Hacker News new | ask | show | jobs
by davidito 2427 days ago
Native speaker of Castillian Spanish here. IMO the most critical sentence in this judicial order was mistranslated by GitHub, changing our understanding of the government's request and casting GitHub's response in a more negative light (if you believe information and communication tools should be free.)

The original language in the last paragraph is "...para solicitar la retención del contenido y los datos relacionados con la investigación antes mencionada."

The Github English translation of this is (caps are mine) "...in order to request both WITHHOLDING the content and data related to the aforementioned investigation."

However, I believe a more accurate translation is "...in order to request PRESERVATION of the content and data related to the aforementioned investigation."

If you look up the Spanish verb "retener", it can have, broadly, two applicable meanings in this case: "withhold" or "preserve".

I don't believe the language in the order should have been interpreted as "withhold", because they are ordering to perform an action on the content AND the "data related to the [...] investigation." This would be IP addresses, access timestamps, GitHub account information, etc., I would presume. It doesn't make much sense to "withhold" that sort of data. It makes more sense to "preserve" it.

To me, this appears to be a request to preserve data, not to remove data.

A case can therefore be made that the GitHub decision to remove content went above and beyond what was expressed in the received written order.

9 comments

"retener" is indeed ambiguous, but in the sentence right before that paragraph they use "retenga" as "to withhold":

> Además, se han creado otros repositorios con la misma información para evitar que el contenido se retenga.

So I'm inclined to believe that the translation is correct.

(I'm also a Castilian Spanish speaker :)

That's a good point. But this is not the actual court order, merely an explanatory letter from a police agency. The language is very "squishy". The preamble to the actual order, where the police explain the rationale, indeed uses the verb "retener" to mean "withhold/block". But the last sentence, if read alone, make more sense as "preserve." I believe this "squishiness" is no accident.

We really need to see the actual court order. Github has not included it in this repo.

Does anybody know how to access Spanish court orders online? Are they public?

Are you literally arguing that both the Spanish native speaker who made the translation as well as the Spanish native speaker who wrote the "explanatory letter" got the meaning of a Spanish word wrong? :/
There is basically no way to parse that as "preserve" without the rest of the document stopping to make sense:

E.g

> judicial authorities are responsible for the supervision and control of websites in order to prevent the dissemination of criminal content

> other repositories with the same information have been created to prevent the content being withheld|preserve.

It is quite clear they meant withhold. The document just does not "click" anymore if you parse it as "preserve".

Also: No one in Spain would use "retener" to mean "preserve" as a synonym of "archive", anyway. "Retener" might loosely translate to preserve as in "to remain unchanged during a flow/process", but not in any of the other senses, and this is not the sense that is used here.

And, if there are was possibly any doubt left, Spain actually blocked the corresponding github.io subdomain (redirecting it to the usual "Your national FBI equivalent has blocked this domain for your protection" page) before sending this Github takedown notice (likely because they did not dare "withhold" the entirety of github.com itself).

Now im confused. Did they order github to take it down, or make sure its not deleted.

Or is everyone confused because a critical meaning is lost in translation.

Native Spanish speaker here. I understood it as preserve.
If that's the case, then github was already compliant with the order as the code is checked in. The Spanish gov't would just need the correct git hash of the repo state they are interested in.
Is it for sure that GitHub did the translation? Not all of the gov't takedown notices have been translated to English. Perhaps the notice came both ways?
Native portuguese here. the confusion may arise from uncommon usage of the word. the word "reter" (portuguese for retain) can also mean "preservar" (preserve), but is more commonly used in a judicial setting. ex: An ISP is ordered to keep logs that would otherwise could be deleted at some point in time

The contents seem pretty clear to me. An investigation is undergoing, the spanish law that makes mandatory keeping logs (also exist in Portugal, 6 months min.) is referred and github.com was asked to keep this data.

So basically the english 'keep the information'. You can keep it as in not give it to someone else, or keep it as in preserve it. For physical objects this makes no difference.
Ok. We know you know a lot about the "Castillian Spanish" language. And maybe about a ton of other things. But it seems kind of pointless, given the note was already translated by the Spanish Guardia Civil, don't you think?
So retener is like retention?
Can be translated also as "to save" or "to keep for yourself only and do not show to other people". Is a polysemic word.
It's also used to mean "withhold" or "block", which is why you have to look at the entire sentence to contextualize.
My (Chileanized) Castilian is pretty rusty but I thought your translation was in line with how I read it. Instead of "withhold" I would say "hold onto for myself" ; as in "Retener mis derechos". The broader context seemed to imply preserve the records as in don't delete anything as we investigate, but I can see how it could be taken the other way. In the sentence as written what word would you use in place of retener if you meant to be clear to stop dissemination or publication of the content?

Edit: In place of "la retención".

It's a verb, so retain, but yes.
In English too, you can retain something to have it, but you can technically also retain to mean "withhold," in the sense that you are keeping it secure or intact and refusing to divulge it.

It's not a very common usage in English though.

Retain in the sense of withhold makes more sense when you're talking about a physical item that can't be duplicated and shared without limit. I don't think I would use it in that sense when talking about data.