4 ms·

Beyond lawcomic[1], what is a good resource for this? If a lawyer wants to learn enough python to create a script that joins PDFs together, they can pick up a c

Beyond lawcomic[1], what is a good resource for this? If a lawyer wants to learn enough python to create a script that joins PDFs together, they can pick up a copy of Automate the Boring Stuff with Python[2] of the law. If an engineer wants to answer a question like “The IRS and HMRC consider these stock options to have a specific tax status if an employee exercised them within 90 days of leaving a company. What defines the date of purchase of a stock? The date the payment is received or the date the stock certificates are delivered?”, what can they do (besides ‘ask a solicitor’) to learn an answer they can be sure of? [1] https://lawcomic.net/guide/?page_id=5 https://lawcomic.net/guide/?page_id=5 [2] https://automatetheboringstuff.com/ https://automatetheboringstuff.com/

md2be 7y ago
Tax law is a lot like coding, the answers are there if you know where to look. Generally, look to regulations for definitions, publications for working examples, and case law for ambiguities and ombudsman/case for unanswered questions. In your case, the exercise date is the date of issue https://www.irs.gov/publications/p525 https://www.irs.gov/publications/p525 https://taxpayeradvocate.irs.gov/ https://taxpayeradvocate.irs.gov/
- devnulloverflow 7y ago
  Yes. In fact the tax law as the authorities apply it probably is code in the sense that they probably shove some inputs into some Excel spreadsheet to calculate your tax return. (I'm exaggerating, they probably have bespoke software which they can use for auto-generating their spreadsheets).
derefr 7y ago
There's a fundamental difference between your two examples. The lawyer only needs to learn a novel skill (Python), and then, said skill in hand, they can read through a few cookbook solutions to their problem on e.g. StackOverflow, and come away with an answer that is guaranteed to work—for at least their particular use-case—because they can test and iterate on it until it does work. The programmer, on the other hand, would need to learn vast swathes of a body of knowledge in order to have the requisite context to understand the "subtext" (i.e. case law) governing any particular "cookbook" solution they find. Or, to put that another way: as programmers, in the decades we've been practicing the craft so far, we've seemingly held as one of our "guiding principles" that implicit context (e.g. DSL/macro "magic") is generally bad: the more you have to know to start working on a new project, the worse a situation you're in. Programmers actively try to lower the "barrier to contribution", by 1. favoring larger libraries and fat language stdlibs over small atomized libraries, as large libraries can be shared as a lingua-franca/Schelling point; and then 2. avoiding abstractions that aren't built into the language itself, instead choosing a language for its built-in abstractions, as then programmers who "know language X" can be guaranteed to be ready to parse those abstractions. Law has no such principle. The "law community" has in no way ever sought to make it easier to maintain or update existing legal text, and therefore, has never really developed an equivalent to the concept of "write-only code." They might make it easier to get started writing the write-only code (with templates) or might discourage writing more law than there needs to be (red-tape reduction) but there's nothing disincentivizing any particular legal document from being as impenetrable—and reliant on knowledge "from the environment"—as possible. --- Or, to reformulate in a more productive way: if the whole body of law of a country were maintained in a reified form as some downloadable expert system knowledge-base, then anyone could learn the "pure skill" of legal analysis, which would be the skill of properly querying said knowledge-base (and, only if the answer isn't clear-cut, asking someone who has experience in predicting the way legal battles over such an issue would go.) That's sort of the way medicine seems to be trending—doctors are more and more becoming people who practice the "pure skill of diagnosis" and less-and-less people who have a particular knowledge-base loaded into their heads. But law hasn't even begun to chart its way down this path.

md2be 7y ago

Tax law is a lot like coding, the answers are there if you know where to look. Generally, look to regulations for definitions, publications for working examples, and case law for ambiguities and ombudsman/case for unanswered questions. In your case, the exercise date is the date of issue https://www.irs.gov/publications/p525 https://www.irs.gov/publications/p525 https://taxpayeradvocate.irs.gov/ https://taxpayeradvocate.irs.gov/

devnulloverflow 7y ago
Yes. In fact the tax law as the authorities apply it probably is code in the sense that they probably shove some inputs into some Excel spreadsheet to calculate your tax return. (I'm exaggerating, they probably have bespoke software which they can use for auto-generating their spreadsheets).

derefr 7y ago

There's a fundamental difference between your two examples. The lawyer only needs to learn a novel skill (Python), and then, said skill in hand, they can read through a few cookbook solutions to their problem on e.g. StackOverflow, and come away with an answer that is guaranteed to work—for at least their particular use-case—because they can test and iterate on it until it does work. The programmer, on the other hand, would need to learn vast swathes of a body of knowledge in order to have the requisite context to understand the "subtext" (i.e. case law) governing any particular "cookbook" solution they find. Or, to put that another way: as programmers, in the decades we've been practicing the craft so far, we've seemingly held as one of our "guiding principles" that implicit context (e.g. DSL/macro "magic") is generally bad: the more you have to know to start working on a new project, the worse a situation you're in. Programmers actively try to lower the "barrier to contribution", by 1. favoring larger libraries and fat language stdlibs over small atomized libraries, as large libraries can be shared as a lingua-franca/Schelling point; and then 2. avoiding abstractions that aren't built into the language itself, instead choosing a language for its built-in abstractions, as then programmers who "know language X" can be guaranteed to be ready to parse those abstractions. Law has no such principle. The "law community" has in no way ever sought to make it easier to maintain or update existing legal text, and therefore, has never really developed an equivalent to the concept of "write-only code." They might make it easier to get started writing the write-only code (with templates) or might discourage writing more law than there needs to be (red-tape reduction) but there's nothing disincentivizing any particular legal document from being as impenetrable—and reliant on knowledge "from the environment"—as possible. --- Or, to reformulate in a more productive way: if the whole body of law of a country were maintained in a reified form as some downloadable expert system knowledge-base, then anyone could learn the "pure skill" of legal analysis, which would be the skill of properly querying said knowledge-base (and, only if the answer isn't clear-cut, asking someone who has experience in predicting the way legal battles over such an issue would go.) That's sort of the way medicine seems to be trending—doctors are more and more becoming people who practice the "pure skill of diagnosis" and less-and-less people who have a particular knowledge-base loaded into their heads. But law hasn't even begun to chart its way down this path.