LARA

Background presentations will take place in week 14. We strongly recommend that you pre-record your presentation. You should upload your talk on SwitchTube (the precise channel will be linked here soon). However, if you prefer, you can also live stream your presentation, but in that case you are responsible if the presentation does not reach your audience due to network quality issues.

The presentation should be 10 minutes long.

Q&A session of 5-10 minutes will follow right after the presentation. Please make sure at least one of you is available for the entire 20 minute slot.

We would like each member of the group to be part of the presentation.

Shortly after, you will receive feedback from us regarding the content of your presentation, as well as some general feedback on the form.

Your presentation should summarize your project. In particular, we'd expect to see

• a basic overview of the features you added to the compiler/language
• some (short) programs highlighting the use of these features, with a description of how your extended compiler behaves on them
• possibly some theoretical background you had to learn about to implement the extension
• an overview of the changes you made to each compiler phase and/or which phases you added

Here are some useful resources on how to prepare and give talks:

• How To Speak by Patrick Winston
• How to give a great research talk by Simon Peyton Jones

Please do not use Viktor's videos as a model for the presentation, but instead incorporate as many points of the talk of Patrick Winston as you believe apply to your presentation. It is an amazing and entertaining talk, despite (or because) it is meta-circular: he does as he says. Note: breaking physical objects or referring to supernatural beings in your video is not required. Use your own judgement and strike a balance in being comfortable with what and how you are saying things and trying out these pieces of advice.

We suggest that the speaker's video shows up when the speaker starts to speak, so that the audience can relate and identify the speaker. Afterwards, the video can be turned off and should come back on for questions and answers. Optionally, a small video can stay on throughout the presentation. The main content of the presentation should be a window showing the material being presented, for example as a PDF to which you can point to and/or annotate it. If the hardware allows you, you can also use a tablet to simulate a blackboard presentation where you write down everything you present, or use a combination or simple slides and a strategy of what you will write on them.

Video upload: please upload your video to this channel (login with EPFL credentials)

For your information and not as a requirement, Viktor's lectures are prepared using this hardware and software setup on Ubuntu 20 OS:

• slides prepared using the `beamer` latex package
• slides annotated using `xournal` PDF annotator in full screen mode on display size 1920×1080
• recording using Zoom, with the following options:
  • screen sharing PDF annotator (`xournal`), without option to optimize for full-screen viewing
  • local recording, with option Optimize for 3rd party video editor
• wacom cintiq pro display as external monitor for annotating PDF's using pen
• video segments are cut and assembled using ffmpeg, which works very fast:
  • cut like this:

fmpeg -i zoom_0.mp4 -ss 00:00:00 -to 00:02:03.00 -c copy mysegment01.mp4

• concatenate like this:

ffmpeg -f concat -i segmentlist.txt -c copy mycombinedvideo.mp4

where segmentlist.txt is a file containing one line per each file to include:

file 'mysegment01.mp4'
file 'mysegment02.mp4'
file 'mysegment03.mp4'

Alternatively, you can also use `obs` open source software. For recording, under advanced options, you may wish to choose a 1 second key frame interval to make cutting the video with ffmpeg work well.