Interactive Software Archival

Over the years, computation has substantially increased our creativity and opportunity. Computers constantly improve along with our ability to create and collaborate toward new art, science, and furthering existing technology. However, there's also a danger in that software becomes difficult to run over time. This fragility leaves us with a tech field that is somehow simultaneously growing and yet narrowing: Word documents become difficult to open, video games on specialized hardware become difficult to play, a researcher's simulator of protein synthesis becomes difficult to find-- these represent a loss of perspective and culture, an inability to revisit the past, and a lack of scientific accountability. In response, librarians have created a collection of digital libraries that archive images, video, and even the most vibrant of GeoCities websites. In the end this is not enough. We have media that is interactive and real-time. We have tools that can create new content at any point. Scientists have a dire need for accountability in research as experiments become harder to replicate. How do we make archives and libraries that can handle the preservation of code while allowing us to -- 100 years from now -- build, run, and archive anything new we create with that code? Where are our museums of interactive art? To answer that, we will take a look at the libraries and archives of the hypothetical, ones that exist now, and those in development. Focusing on the motivations and technology behind them, we will look at what they could do better and how. You will leave with an appreciation for the concept of provenance and how librarians can teach us the importance of systems that can trace an object back to the code that generated it. Finally, a look at the state-of-the-art and how Pittsburgh is leading the effort with two such interactive archives in development centered around using various levels of virtualization to preserve interactive art, games, and scientific experimentation. Our discussion will culminate with a demonstration of these new interactive archive systems, how they work, and what the future looks like.

Transcript

1. Interactive Software Archival wilkie @ wilkieii wilkie@status.wilkie.io (cc) Transcript: blah/blah/etc
2. Me Computers Make Me So Happy Sometimes
3. Me Computers Make Me So Grumpy Sometimes
4. Systems + OSes Application File System Abstraction Drivers IPC Kernel Software Is Too Complex
5. Systems + Archives Application Figure It Out, Genius Reliable Storage Hardware The Ideal System Is an Archive
6. Archival
7. Artifacts Mona Lisa Rosetta Stone
8. Leaning Tower of Pisa 3.99° of PRESERVED TILT
9. Provenance • Even if you’ve never heard the word, we developers have created a culture around using provenance. • Software development is sometimes aptly referred to as a Garbage Fire. • Bundler, npm, pip, CPAN… etc… all solving the problem of the chaos of development through maintaining a record of where and when and who dependencies come from.
10. Digital Artifacts 0010 1011 0110 Which Lives Longer?
11. Digital Archives
12. HD-Rosetta Disks Rosetta Disk Documentary: scottoller.com/rosetta
13. Software As An Artifact
14. What is Software? < ? >
15. Ada Lovelace ...the Analytical Engine might act upon other things besides numbers… the engine might compose elaborate and scientific pieces of music of any degree of complexity or extent. -- "Notes" Section A, 1842
16. Ada Lovelace … scientific pieces of music of any degree of complexity or extent.
17. Code as Art
18. Software Piracy Becomes Art
19. Demoscene Second Reality by FUTURE CREW
20. Software as an Artifact Is software preserved if it cannot be executed?
21. The Internet Archive https://archive.org
23. نب نسحلا ،يلع وبأ مثيهلا نب نسحلا Abū Alī al- asan ibn ʿ Ḥ al- asan ibn al-Haytham Ḥ c. 965 – c. 1040 CE
24. Scientific Method Form a Question Propose Hypothesis Predict Outcome Test Analysis Record, Reflect, Repeat
25. Repeatablilty Running the same experiment
26. Reproducibility Running the same experiment but altering the conditions
27. Science + Software Is science reproducible if you can't run the code?
28. Sorry... Pessimistic CS Statistics Reproducibility Rate 32% Of Scientists Believe Slight/Significant Reproducibility Crisis 90%
29. Dang it Code Publicly Available 35% Asked and received code 22% Asked but told NO 36% Asked but no response 7%
30. That’s a Little Better Got it to Build: <30 minutes 58% Got it to Build: >30 minutes 28% Couldn't build it; told it should 10% Couldn't build it; told it sucks 4%
31. Excuses “ Student graduated; Nobody knows how it works ”
32. Science is an Interactive Narrative
34. Solution: Archival
35. Interactive Archival
36. Solution: Virtualization “Not a Graveyard” – Erika Linke and Dan Ryan introducing the concept of Olive and executable archives.
37. Olive Archive CMU (Pittsburgh!) QEMU+KVM Emulation On Top olivearchive.org
38. VmNetX: Internet Suspend/Resume Virtual Machine Network eXecution; GPLv2 VM Archive Your Computer Executes VM Stream VM (HTTP)
39. bwFLA University of Freiburg
40. bwFLA VM Archive Your Computer Sees Video Stream Stream Video (SPICE, etc)
41. http://bw-fla.uni-freiburg.de/demo-cloud.html
42. Olive/bwFLA: Drawbacks Inflexible What if VM tech is obsoleted?
43. Solution: Satisfaction Application Figure It Out, Genius Reliable Storage Hardware
44. Ex: Commander Keen Reliable Storage Hardware 1991
45. Native DOS -> Impractical Figure It Out, Genius Reliable Storage Hardware 2016
46. DOSBox Hardware
47. Decide How To Run Hardware
48. University of Pittsburgh https://occam.software
49. React To The Future x86-64 Javascript
50. Javascript Emulation Reliable Storage Browser
51. Adaptable (NES, SNES, etc) x86-64 Javascript XNES OR
52. Scientific Workflows
53. Active Curation Image Filter
54. Lessons Think about how the heck your code works in the future
55. AND HAVE FUN! http://wilkie.io /abstractions