"Custom software is significantly more problematic to open than generic, reusable components. Often integrally tied to an organization’s way of working, custom software tends to be both hard to use in other environments and also hard to open without exposing some kind of sensitive information."
Panu Kalliokoski,
Senior Software Developer, Solita
International Technology Conference Mindtrek 2017
20th - 21st of September, 2017
Tampere, Finland
Axa Assurance Maroc - Insurer Innovation Award 2024
Panu Kalliokoski, Solita “Why Custom Software Should Be Open Source” - Mindtrek 2017
1. WHY CUSTOM
SOFTWARE SHOULD
BE OPEN SOURCE
Panu Kalliokoski, senior software developer, Solita Oy
Mindtrek 2017, Tampere
2. INTRODUCING: ME
› Programmer, activist, IT admin, manager
› Procurements as both buyer and seller
› Project lead / product owner in both closed and open source
development projects
› Software development and operations experience in new product
creation, maintenance, and phase-out
4. CAVEATS / RESTRICTIONS
› Chosen point of view: owners and developers of custom software.
› Custom software: software developed for a specific, in-house need.
› Software: anything that makes machines behave in a well-defined,
specific way.
TL;DR: Open source focuses effort where it has most impact
6. WHAT YOU ARE ABOUT TO HEAR
› Software ROI assessment is misfocused
• hard facts to show this
› Benefits of opening source
• soft facts to show this
› No harm in opening source
• only anecdotal evidence
7. WHAT IS THE BIGGEST
ASSET YOU GET WITH
DEVELOPMENT MONEY?
(in the sense that it will cost you the most to replace if you lose it)
9. EVIDENCE
› Software TCO breakdown*
• The real cost is not the software but making people work with it
› New tools and processes take time
• Time lost easily exceeds software development time
› Developer orientation time
* According to 84 TCO evaluations, reported at
https://www.computerworld.com/article/2570233/enterprise-applications/the-true-costs-of-software.html
10. SO?
Maximise the utility of
social/institutional change.*
Open source is an answer – or at
least, a piece in the puzzle.
* ”TCO is really about process improvement”,
https://www.computerworld.com/article/2570090/it-management/gartner--debunking-five-myths-of-tco.ht
ml
13. EFFECTS ON DEVELOPERS
› Pride
› Contact with users
› Focus on reusability, documentation, ease of deployment
› General-purpose components
› Compatibility
That is, maximum impact for the software.
14. EVIDENCE
Answers to “how will it affect your attitude or way of working if you
write open source?” with my colleagues
› “I put more effort in documentation, changelogs and compatibility”
› “It’s much more tempting to take code aesthetics seriously”
› “I think on a more high level, how it would be best done”
› “It feels good if I can publish something actually useful”
So… it makes you work in the way you should anyway?
16. EFFECTS ON USERS
Quite few, really. But:
› Empowerment
› Will to contribute
› “Team spirit”
› Gratitude
More people able to use and fix your
software!
17. SO?
Getting people to use your
software effectively is hard.
Open source has the
correct focus.
18. NO HARM IN OPENING
SOURCE
This is where it gets complicated :(
19. PUBLISHING SOURCE CODE
Nothing happens.
› No one will even notice
› Code no better/worse
› It’s the development process that counts
• better prepare right from the start
21. THREATS AND ANSWERS
› Security problems?
• Fix them. Now.
› Taking advantage?
• Licenses for mutual benefit
› Helping your competition?
• They don’t have your integrations, user base, and experts
• Maybe you should be selling your software as a service :)
• Maybe you should hire their experts.
22. SO?
The threats are imaginary.
OTOH, opening old codebases might
be useless or too much work
(depending on the code).
24. DOES IT PAY OFF?
› You want the psychological / social effects
› Best effect by actually opening the source
› Doesn’t help old projects much
Someone: “It’s not realistic to put all that extra effort in...”
Me: “How idealistic is it to do your work properly?”