Casetext (YC S13): New Angle on Legal Research With Wikipedia-Style Annotations | Hacker News Reader

Casetext (YC S13): New Angle on Legal Research With Wikipedia-Style Annotations

techcrunch.com

68 pointsjoannahuey23 comments

J/J: I'm a deal lawyer, so I rarely look at caselaw. But regs. and statutes are my bread and butter. Any thoughts about how you would approach annotating these kinds of materials?

There's of course an interesting version control problem when statutes get amended--how do you merge your annotations which are anchored to the old text into the new text?

We're planning on adding regulations and statutes, and the annotation platform should generally work for these texts as well. As you noted, there's a version control issue, and we're still thinking about the best way to deal with it. One possibility would be to have the text appear as a redline, which would make keeping all annotations alongside the text possible. Alternatively or in addition, we could have a revision history of the text (allowing you to view and compare older versions) -- then, the easiest way to manage annotations might be to keep annotations for paragraphs that haven't changed in the newer version, but to have annotations for changed paragraphs only appear in the older versions.

Regulations and rulemaking info would be very useful for the area of law I'm interested in. Though annotations would be a great addition, I think just having a simple UX for searching and accessing these documents (Orders, filed comments, etc) would be a huge win. This is a problem I've been working on mostly as a hobby to scratch my own itch, and I'd be happy to discuss what I've learned.

One big pitfall I've noticed with taming regulatory data is that often each agency has its data siloed differently. Regulations.gov has been great for harmonizing many, many agencies, but here's a list of all the agencies that are currently "non-participating": http://www.regulations.gov/docs/Non_Participating_Agencies.p.... This means you'd need a parser/scraper for each agency, assuming its rulemaking data is even made available in a sane format (like a spreadsheet).

Anyways, best of luck. Successfully wrangling legal data seems more difficult than building the Hyperloop.

Thanks for the information! It'd be great to chat some more about what you've found out.

Nice--as the above poster notes, linking statutes and their interpreting rules and regs. in a sensible way could be a nice differentiator. E.g. if a regulation or another code section gives a special definition for a term used in the statute, it shouldn't be too hard to generate a default annotation linking to that definition, especially if you're already doing this for inline citations in cases.

Definitions are not as uniform as citations, but they are still pretty formulaic. "_________ shall mean ________" etc.

That's a great idea, thanks!

You can annotate statutes and regs on PlainSite, and you can link them to case dockets:

http://www.plainsite.org/laws

Nice implementation!

This sounds awesome. I've done a little digging (probably 6 months of probing worht) into technology tools for lawyers (mainly in the UK) and the it's a massively underserved and a big mess.

> Joanna Huey attended Harvard Law School, where she was president of the Harvard Law Review, and Jacob Heller attended Stanford Law School, where he was president of the Stanford Law Review. They later served together as clerks for Judge Michael Boudin at the U.S. Court of Appeals for the First Circuit in Boston, and worked at law firms.

Joanna/Jacob - I'm sure you know this already, but get some old school, grey-haired advisors for you company. Trust me. You need to understand how these people eat, breathe, and sleep. If there's one thing I've learned about lawyers (especially ones who have been in the business for 20+ years) they are very picky about how they do business. These guys happen to be the guys with all of the money as well.

Glad you're excited! We agree that there's a ton of room for innovation in the area. And thanks for the advice -- definitely a great idea to better understand the market and add to our first-hand experience.

Call me crazy, but it seems like a marketplace exactly like ThemeForest/CreativeMarket so I wonder why their business model does not follow the ThemeForest business model...where the law firm would buy bulk credits in $100/$200/etc increments and then they would have clerks read and review and annotate cases which can be purchased on a micro level (like say $5 for a case) so that CaseText can provide high quality content for lawyers bc the creators of the content would be getting a kickback.

Then, maybe, the cases could be bundled together in logical formats so that they could distinguish themselves from the big guys by not only providing a better tool, but logically grouping cases to ease the burden of the lawyer's ability to search effectively!

Not crazy at all! But, we think we can get higher quality annotations by focusing on building a community of lawyers interested in demonstrating their expertise (to build their online reputation or get referrals) than by hiring clerks.

We are working on improving search by grouping cases by issue -- users can already tag cases with various legal terms, and we'll incorporate those tags into search to improve the results.

Fair point. I just want y'all to make money so just throwing ideas out there! :) Good luck on y'alls journey!

btw I need to learn the lingo, bc I didn't actually mean clerks, I meant the lawyers you were referring to...I think clerk may mean low wage unskilled worker.

I'll reiterate a question appearing in the comments to the article. Is anyone familiar with how Casetext will handle shepardizing case law ?

Thanks for asking! At the moment, we automatically link all citing cases and then allow users to categorize them (by negative, distinguishing, and positive treatment) and to write a bit of context explaining how the case is relevant. In the future, we'd also like to automate some of the categorization.

To follow up on this -- we believe that over the long-run this will produce better results than Lexis and Westlaw. Because users can add any related cases (not just those that cite the case), you get a more complete picture of other cases you should be looking to in your research. For example, in some circuit splits, neither circuit cites the court that it is splitting from (and so doesn't appear in West/Lexis' database), but would appear on our site.

Users can also vote up or down a case based on relevance. Voting will help you quickly find the most relevant resources based on the thoughts of dozens or hundreds of other attorneys.

Is there any information or policy on third-party integration / API? There are other tools for legal writing and research that could benefit from integration with Casetext.

Interesting. We haven't implemented third-party integration, but we'll definitely think about it. Feel free to contact us if you have particular suggestions.

Also, justification might look more "legal" in that it's a closer match for typeset cases but I would be curious whether your users might prefer ragged right.

Interesting -- we haven't gotten other requests for this change, but we'll keep it on our radar. One possibility would be to add settings for text display that allow people to choose justification (along with font size and other such things).

Congrats on getting some good press, Joanna and Jake. He never mentioned you guys came out of YC - we'll be following you up then soon I hope.

Matt, CommentLaw

Good luck on getting into YC! We love innovation in the legal space and think we need more of it, so we'd be happy to advise you through the process -- shoot me an email or PM.

Go Harvard-Westlake!

Reply on news.ycombinator.com