need an exhaustive list of event types.

Example of issue:

https://anthillinside.in/2017-nlp-workshop/
Steps:

• Clicking on one of the cards under instructors brings up more information
• Clicking anywhere outside the white card dismisses the card
• Clicking on the 'X' button at the top right corner does nothing

It'll be useful to programmatically update bio, profile and cover photos, url, etc - based on what we consider the most important relevant event.

Tweepy should be simple enough for this use case.

The update_profile(), update_profile_image(), and update_profile_background_image() APIs are what we'll need to use.

http://amitkaps.com/stories/

The current implementation of the service worker uses sw-precache to cache resources.

This approach caches all the pages assets and updates the cache in the background.
This leads to a problem when you first visit the site after an update. You get the old version of the assets and need to reload to get the latest version. This is less than ideal on something like event websites which tries to be a single page site, with no actual page reloading happening until you visit the site again.

Possible fix

Notify the user that the site's content needs to be updated (and prompt for a reload). It's also the strategy suggested by Google

Using separate files saves you from merge conflicts if two people file PRs at the same time.

Run tests after a successful Jekyll build to validate the API

Since hasgeek/boxoffice#167 has been merged (❤️ @shreyas-satish) we can specify the items_ids we want to render in the widget, and how to order them.

A side effect of this is that we can use boxoffice's /order/kharcha API to get the current price for a ticket and render it.

This will require us flagging a specific item id as the primary one, which is usually the most common ticket people may purchase.

We have numerous events coming up, but they are everywhere - newsletter, Twitter, Google calendar, Trello - except the website. Need to find a way to feed all the output sources from a single input.

A schedule section that is built using funnel's schedule API.

Reference: ffi/ffi#608

Cerberus, the validation library, has a new release 1.2 which introduces a bug around custom rules - pyeve/cerberus#379

The current fix is to lock it to version 1.1 - 45fbdb7

Once this issue been patched, we should test and update back up to the latest version.

It's good to provide value to visitors (potential attendees) without requiring them to buy a ticket. For instance, https://fifthelephant.in/2018/ does not provide any value to a visitor save for informing them about the conference agenda.

An easy way to start doing this is by embedding the YouTube playlist for the previous editions. It is also be useful to measure the increase in website engagement and conversions, this change produces.

This will also be useful internally at hasgeek.

Similar to sponsors, venue data should be centralized and referenced by a unique name, while still allowing per-event venue information if needed.

Add a new front matter value called instructors/speakers (both should be allowed) section to pages. Useful for workshop/talk pages.

name,image_url, bio fields are compulsory.

layout: workshop
instructors:
- name: Rahul Mathur
  image_url: https://images.hasgeek.com/embed/.....
  bio: |
  Engineer @FreeCharge
  Windows and iOS dev
  Alpacas are evil
  url: https://twitter.com/weemundo

Hypothesis: Seeing what jobs in the domain pay, will help justify paying the cost of the ticket.

For people who are new to a domain, how do they justify attending a conference that costs several thousands of rupees? This is especially a problem where the attendee's ticket is not reimbursed by the company.

I propose we import relevant job posts from Hasjob with the pay graph to help provide some context around the domain.

Test: Does showing jobs result in better conversions?

Right now, there are several CfPs open, for example. But there's no visibility on the site about that.

Events can target specific interest groups, such as data science, devops, frontend engineering, etc.

These can be fixed list of tags (in a tags.yml) you can attach to an event, and limit to 2 or 3.

These can then be used for sorting/filtering.

Travis is moving the default linux image for builds from Precise to Trusty

This seems to have broken a dependency in the build.

The new image seems to be missing the gulp npm package.

Possible solution

Create a new branch and install gulp manually in before_script on .travis.yml, and then solve any subsequent errors one by one.

Sites should allow favicons in the config and default to the HasGeek one if none are specified.