[udpheaders]

As if Python was the most powerful language, heh. :)

Maybe he was going for a world record of most citations in a single paper. Who's to say he wasn't just doing research? How many downloads is too many?

I would be willing to bet that the JSTOR TOS do not give a specific number. e.g. "You may not download more than n papers in 24 hours." And if they don't state a maximum in the TOS, then why shouldn't they, for clarity?

[danielweber]

It sounds like you are trying to invoke Loki's Wager -- since you cannot define N where downloading N is too many and N-1 is not too many, there must not be such a concept as downloading too many.

People who deal with the law don't have much patience for this.

[udpheaders]

I'm not sure if I am interpreting your last sentence correctly. Can you rephrase it?

[danielweber]

Courts have a very long history and lots of people have tried lots of really weird things over the years. Swartz will hopefully have a good lawyer, and a good lawyer won't even try to something along the lines of "there wasn't a preset limit on X therefore my client didn't do anything wrong" when his use of X was over a hundred times the combined consumption of all the legitimate users over two months.

Judges are not computers. If counsel presents them with a bad enough argument, they might get insulted that counsel thinks the judge is dumb enough to fall for it. Things that depends on the judge's mood (like purposefully obtuse arguments) are not a good courtroom strategy.

[udpheaders]

You are jumping around a bit. We were talking about TOS and now we're in a court room and using the words "judge" and "dumb" in the same sentence. I was kidding about doing research. Humor. We all know what he was doing. But the truth is I'm serious about these types of TOS. And I'm looking at this mainlly from the end user's perspective. You see the same type of ambiguous TOS language everywhere on the web. Let's stay focused on TOS for a moment, and leave aside the Swartz case. Do you think ambiguous contracts (TOS) are "better"[1] than unambiguous ones? For example, would reducing ambiguity lower the probability of (costly) disputes?

1. better for who?

[danielweber]

ToS might be interesting in some cases, but not in this one. JSTOR and MIT kept on denying access to Swartz and he kept on working around their defenses.

[udpheaders]

What if we looked beyond the Swartz case? Then what do you think about these types of TOS?

Maybe another example would be more interesting. Say you have a choice between an API that allows a "reasonable" number of requests in any 24 hour period and one that allows n number of requests in any 24 hour period. Which one would you prefer?

First assume you're an API user. Then assume you're the API provider.

Anyway, this kind of question is what I was getting at. What is reasonable? I don't know what their server capacity is.

I like using automation, I prefer non-interactive to point and click, and I have always found TOS on academic databases, not to mention most websites, interesting. Because they fail to account for anyone who might want to use automation (reasonably, having respect for the resources of the server). But maybe I'm the only one who finds this question interesting.