1 of 52

World of Code hackathon 2023

R

2 of 52

Before we get started…

https://bit.ly/woc2023github

Name badge / Discord name
Discord role: #select_role
Wifi access (eduroam / UT-open)
Zoom access
WoC access
WoC Github (https://github.com/woc-hack)

R

3 of 52

Localities and links

Zoom: https://bit.ly/woc2023zoom

Github: https://bit.ly/woc2023github

Discord: https://bit.ly/woc2023discord

Main room: MK 435

Breakout rooms: MK 434, MK 354, MK 235 (for working time)

R

4 of 52

https://conduct.tennessee.edu/

R

5 of 52

Team presentations will be recorded

R

6 of 52

#knoxville2023-hackathon channel in Discord

Github Issue tracker

or ask one of the WoC experts online or in the room

R

7 of 52

05:00pm: Opening Ceremony, Schedule, Logistics, Getting to know each other

05:45pm: Idea pitches

06:15pm: Idea voting

06:45pm: Team formation

08:30pm: Dinner @ Carrabba's Italian Grill

(leave at 08:00pm)

Schedule Friday (November 17)

(all times are in EST)

R

8 of 52

Schedule Saturday (November 18)

(all times are in EST)

09:00am: Breakfast

09:30am: Roll-call

09:45am: Team formation finalization

10:30am: Team presentations

11:00am: Working time

12:30pm: Lunch

01:30pm: Checkpoint 1

02:00pm: Working time

04:00pm: Break

04:30pm: Working time

06:30pm: Checkpoint 2

3:30pm EST

Dinner on your own

R

9 of 52

09:00am: Breakfast

09:30am: Roll-call

09:45am: Working time

11:30am: Break

12:00pm: Final presentations and closing

01:00pm: Lunch

Schedule Sunday (November 19)

(all times are in EST)

R

10 of 52

Project ideas

30 second pitch
I am…
I would like to…
Because…
And for that I need…

R

11 of 52

Examples of submitted ideas:

Using World of Code to identify scientific software repositories to assemble a corpus with the help of AI/ML LLM technology
Analyze the "halflife" of code for languages over time
Correlate README features to the number of GitHub stars

(https://github.com/woc-hack/hackathon-knoxville-2023/blob/main/ideas.md)

Examples of prior projects:

BotBusters - Detecting and Characterizing Bots that Commit Code
Inflection points - Studying changepoints in open source software evolution
TrackHack - Tracing hackathon code
KJD - Studying the vulnerability duration across projects

(https://github.com/woc-hack/hackathon-knoxville-2023/blob/main/prior_projects.md)

R

12 of 52

Idea voting

https://bit.ly/woc2023vote

R

13 of 52

Team formation

Most voted ideas get to form a team
Idea proposer find a space and join their breakout room in Zoom
Remaining participants choose which project they want to join
Teams need to have between 3 and 5 members from different institutions and at least one on-site participant
Pay attention to potential time zone issues!
Having an experienced WoC user in your team is encouraged

R

14 of 52

Team presentations

Tomorrow at 10:30am in MK 435 and main Zoom
Github repo in https://github.com/woc-hack
Short presentation (3 minutes per team):

Team name and members
Concrete project idea
How are you planning to approach your project?
What can you achieve during the hackathon?
What if anything do you think you might need help with?

Decide for presenter!

R

15 of 52

Deliverables

Presentation at the end of the hackathon
Github repo in https://github.com/woc-hack
Project report:

Introduction
Methodology
Preliminary Findings
Challenges
Future work

Decide for presenter!

(template: https://github.com/woc-hack/hackathon-knoxville-2023/blob/main/project_template.md)

R

16 of 52

It is ok to help each other!

R

17 of 52

Enjoy the ride!

Share your experience: #worldofcode2023

R

18 of 52

Research on hybrid hackathons

https://hackathon-planning-kit.org/

Abasi-Amefon Obot Affia (amefon.affia@ut.ee)

Organizer observation and interviews
Participant surveys and interviews

R

19 of 52

Getting to know each other…

For experienced WoC users:

I can help you with…

Name, Affiliation

Why are you participating?

R

20 of 52

Project pitches

30 seconds
I am…
I would like to…
Because…
And for that I need…

R

21 of 52

Voting and break

https://bit.ly/woc2023vote

R

22 of 52

Team formation

Most voted ideas get to form a team
Idea proposer find a space and join their breakout room in Zoom
Remaining participants choose which project they want to join
Teams need to have between 3 and 5 members from different institutions and at least one on-site participant
Pay attention to potential time zone issues!
Having an experienced WoC user in your team is encouraged

R

23 of 52

Housekeeping

Breakfast at 09:00am (MK 435)
Roll-call at 09:30am (MK 435 and main Zoom)
Team formation finalization 09:45am

Make sure that you are in a team!
Exchange contact information before you log off today

To get into the building tomorrow morning:

text Mahmoud or Luis on Discord

R

24 of 52

Team presentations

Tomorrow at 10:30am in MK 435 and main Zoom
Github repo in https://github.com/woc-hack
Short presentation (3 minutes per team):

Team name and members
Concrete project idea
How are you planning to approach your project?
What can you achieve during the hackathon?
What if anything do you think you might need help with?

Decide for presenter!

R

25 of 52

Dinner @ Carrabba's Italian Grill at 08:30pm

Departure: 08:00pm from MK 435

R

26 of 52

Day 2

R

27 of 52

Schedule Saturday (November 18)

(all times are in EST)

09:00am: Breakfast

09:30am: Roll-call

09:45am: Team formation finalization

10:30am: Team presentations

11:00am: Working time

12:30pm: Lunch

01:30pm: Checkpoint 1

02:00pm: Working time

04:00pm: Break

04:30pm: Working time

06:30pm: Checkpoint 2

3:30pm EST

Dinner on your own

R

28 of 52

Team presentations

At 10:30am in MK 435 and main Zoom
Github repo in https://github.com/woc-hack with team name and members
Short presentation (3 minutes per team):

Team name and members
Concrete project idea
How are you planning to approach your project?
What can you achieve during the hackathon?
What if anything do you think you might need help with?

Decide for presenter!

R

29 of 52

Physical breakout rooms:

MK 434, MK 354, MK 235

R

30 of 52

Team presentations

(10:30am in MK 435 and main Zoom)

R

31 of 52

Schedule Saturday (November 18)

(all times are in EST)

09:00am: Breakfast

09:30am: Roll-call

09:45am: Team formation finalization

10:30am: Team presentations

11:00am: Working time

12:30pm: Lunch

01:30pm: Working time

03:30pm: Break

04:00pm: Checkpoint

04:30pm: Working time

3:30pm EST

Dinner on your own

R

32 of 52

Checkpoint

At 04:00pm in MK 435 and main Zoom
Short presentation (3 minutes per team):

What have you achieved so far?
What were your biggest issues?
What are your plans until the

final presentation?

What do you need help with?
Slides are ok but not required.
Decide for presenter!

R

33 of 52

12:30pm

Lunch in MK 435

R

34 of 52

03:30pm

Break in MK 435

R

35 of 52

Checkpoint

(04:00pm in MK 435 and main Zoom)

R

36 of 52

Checkpoint

Short presentation (3 minutes per team):

What have you achieved so far?
What were your biggest issues?
What are your plans until the

final presentation?

What do you need help with?
Slides are ok but not required.
Decide for presenter!

R

37 of 52

Final presentations

Tomorrow at 12:00pm in MK 435 and main Zoom
Upload your code to your Github repo
Document your project

(Template: https://github.com/woc-hack/hackathon-knoxville-2023/blob/main/project_template.md)

Presentation (max. 5 minutes per team):

Project description
Methodology
Findings
Challenges
Future plans

R

38 of 52

Housekeeping

Breakfast at 09:00am (MK 435)
Roll-call at 09:30am (MK 435 and main Zoom)
To get into the building tomorrow morning:

text Mahmoud or Luis on Discord

Dinner on your own
Flexible end of day

R

39 of 52

Day 3

R

40 of 52

09:00am: Breakfast

09:30am: Roll-call

09:45am: Working time

11:30am: Break

12:00pm: Final presentations and closing

01:00pm: Lunch

Schedule Sunday (November 19)

(all times are in EST)

R

41 of 52

Final presentations

At 12:00pm in MK 435 and main Zoom
Upload your code to your Github repo
Document your project

(Template: https://github.com/woc-hack/hackathon-knoxville-2023/blob/main/project_template.md)

Presentation (max. 5 minutes per team):

Project description
Methodology
Findings
Challenges
Future plans

R

42 of 52

11:30am

Break in MK 435

R

43 of 52

Final presentations

(12:00pm in MK 435 and main Zoom)

R

44 of 52

Team presentations will be recorded

R

45 of 52

Project To License Matching

World of Code 2023

46 of 52

Sampling The Data

A mapping from project license blobs to known official license blobs had already been gathered.
We validated the mappings by reviewing the project licenses and the matched official license by hand.

First we sampled 20 project licenses at thresholds of 0%-20%, 20%-40%, 40%-60%, 60%-80%, and 80%-100% matches of hashed windows against the official license blob.
We found that a threshold of 80% worked well, so we reviewed 20 more samples each of smaller buckets: 80%-85%, 85%-90%, 90%-95% , and 95%-100%.

A manual review using a threshold of >=95% matches exclusively had only name and year changes against the license template.
Using a threshold of >=80% matches, we can reasonably assume that the project is using the official license, but there may be additional formatting changes in the document.
The 80% threshold still includes 82% of the matched licenses; we still can include most of the data with this method.

47 of 52

Sampling Analysis Results

Projects with	>100 signatures
80%-85% Match	85%-90% Match	90%-95% Match	95%-100% Match
9/20 had additional licenses but no modifications, 10 were unchanged, 1 had added clauses.	2/20 had additional licenses but no modifications, 10 were unchanged, 1 had added clauses.	2/20 had additional licenses but no modifications, 13 were unchanged, 5 had added standard exceptions (SPDX)	All 20 were unchanged.

Projects with	<=100 signatures
80%-85% Match	85%-90% Match	90%-95% Match	95%-100% Match
3/20 had additional licenses but no modifications, 17/20 were unchanged.	3/20 had additional licenses but no modifications, 17/20 were unchanged.	All 20 were unchanged.	All 20 were unchanged.

48 of 52

What We Learned From Sampling Project Licenses

License blobs may include multiple licenses, so matching against one official license may not be the best strategy.
Oftentimes licenses contain a disclaimer by the author that the software may be subject to other licenses (which does not change anything for our purposes).
Some licenses point to other files which contain third party licenses (which also does not change anything for our purposes).
Some licenses, like the Common Attribution license, are more likely to be modified with additional clauses than others.

49 of 52

Finding Conflicts In Copied Code

Using this data, we are able to find potential license conflicts by checking the licenses of downstream and upstream copied code.
We created a Jupyter Notebook to perform this search for a sample of 10,000 projects.

We created a mapping from downstream projects and their licenses to a list of upstream projects labeled with their licenses as well.
Using this mapping we found 9 instances of the 10,000 where an upstream project is determined to have a copyleft license (like GPL), but is re-licensed an incompatible non-copyleft license.
This was just performed in the first 10,000 projects sorted in alphabetical order; there are likely to be many more instances of these conflicts in a larger sample.

50 of 52

Conclusion

What did we find?

Matching licenses with a threshold of >=80% matched windows is an effective way to predict the license for a project.
Using this method, we can map the licenses of projects and the projects they’ve copied code from.
We can find a statistically significant number of projects that potentially copy copyleft-licensed code, but are re-licensed with an incompatible non-copyleft license in the downstream project.

51 of 52

Post-event survey and interviews

https://hackathon-planning-kit.org/

Abasi-Amefon Obot Affia (amefon.affia@ut.ee)

Post-event survey: https://bit.ly/woc2023survey
We will contact you to schedule a 30-minute post-interview

R

52 of 52

Thank you for participating!

World of Code hackathon 2023

R