Software is about promises(bramadams.dev)
bramadams.dev
Software is about promises
https://www.bramadams.dev/software-is-about-promises/
23 comments
Right. The work that Mark has done for more than three decades is amazing.
Unfortunately I think it has been unappreciated and largely unknown, leading to people either re-inventing things already known, or, even worse, trying to create something that essentially contradicts the theory and results and therefore failing.
Unfortunately I think it has been unappreciated and largely unknown, leading to people either re-inventing things already known, or, even worse, trying to create something that essentially contradicts the theory and results and therefore failing.
Not long ago I discovered that there was a https://en.wikipedia.org/wiki/Evaluation_strategy#Call_by_fu... term
While I can apply promise theory to call-by-future, call-by-future by itself does not really bring out the richness of promise theory.
Promise theory isn't really about computation. It is about voluntary cooperation among agents. Agents, for the purpose of this semi-formal language, is defined as something that can make its own promises -- that is, communicate intent to a set of observers, some of whom might also be agents.
Promises are not obligations, and as such, the intent to do something is not the obligation to do something. There is no guarantee that the intent will be executed at all. Maybe other agents are actually doing this by best-effort. Sometimes external circumstances can trigger a failure. Sometimes the agent is not able to execute the promise well. Sometimes, the agent may deceive, and deliberately communicate an intent when it is intend to do something else.
How well an agent kept is promises is an assessment -- crucially, assessments are not global determination of an agent's ability to keep a promise. Each agent makes its own subjective assessment on how well another agent keeps its promises. Understood in this way, this can model both centralized and decentralized systems of voluntary cooperation.
Promise theory isn't really about computation. It is about voluntary cooperation among agents. Agents, for the purpose of this semi-formal language, is defined as something that can make its own promises -- that is, communicate intent to a set of observers, some of whom might also be agents.
Promises are not obligations, and as such, the intent to do something is not the obligation to do something. There is no guarantee that the intent will be executed at all. Maybe other agents are actually doing this by best-effort. Sometimes external circumstances can trigger a failure. Sometimes the agent is not able to execute the promise well. Sometimes, the agent may deceive, and deliberately communicate an intent when it is intend to do something else.
How well an agent kept is promises is an assessment -- crucially, assessments are not global determination of an agent's ability to keep a promise. Each agent makes its own subjective assessment on how well another agent keeps its promises. Understood in this way, this can model both centralized and decentralized systems of voluntary cooperation.
Yeah I didn't skim enough through the article and since I just found out about call-by-future .. I premacommented.
Huh, interesting. I recognised the name from my time working with CFEngine – a very impressive piece of software in a space full of crap – but did not know there were books on it. Colour me intrigued.
Is the Thinking in Promises book too shallow or is it a good start?
Is the Thinking in Promises book too shallow or is it a good start?
I am not familiar with that book and that sounds interesting. I have not read that book so I won't make a recommendation on where to start.
I was working through "Promise Theory: Principles and Applications" and read through the first handful of chapters. For me at least, it was more accessible than math-heavy books. Yet, I can tell that it helped sharpened my reasoning ability about promises. I only made it through the first handful of chapters and yet it set a foundation that I was able to use over the years. Reading through some of the comments and reviews again, I see there are even more ideas that I could understand if I make my way through the book.
I was working through "Promise Theory: Principles and Applications" and read through the first handful of chapters. For me at least, it was more accessible than math-heavy books. Yet, I can tell that it helped sharpened my reasoning ability about promises. I only made it through the first handful of chapters and yet it set a foundation that I was able to use over the years. Reading through some of the comments and reviews again, I see there are even more ideas that I could understand if I make my way through the book.
> When publishing software, you make a promise to your users.
Just to add on to that. Beyond a promise, it's a contract, and someone has to be responsible and accountable for it.
Like when you're walking or driving and see a traffic light... you don't stop to wonder if there's a race condition or if another signal is out of sync. You trust it and act.
Unfortunately, it feels like in software today, promises are made... but rarely kept. And worse, most people just seem to accept that. If traffic lights were broken, we'd just need to upgrade to the next version right?
Just to add on to that. Beyond a promise, it's a contract, and someone has to be responsible and accountable for it.
Like when you're walking or driving and see a traffic light... you don't stop to wonder if there's a race condition or if another signal is out of sync. You trust it and act.
Unfortunately, it feels like in software today, promises are made... but rarely kept. And worse, most people just seem to accept that. If traffic lights were broken, we'd just need to upgrade to the next version right?
By definition the vast majority of software publishers are not that virtuous or credible.
Hence why large businesses pay a lot of money to get guarantees in writing from reputable firms, even though the nominally same software may be available for free.
Hence why large businesses pay a lot of money to get guarantees in writing from reputable firms, even though the nominally same software may be available for free.
I disagree. A lot of open source software literally come with "it's what it is" clause in the license, meaning that the maintainer has ZERO responsibility towards you, the user. No promises were made, you take it as is or leave it. Just because the dev generously decided to open source the code doesn't give you, the user, any "rights" to inundate the dev with issues.
> A lot of open source software literally come with "it's what it is" clause in the license, meaning that the maintainer has ZERO responsibility towards you, the user.
There's a loophole in that scenario, which is created by the fact that in FLOSS there is no clear distinction between users and mantainers. Meaning, the whole concept of FLOSS is based on community modifying and redistributing their own contributions, as a form of giving back to the community.
The same principle applies to wikis. You can take it or leave it, but the whole notion of a wiki lies on your ability to make it your own.
There's a loophole in that scenario, which is created by the fact that in FLOSS there is no clear distinction between users and mantainers. Meaning, the whole concept of FLOSS is based on community modifying and redistributing their own contributions, as a form of giving back to the community.
The same principle applies to wikis. You can take it or leave it, but the whole notion of a wiki lies on your ability to make it your own.
WARNING: Putting some code on the internet may yield curiosity and contact from other like-minded humans. The more potentially useful it is, the more people you may hear from. Gasp!
A simple "I put this up but please do not contact me" solves the problem.
A software license is not a social contract but that doesn't mean you can't bring some kindness and common sense to the situation.
A simple "I put this up but please do not contact me" solves the problem.
A software license is not a social contract but that doesn't mean you can't bring some kindness and common sense to the situation.
Yes, exactly. Especially when a lot of open source projects seem to be resume padders. They're presented as if they're the most useful library in the world with implied support, and then when your first issue is closed immediately with "yea, we don't care about that", you realize the reality.
It goes a long way if a "maintainer" simply states upfront about their mode of operation and sets expectations accordingly.
It goes a long way if a "maintainer" simply states upfront about their mode of operation and sets expectations accordingly.
> that doesn't mean you can't bring some kindness and common sense to the situation
It doesn't mean you have to either. There's no social contract either way, so there's nothing to debate. Doing whatever you want is the correct thing to do regardless.
It doesn't mean you have to either. There's no social contract either way, so there's nothing to debate. Doing whatever you want is the correct thing to do regardless.
Setting expectations is basic human decency.
Setting expectations is decent. However I suspect it to be advanced human decency.
Basic human decency is probably something more along the lines of 'live and let live'.
Basic human decency is probably something more along the lines of 'live and let live'.
`software can be anything in imagination, but must be something in reality.` - I like that. I should stick it on my wall.
This truism can be applied to any object that exists in physical reality.
The central insight is understanding that promises are not obligations, and why and how that matters. From there, interesting things can be analyzed -- using types and contracts in a development team, unit tests, integration tests, specs, user interface and user experience, compliance, signaling, APIs, etc.
I think it is particularly useful now in the age of LLMs, agenic AIs, and autonomous robots that have to navigate spaces shared with humans.
https://markburgess.org/promises.html