4 ms·

> no wonder a programmer sees the law as inneficient code to be refactored asap! ;-) Almost every time I hear programmers talk about the law arguing that it sh

> no wonder a programmer sees the law as inneficient code to be refactored asap! ;-) Almost every time I hear programmers talk about the law arguing that it should be treated like software and be far more precise and absolute than it currently is I'm reminded of this comic: https://xkcd.com/1831/ https://xkcd.com/1831/ There's good reason why most laws contain measured language such as the infamous "reasonable person." They are purposefully written to support changing interpretations as societies expectations of reasonableness and acceptableness change. Many programming based approaches to the law treat it more like a codification of exactly how a person should behave in every situation, and precise penalties should they violate strict protocol. It also reminds of the of Etherium DAO. How the mantra for Etherium was "the code _IS_ the contract" for so long, until it turned out the code had unintended behaviour and the so called "contract" was rolled back.

mrandish 7y ago
> Almost every time I hear programmers talk about the law arguing that it should be treated like software and be far more precise and absolute than it currently is I'm reminded of this comic Yep, it's hard. However, there are two things going on. The first is that there are people who are highly motivated to reinterpret or misunderstand a law they don't like or think is wrong. Even when it's pretty damn clear. The second thing is there are some passages in both the Bill of Rights and the law that aren't as clear as they could be (and not because they're intentionally written that way to provide latitude). I think these are "bugs" that could have been addressed with better drafting. I often ponder how one would re-write the constitution to be more clear, more effective and yet still resilient to change over centuries. It's a hard problem but I think there are certain meta concepts that could be applied that would be net improvements.
- datashow 7y ago
  American law system does evolve itself through case law. Things which are not clearly defined in the original language are clarified and adjusted through cases. Does it always get everything correct? No. But it's one of the best systems you can get on this planet.
  - int_19h 7y ago
    By what metric do you judge that system as "best on this planet"?
supernova87a 7y ago
Love the idea of programmers writing the Constitution. "It works for 95% of the cases, should be fine!" "Just write up a first working version and we'll iterate from there!" The writers of the Constitution were experts in history, philiosophy, religion, political theory, war, and even they hesitated and put multiple safeguards in case they got things wrong. What a laugh that the average computer science grad (or worse, just brogrammer) could even understand the range of considerations to be encoded into a country's operating principles...
- bllguo 7y ago
  just as much a laugh to deify the original writers of the constitution. do you really think we havent learned anything since? and why are you talking about average programmers being involved in this hypothetical process? i have no doubt we as a people could write a far improved constitution; the problem is that there is no suitable mechanism to get there, no way to avoid corruption
- K0SM0S 7y ago
  As someone who studied law, I agree with your general sentiment. I was not implying that programmers do that. ;-) I would however temper your dismissal of "average grads": in lawmaking too, drafts may be written or refined by young people who need to learn! But not without ultimate supervision, of course. Just like young code monkeys are totally able to touch production code in some settings. As for the deeper matter, "the range of considerations to be encoded into a country's operating principles..." — love the wording, btw. There's a long way until we are anywhere near expressing the full range of human languages in 'machine compatible' code, programming languages. One could argue there's not even a point in that, since you essentially attain redundancy; and there are more promising ways like ML to tackle the linguistic problem. However it's not black-and-white, the entire sum of norms that apply to us is a crazy vast space, and one that all (individuals and companies) should "optimize" to their particular case. In this as in many things, computing tools can help — and we certainly don't replace surgeons, senators or judges, we should already be content to immensely enhance their capabilities through technology. IMHO. "Refactoring" was my tongue-in-cheek way of saying that there's much to do in the way of organizing information in the legal space: datasets from which to base logic or standards upon are basically non-existent beyond non-computable literature (that could be modeled, we're talking structure of the law not content, hierarchy and structure of norms, not what citizens put into it). Morphing said structure to become mappable is nigh-impossible to do without concerted effort of most parties. I vetted first-hand some ideas with people working in the field, but the general obstacle is more economical, it's a business strategy/politics problem more than a technical one. I have no idea myself but I doubt this quadri-centennial culture is in any way, shape of form less entrenched at the highest public/federal/international levels.
K0SM0S 7y ago
As someone who took actual law courses, I can only share your sentiment. Briefly, - that xkcd is (as usual) pretty spot-on - the most ancient piece of "law" that we have is Hammurabi's (~3200 BC but don't quote me on that), and it's translated to "code". In a weird turn of linguistics, I guess modern incarnation of "code" as a concept (i.e. computer programs) is a very specific subset of "ordering reality", applying "norms", defining "systems". But we should not confuse the two, there's a reason we write law in human languages and code in computer languages and the two have proven quite unable / unsuited to reach the other's purpose. - a very active domain of lawmaking activity is "harmonization" (as called in the EU), the idea of making different law systems interoperate better with others (e.g. between States within a federation, countries in a Union, internationally altogether). The benefits (in litigation, cost) are enormous, and it's a massive undertaking for that sector. This is the kind of domain optimization I'm thinking about, where computing as a means can help humans better perform the job by orders of magnitude. Think e.g. tackling complex comparative graphs, effect/impact, etc. > Many programming based approaches to the law treat it more like a codification of exactly how a person should behave in every situation, and precise penalties should they violate strict protocol. Indeed, a chilling hell that hopefully we'll never be foolish enough to implement. But you know, I know, that the mere existence of the concept means someone, somewhere, sometime will do it. Note also that, in a "positive" way — e.g. to define a meritocracy in a private environment like a company or club — it's been proven viable (the case of Dalio is probably the most famous, but you could argue most corporations are/become such programmatic systems). - The whole notion of "blockchain as law", "code = contract" is, AFAIK (I read a lot about that in recent years), idealistic. Wishful thinking. It is no more no less "law" or "contract" than your run-of-the-mill compliance-motivated feature. It's born of the will to define a new space from scratch, to 'evade' current systems — physical and digital: banking, surveillance, etc. I don't wish to state about the intent, but the constitutional view is quite clear: such a departure from the accepted "norm" is secession or revolution, and the only way to change that is to become the norm and thus define itself as legal. It does not matter if it's a 1776 paper or a 2019 state of the art technology, the concept of legality remains. TL;DR: everything is/becomes illegal from the standpoint of preserving the existing regime / order / norms; in effect no new regime is ever "legal" under the ancient: you effectively revoke the old, then create a new one. I'm not sure how versed in public/state legal matters and the history of these things most blockchain proponents are when they believe creating a new legal domain is in any way different or easier because you write it in Java instead of English. The business angle is more legit, literally so when it conforms to "good practice" (regulations etc.); however self-defeating it might seem to the idealistic "new law / regime" proponents. Short story, all the talk about evolving law with/using blockchain technology is fluff. It's like saying "internet" or "social networks" will solve "social" altogether. Confusing pipes for content.

mrandish 7y ago

> Almost every time I hear programmers talk about the law arguing that it should be treated like software and be far more precise and absolute than it currently is I'm reminded of this comic Yep, it's hard. However, there are two things going on. The first is that there are people who are highly motivated to reinterpret or misunderstand a law they don't like or think is wrong. Even when it's pretty damn clear. The second thing is there are some passages in both the Bill of Rights and the law that aren't as clear as they could be (and not because they're intentionally written that way to provide latitude). I think these are "bugs" that could have been addressed with better drafting. I often ponder how one would re-write the constitution to be more clear, more effective and yet still resilient to change over centuries. It's a hard problem but I think there are certain meta concepts that could be applied that would be net improvements.

datashow 7y ago
American law system does evolve itself through case law. Things which are not clearly defined in the original language are clarified and adjusted through cases. Does it always get everything correct? No. But it's one of the best systems you can get on this planet.
- int_19h 7y ago
  By what metric do you judge that system as "best on this planet"?

supernova87a 7y ago

Love the idea of programmers writing the Constitution. "It works for 95% of the cases, should be fine!" "Just write up a first working version and we'll iterate from there!" The writers of the Constitution were experts in history, philiosophy, religion, political theory, war, and even they hesitated and put multiple safeguards in case they got things wrong. What a laugh that the average computer science grad (or worse, just brogrammer) could even understand the range of considerations to be encoded into a country's operating principles...

bllguo 7y ago
just as much a laugh to deify the original writers of the constitution. do you really think we havent learned anything since? and why are you talking about average programmers being involved in this hypothetical process? i have no doubt we as a people could write a far improved constitution; the problem is that there is no suitable mechanism to get there, no way to avoid corruption
K0SM0S 7y ago
As someone who studied law, I agree with your general sentiment. I was not implying that programmers do that. ;-) I would however temper your dismissal of "average grads": in lawmaking too, drafts may be written or refined by young people who need to learn! But not without ultimate supervision, of course. Just like young code monkeys are totally able to touch production code in some settings. As for the deeper matter, "the range of considerations to be encoded into a country's operating principles..." — love the wording, btw. There's a long way until we are anywhere near expressing the full range of human languages in 'machine compatible' code, programming languages. One could argue there's not even a point in that, since you essentially attain redundancy; and there are more promising ways like ML to tackle the linguistic problem. However it's not black-and-white, the entire sum of norms that apply to us is a crazy vast space, and one that all (individuals and companies) should "optimize" to their particular case. In this as in many things, computing tools can help — and we certainly don't replace surgeons, senators or judges, we should already be content to immensely enhance their capabilities through technology. IMHO. "Refactoring" was my tongue-in-cheek way of saying that there's much to do in the way of organizing information in the legal space: datasets from which to base logic or standards upon are basically non-existent beyond non-computable literature (that could be modeled, we're talking structure of the law not content, hierarchy and structure of norms, not what citizens put into it). Morphing said structure to become mappable is nigh-impossible to do without concerted effort of most parties. I vetted first-hand some ideas with people working in the field, but the general obstacle is more economical, it's a business strategy/politics problem more than a technical one. I have no idea myself but I doubt this quadri-centennial culture is in any way, shape of form less entrenched at the highest public/federal/international levels.

K0SM0S 7y ago

As someone who took actual law courses, I can only share your sentiment. Briefly, - that xkcd is (as usual) pretty spot-on - the most ancient piece of "law" that we have is Hammurabi's (~3200 BC but don't quote me on that), and it's translated to "code". In a weird turn of linguistics, I guess modern incarnation of "code" as a concept (i.e. computer programs) is a very specific subset of "ordering reality", applying "norms", defining "systems". But we should not confuse the two, there's a reason we write law in human languages and code in computer languages and the two have proven quite unable / unsuited to reach the other's purpose. - a very active domain of lawmaking activity is "harmonization" (as called in the EU), the idea of making different law systems interoperate better with others (e.g. between States within a federation, countries in a Union, internationally altogether). The benefits (in litigation, cost) are enormous, and it's a massive undertaking for that sector. This is the kind of domain optimization I'm thinking about, where computing as a means can help humans better perform the job by orders of magnitude. Think e.g. tackling complex comparative graphs, effect/impact, etc. > Many programming based approaches to the law treat it more like a codification of exactly how a person should behave in every situation, and precise penalties should they violate strict protocol. Indeed, a chilling hell that hopefully we'll never be foolish enough to implement. But you know, I know, that the mere existence of the concept means someone, somewhere, sometime will do it. Note also that, in a "positive" way — e.g. to define a meritocracy in a private environment like a company or club — it's been proven viable (the case of Dalio is probably the most famous, but you could argue most corporations are/become such programmatic systems). - The whole notion of "blockchain as law", "code = contract" is, AFAIK (I read a lot about that in recent years), idealistic. Wishful thinking. It is no more no less "law" or "contract" than your run-of-the-mill compliance-motivated feature. It's born of the will to define a new space from scratch, to 'evade' current systems — physical and digital: banking, surveillance, etc. I don't wish to state about the intent, but the constitutional view is quite clear: such a departure from the accepted "norm" is secession or revolution, and the only way to change that is to become the norm and thus define itself as legal. It does not matter if it's a 1776 paper or a 2019 state of the art technology, the concept of legality remains. TL;DR: everything is/becomes illegal from the standpoint of preserving the existing regime / order / norms; in effect no new regime is ever "legal" under the ancient: you effectively revoke the old, then create a new one. I'm not sure how versed in public/state legal matters and the history of these things most blockchain proponents are when they believe creating a new legal domain is in any way different or easier because you write it in Java instead of English. The business angle is more legit, literally so when it conforms to "good practice" (regulations etc.); however self-defeating it might seem to the idealistic "new law / regime" proponents. Short story, all the talk about evolving law with/using blockchain technology is fluff. It's like saying "internet" or "social networks" will solve "social" altogether. Confusing pipes for content.