Ever wanted to programmatically file a lawsuit? In Poland, you can. | Hacker News Reader

Ever wanted to programmatically file a lawsuit? In Poland, you can.

blog.danieljanus.pl

60 pointsnathell35 comments

Excellent, soon computer-controlled companies can automatically sue each other.

"The president of agalmic.holdings.root.184.97.AB5 is agalmic.holdings.root.184.97.201. The secretary is agalmic.holdings.root.184.D5, and the chair is agalmic.holdings.root.184.E8.FF. All the shares are owned by those companies in equal measure, and I can tell you that their regulations are written in Python. Have a nice day, now!" (http://www.antipope.org/charlie/blog-static/fiction/accelera...)

We all went there pretty much immediately, didn't we? I know Charlie reads HN occasionally; I assume if he hasn't already seen this, he soon will...

It's almost as if Douglas Adams had written Dickens' Bleak House.

I could have sworn it rather mentioned Lisp. Maybe I'm imagining it though.

On a pro-python note, I'm also reminded of the not-that-long-ago idea by the SEC to require certain securities to include executable documentation: https://jrvarma.wordpress.com/2010/04/16/the-sec-and-the-pyt... and http://www.itworld.com/government/105031/will-wall-street-re... for example.

This court is in my city, I even worked for a while with people that manage the system :)

This is only for very repetative cases about debt collection enforcement - courts were overwhelmed by these before, and now this specialised court handles them.

Import this into the USA, and the RIAA/MPAA will integrate the API into their torrent detection tools before you can say dmca. Then there will be an automated interface to subpoena ISPs for subscriber information (GET /subscriber/info?ip=123.45.67.89&time=1234567890&judge_signature=b76c8a23), and the only thing missing will be an automated interface for customers to be notified, view, and quash subpoenas for their information. The last one is crucial when the lawsuit is being filed thousands of miles away, but I doubt that the MAFIAA will allow that to happen.

This short piece of fiction seemed so far off before. Now it seems closer: http://www.ftrain.com/nanolaw.html

Dang, I've been mistakenly attributing this to Stross in my head this whole time.

Not mistakenly: it's in Accelerando - see the first comment.

Once lawsuits can be filed by algorithm, any court system would be completely overwhelmed unless it had a) some system for limiting the number of lawsuits and/or b) some system for settling lawsuits that would involve an algorithm.

Some system for limiting the number of lawsuits could be as simple as charging filing fees. Ones that scale in proportion to the number of lawsuits you file would be ideal (first case: free; 1,000th case: 10 million euros).

Some system for settling lawsuits sounds scary, but I actually kind like the idea of an algorithm that checks the number of lawsuits you have filed, and automatically returns a case dismissal if it is above a certain threshold. I'm sure there are some grave implications for having such a cavalier approach to dismissing lawsuits, but for some reason none are occurring to me at the moment.

And thus begins a competition between legal quants. I don't know if this is better, worse, or both.

I think that falls into the "effect-free regulation" pile, rather than the "unanticipatedly harmful regulation" pile. Algorithmically spinning off legally distinct entities to file more lawsuits can't be that difficult.

Some system for limiting the number of lawsuits could be as simple as charging filing fees.

It seems that there is already a fee:

To give you an example, the FileLawsuits method returns a structure that consists of, inter alia, the amount of court fee to pay, the value of the matter of dispute (both broken down into individual lawsuits), and a status code with a description.

I assume that you can't actually settle the suit programmatically, though - otherwise, you could probably gamble on winning some percentage of lawsuits in order to pay for the fees of the losing ones.

Stock markets are so 2000's.

(As an aside, can we please get some sort of sane mechanism for quoting text?)

> (As an aside, can we please get some sort of sane mechanism for quoting text?)

I've always done it like email.

Or you can be a journalist, and put quote marks around what you are quoting.

That gets messy for multi-line quotations, and paragraphs. :(

> (As an aside, can we please get some sort of sane mechanism for quoting text?)

I use email-style quoting with italics.

But full-blown Markdown support would be so much better, especially since this is a programming-oriented website.

If you focus too much on the lawyer filing the lawsuit then smart lawyers will prey on young graduates. If you focus too much on the plaintiff filing the lawsuit then smart lawyers will probably find a way to make special corporations which can cheaply make the complaints before being dissolved.

Given the precedent of the DocX foreclosure signature fraud, the DirecTV smart card lawsuits, and the RIAA lawsuits, I think bulk lawsuits are a ripe territory for a combination of fraud and lawsuits with poorly thought-out evidence that only win because settling is cheaper. With the current system, there is an assumption that a human at least looked over the lawsuit before filing, but if programmatic lawsuits are made possible, we may see the first lawsuit that literally was not thought out at all before filing. In this case, I think it's best to retain a human element to the filing process, much like we have dead man's switches on trains.

This is "payments claims court" - presumably it's a fairly deterministic bureaucratic process that's required to go through to deal with debtors in default, perhaps before you can start chasing them properly.

That's correct. Combined with a high barrier of entry (you need to create an account of the appropriate type AND pay a fee -- 1,25% of the value of the matter of dispute -- for the case to be processed) this is what makes it viable in the first place.

Yup, that fees quickly add up. I work for a company which builds interfaces between clients' systems and e-court (http://www.infisolus.pl/oferta/elektroniczne-postepowanie-up... - sorry, polish only) and I can attest that they couldn't actually "spam" e-court with lawsuits - their budgets for this part of collections process traditionally are not so big.

This actually looks like - let's build an api "because we can". Not "because it's reasonable". I presume, that process of suing still requires some kind of human approval and post-processing. An idea of e-court is great, if restricted to small and simple claims, though.

Edit: I am astonished by these numbers. I don't believe that this is court per se. Can somebody elaborate? 1.6 million issued payment orders is huge number even for Poland :)

The adjudicating personnel of the e-court comprises four judges and 96 court officials (referendarze-officials adjudicating independently in certain categories of cases) who are supported by 70 clerks. As of October the 1st 2011, 2 000 000 lawsuits have been lodged in the e-court, and in some 1 600 000 cases payment orders have been issued.

API serves as a way to file lawsuit. It is humans who process them. This court deals with standard claims like businesses not paying for the service/goods but also consumers not paying their bills.

API is useful for companies like mobile operators or utility providers who have thousands of subscribers who do not pay.

1.6M payment orders, 21 months, 200 staff = ~17 cases per day per person (no holidays and vacations). Sounds doable given it is likely to be batches of claims as described above.

I have mixed feelings about this: "The methods include FileLawsuits, FileComplaints, SupplyDocuments, GetCaseHistory and so on (the actual names are in Polish)." I'd never even think of creating an API that doesn't use English method names (even though it's not my first language). It doesn't matter whether this is used only internally at the moment. At some point you'll need to integrate with others in one way or another.

Then again... it was created for Polish people and it probably contains some law-related vocabulary which not everyone will be familiar with if it's translated.

But if they ever make it stable enough to license the system to other countries, it would be a hard thing to change. Can't decide what I think of this idea.

I agree, but as someone who's tried to implement an application modelling a specific legal (financial) domain in Danish, I can attest that it's painful to forcefully map local terms to English one where there, more often than not, are no direct equivalents.

I'm Polish and I have worked on a project with a very specific locally oriented domain (politics and administration) a few years ago, and I must say that we tried to do it in English and abandoned the idea after a few hours during the design phase. It was just counter-intuitive to translate every institution name and process to English, even though we all were fluent with the language. Not all of the terminology can be translated 1:1, and we'd have to look up not only the words, but even definitions of the respective English term to make sure it's applicable, constantly check how various related entities' names translate, etc. And all of this for nothing, it would be useless abroad anyway! Huge waste of time and brain resources. We ended up coding in Polish. One serious downside: after working on such a project for months, you get to the next one and start coding it in Polish - ugh.

We should just write the laws in Prolog and call it Prolaw. Then we no longer would have to call the web site optimization peoples by the disparaging term "spammers" but could instead call them by the marginally less disparaging term "lawyers," and they could command much higher salaries.

Is there anything good about this? (serious question)

Lots of adventages. The most improtant is - one small institution serves big amount of traffic, freeing rest courts in Poland to judge regular cases. This makes long wating time for each case shorter, and this is very important in many cases.

It's the difference between functional law system, that people treat as one of the options to solve problems, and dysfunctional one, where nobody bothers suing anybody, because it will take 5 years and cost fortune.

I don't think we in Poland have good law system, because waiting time is still too long. But it's step in the right direction.

As far as I know, you cannot defend yourself in the same way the person or institution attacks you. That is, if you want to respond, to defend yourself in court, you have to fight using normal court procedures, not e-court procedures. Proceeding in normal (pen and paper) way is more expensive than proceeding in electronic form, though the former is the only way to cancel the decision of e-court.

Moreover, the vast amount of cases solved by such small number of workers suggests that processes in such e-courts are short and irrelevant. The sentence is automaticaly set to 'guilty'.

EDIT: clarification

Yes, but that's exactly the idea - to speed up the irrelevant obvious cases where sued person don't even try to defen himself - there's many such cases, and before they were solved in normal procedure, with tons of buerocracy, and that used up judges and other workers time, making more complicated cases longer as well.

Sorry, but this is not the level where whole case should be solved. The court is by definition a place where both sides speak, and e-court is a place where only one side speaks, and she doesn't even has to prove she's right.

http://www.dziennikwschodni.pl/apps/pbcs.dll/article?AID=/20...

"Elektroniczna procedura przeznaczona jest dla tzw. spraw upominawczych, które wytaczają głównie dostawcy masowych usług, np. zakłady energetyczne czy operatorzy komórkowi. Chodzi w tych sprawach o należności dobrze udokumentowane, a dłużnikami są zwykle osoby fizyczne, które zazwyczaj nie kwestionują sądowego nakazu. Gdy nie ma odwołania, nabiera on skutku prawomocnego wyroku."

It means e-court is for cases where there is good documentation, and persons that were judged can appeal, but if they don't (and they usually don't), these cases are processed automatically.

"Tylko w 4 tys. spraw e-sąd nie wydał nakazów i np. oddalił pozew, umarzając postępowanie. Sprzeciwów złożono zaledwie 3229, a więc w nieco ponad 2 proc. spraw. W takich sytuacjach sprawy przekazywane są do zwykłych sądów."

Only in 4 tousand cases e-court changed default decision, there were also 3229 appeals (2% of cases - this artcle is old a few months so this won't add to 2 000 000 cases). In such cases the matter is transferred to regular court.

What's wrong with that?

Appeal should be possible in e-form too.

yet another way to log jam the justice system

http://www.nytimes.com/2012/03/11/opinion/sunday/go-to-trial...

Reply on news.ycombinator.com