Re: Non-English posts on Perlmonks

[Abigail-II]

The only conceivable problem I see with native language postings is moderation. What if something that appears to be another language is really just semi-randomly generate strings? Or something offensive (or in violation of copyright laws, etc) is posted and nobody else hear speaks that language?

There are three different issues. First, what happens if someone posts something offensive, and there's noone here that understands the language? Let me answer that with an old riddle If a tree falls in a forest, and there is noone to hear it, does it make a sound? Can something be insulting if there's noone insulted? Does it matter?

The second issue is, what if someone posts something that's a copyright violation. Well, what happens if someone posts something in English that is a copyright violation? My bet is that nothing happens, until someone notices and informs the right people. My bet is, the same happens if someone posts a copyright violation in a language other than English.

As for the third issue, it's far from trivial to generate random strings that look like a language but isn't. And suppose among the thousands of monthly posts here, one or two are a clever spoof. I don't think that's a problem. It hasn't been a problem in all the years this site existed, so I wouldn't worry about it too much.

Abigail

[Anonymous Monk]

so I wouldn't worry about it too much.

Neither would I. In fact, since I don't run this site, I wouldn't worry about it at all :)

That said, it's all a matter of cost vs. benefit. The probability of any damage actually coming from copyright infringement is next to none, worst case you get a cease and desist letter (and even that has a very low probability).

As for offensive content and falling trees, you never know who might walk along after the tree has fallen. It would reflect poorly on Perlmonks, but the odds are slim and definately wouldn't outweigh the benefits.

it's far from trivial to generate random strings that look like a language but isn't.

I've never done it, so I can't say for certain, but in theory it would appear to be remarkably easy. A simple example could be to just rot13 an english post, add/subtract a specific amount of padding from each word, and vary the punctuation slightly. More complex examples could be done in under an hour as well. Granted, I doubt it would ever cause a problem, but it certainly doesn't appear difficult.

[The Mad Hatter]

I don't know about you, but to me (and others I know), rot13 is blatantly obvious, and I doubt padding/punctuation would change that. I don't think making up a plausible language within an hour is possible.

[Anonymous Monk]

rot13 is blatantly obvious

Okay, rot14 then ;-P

Really though, languages follow many simple patterns. What you could do would be to take a couple hundred posts from Perlmonks (or anywhere else) and analyze them for average word length and order. So taking your post I could use the notation 'A' for an alphabetical character and 'P' for punctuation (you could obviously get more specific here) and you get:

A AAAPA AAAA AAAAA AAAP AAA AA AA PAAA AAAAAA A AAAAPP AAANN AA AAAAAA
+AAA AAAAAAP ...

You then average out the structure of the words and create general rules like 'a one-letter word is seldom followed by another one-letter word' and 'this type of punctuation occurs every X letters.' You use these rules to create an acceptable level of variation and then use some random generator to generate numbers in this variation. You then account for certain letters occuring more often than others and assign them accordingly. Dead simple.

Again though, it's hardly worth worrying about, but it is a neat (SIMPLE) academic exercise.