Hacker News new | past | comments | ask | show | jobs | submit login

I haven't commented on HN in a while, but this article actually pretty much complains about me, since I wrote the text in question and enabled the policy to be released.

So i'm just going to say:

1. The author claims the Google states something, then doesn't actually quote anything google stated, but instead writes their own interpretation of the words and a made up example. That's not a good start.

2. Having set up their own example, the author then proceeds to say it's wrong because of an appeal to authority, asserting no reasonable disagreement is possible - when not only is it possible, but large numbers of IP lawyers disagree on what the AGPL requires. It's true that someone wrote it, and there are FAQ's, but depending when and where, their opinion on what it means is not as relevant as one might think.

In truth, lots of people disagree over interpretation of the AGPL - many more than disagree about GPL interpretation or LGPL interpretation (which are fairly settled at this point). Searches over any legal licensing list in existence will you this.

When I wrote the interpretation you see here, it was the best available info at the time, guided not just by own views, but by listening to a lot of smart lawyers, counterparts at other companies, etc.

Even if you ignore all the lawyers and whatever as useless, the author has the huge problem that There are people who make AGPL software that take the view listed in the google policy

So it's not just Google or lawyers, it's software authors. Not just a few, either.

While companies are generally happy to ignore one or two people whose interpretation is outside the norm (and just not use their software), that's much harder when there is such widespread difference in view of the AGPL.

3. For no particular reason at all, the author then decides to assert a tremendous amount of bad faith in interpretation of the AGPL and reasons for publishing such policies.

We published our policies because over the years I (and others) were repeatedly asked by counterparts at other companies, various communities, and others to understand what our policies look like, for a variety of reasons (to understand for themselves, to use as a template, etc)

Period. That's the whole reason. It even says this: https://opensource.google/docs/

The author here has decided that's clearly all lies or deliberate misinformation, without even bothering to even ask anybody.

I can confirm what Danny says here.

It's a shame the original article has such flaws in its argumentation, because at the core of it I think there are some good points. Notably, other people don't have to blindly follow Google's example and reasoning. In fact, as discussed here, there are good reasons not to. For one, because Google is so highly integrated (with Borg and its monorepo, not to mention custom approaches to security), it's harder to use AGPL software without risking linking. Most other users outside Google would just apt-get a package or the like. And of course the business risk to Google is large simply because of their scale.

You can even disagree with Google's decision not to use AGPL, but they have openly explained their reasoning, and I think that adds to the discourse.

Disclosure: I used to work at Google, and was fairly actively involved with open-source activities in different capacities while I was there. I am also on record as sharply criticizing Google when I believe that is warranted. This is just not one of those times.

Thanks for having publiished these policies publicly! You gave a pretty clear summary of Google's motivations for banning AGPL on HN five and three years ago two of the previous times this came up, and having worked at Google 2011-2015 (including some interactions there with you), your public explanations of why from back then are extremely credible and rational given the logistics of what's easy and hard inside Google.



Not that I'm surprised, as you've always been pretty forthright and honest. Thanks for continuing to be that way.

Of course. I appreciate the thoughts. I try my best to be transparent and straightforward.

I actually do think it's reasonable to expect companies to re-evaluate every few years (at a minimum) because the world changes. At some point, the reasoning from X years ago will not make sense anymore.

One of the top rules of sane policy making is to write into the policy the conditions that should cause you to re-evaluate it, whether it's time, or change in evidence you relied on, or ...

You want people to understand your rationale and what might change your mind, otherwise you shouldn't expect them to go along with it.

Even though such a rationale exists in various emails, etc, I actually failed to do that in the resulting police here, unfortunately (the internal version is not any better in this respect). That one is on me and I think would be a more legitimate complaint here.

I don't think it would have changed yet (if anything, the AGPL has become more contentious as the economy has gone south), but that's just an at-a-glance view.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact