Looks like I'm going to be spending some time with the Privacy Act over the next couple of months, which is fine by me. #RulesAsCode

If you have some experience with #langchain and wouldn't mind some newbie questions, let me know. I could use some help wrapping my head around where "memory" lives, and what it's for.

Here's some promising results: under the hood a GPT-powered agent chose a strategy of 1. Collecting an ontology from a Blawx endpoint, 2. Encoding facts in that ontology, 3. Sending those facts to the "interview" endpoint, and 4. Generating a natural language statement of the answer. The interface still looks like ChatGPT, but by giving it the ability to use Blawx as a tool to solve problems, there is no risk of hallucination, and the reasons are explainable. #RulesAsCode #LawFedi #LegalTech

@hrefna I'm curious... I work on a visual interface for encoding statutes in a declarative asp language called s(CASP). I've recently discovered #rego as a place that a similar approach to #rulesascode has been taken. I wonder if you think there is any need or use for more friendly user interfaces in that space. Maybe in the context of service mesh?

I have found that lawyers who would be unwilling to try Prolog are excited to learn Blawx. Any similar disconnect in security-policy-as-code?

This new book celebrating 50 years of #Prolog has two chapters about #law, specifically. Another chapter is written by the people who wrote the Prolog variant Blawx currently uses. Yet another by the people who wrote the variant it previously used.

If you do #legaltech and you don't understand the relationship between legal rules and logic based knowledge representation, you are going to be very surprised by what's JUST around the corner.

#RulesAsCode

https://link.springer.com/book/10.1007/978-3-031-35254-6

Learned something very important about using Blawx to have non-programmers validate #RulesAsCode encodings, today. A process problem that I would never have anticipated, and it's going to take some long and hard thought in order to figure out how to get around it... Lots of useful feedback on the design of the tool, too, but the process thing has sent me for a loop...

Have docker installed? Wanna do me a favour?
Do a `docker run lexpedite/blawx` and shoot me a ✅ if it loads for you.

Thanks!

#LawFedi #LegalTech #RulesAsCode

Pushed a new version of Blawx, today, v1.6.16-alpha. Some small improvements to the documentation, plus some changes to the URL patterns that were requested by people using it as a back-end tool inside #GC. First time I've had to make changes to the API for users who aren't writing code, just consuming it.

Two more such changes to make later this week. Expecting another version to be pushed by the end of the week.

#RulesAsCode #LawFedi #LegalTech

Every once in a while, someone asks "what's new since last time," and I take a look at the changelog since then, and I'm genuinely impressed with how far things have come and how fast.

From the inside, I'm always looking at the next thing. But since the last GDE workshop I have made 6 major improvements to the language, 2 major improvements to the scenario editor, completely revamped the UI, and integrated with LLMs for summarization.

I feel pretty good about that.

#RulesAsCode

Going to be doing a session on #RulesAsCode at CaliCon 2023 in about half an hour! Can't wait.

So I'm coming to the realization that the perfect target audience for #RulesAsCode is Data Governance people.

Most of what they do is write, communicate, and implement rules, so they have all the same problems as legislative drafters. But ...

1. Their rules are about data, so they are already concerned with using well-defined terms.
2. Every rule is implemented in software.
3. Ease and scaling of compliance is their problem.
4. No technophobia or status-quoism.

Where do I find these people?

Today, I have a version of lawMirror that can generate all the same XML elements as CLEAN, including arbitrarily nested spans with custom ids, headings, and sandwich text. ProseMirror doesn't "do" nested inline elements, so the next question to answer is whether we can use some CSS magic to make the XML look right without breaking how navigation works. Then it's just a matter of cleaning up the UI, which will be a lot of work, but there are no roadblocks, there. #LegalTech #RulesAsCode #LawFedi

In a perfect world, I don't think this would be necessary to add to Blawx. But people (understandably) do not anticipate or appreciate the layers of abstraction that are required to separate legal logic from application logic. And the entire principle behind Blawx is "show, don't tell". So we'll just build an example of how to do it right.
7/7
#RulesAsCode #LegalTech #LawFedi

Prototype WYSIWYM web editor for #AkomaNtoso (aka #LegalDocML): https://github.com/Lexpedite/lawmirror

think "google docs for laws"

Please boost, and feedback welcome.

pinging @verbman

#RulesAsCode #LegalTech

For a couple of years now I've been looking for an open source, WYSIWYG, web-based XML schema-compliant editor for statutes. Everything I have been able to find is either closed source (LegisPro, GPSC), or text-based (LEOS, Cobalt). That is why I ended up building my own text-based solution specifically for Canada, called CLEAN as a stop-gap measure.

Today, I finally found a solution that can work. Open source, looks great, and it generates only valid LegalDocML. #RulesAsCode #LegalTech

The Blawx.com domain is down, which I'm working on. Apologies.

In the meantime, v1.6.15-alpha was released on GitHub today, adding ChatGPT integration to summarize explanations displayed in the scenario editor.

Symbolic and generative legal AI, together at last.

I'll make a demo video soon. Let me know if you want a tour.

#LegalTech #RulesAsCode #LawFedi #ChatGPT

This week has been absolutely head down coding like mad. Probably released 6-7 fixes for Blawx. Currently working on one to solve a problem with drop downs that look fine, but generate invalid code, because of a weakness in how updating dynamic option lists works.

I wanted to be doing more work on immunizations this week, but bugs got in the way. My "writing code IN/FOR Blawx" ratio has been about 1:20, this week. Hoping for much better next week.

#RulesAsCode #LegalTech

Finding a few bugs in the new interface, and in the date library. But I'm really liking the way the code reads with the new relationship feature. Way easier to use.

#RulesAsCode

v1.6.2-alpha of Blawx is now live at https://dev.blawx.com. No feature changes, it just looks less childish, now. Still childish, but "less" childish.

And only one new bug discovered so far!

#LegalTech #LawFedi #RulesAsCode

Good progress on adding "relationships" to Blawx today. Block generation and type checking and code generation is now working.

You can see in the Rock Paper Scissors example where "throw" is now a "relationship" between a player, a sign, and a game. That's a ternary (n=3) predicate, but the system supports up to 10-ary.

Just the API and scenario editor left to update. Scenario editor will be the hard part, but we might have it done this week.
#LegalTech #RulesAsCode #LawFedi

If you are interested in #RulesAsCode, you need to be aware that ErgoAI from Coherent Knowledge is very recently open sourced under Apache v2.

I don't know if I would say that makes it my favourite programming language for Rules as Code. I've become attached to multiple world semantics.

But it is at least #2.

Logical, non-monotonic, object-oriented, higher-order, argumentation theory, natural language explanations, lots of integrations with Java, Python, RDF, OWL, JSON, XML... 🔥 #LegalTech