Hi,
<tl;dr> I do not really enjoy the way the mandatory-for-editors HTTPS was introduced, mainly for time frame and communications (still) reasons, although I’m globally really enthousiastic about a better security and particularly the activation of HTTPS. Generally speaking I do _hope_ in the future WMF will give more time and more discussion space to handle major changes. <end tl;dr>
History: (I concede I may lack some readings, but I think I have the big picture)
After the PRISM scandal in June (2.5 months ago) everybody condemned that program and the Internet security became a major concern for Internet users. HTTPS is in important means to improve the security (although concerns about the protocol and the way it is implemented appear) and since it was a matter of time before it could be globally activated the blog post published on August 1st announced HTTPS will be activated for logged-in users 20 days after, with solutions about the blocked China HTTPS to be found [1], after a discussion on wikitech-l [2].
Some Chinese editors made petitions [3] (starting on 08/08) and Iranian users raised a similar problem [4] (on 14/08). In parallel these last two weeks there were discussions on wikitech-l about some way to opt-out by user and/or geographically. And in parallel the last two weeks there were discussions on wikitech-l whether some opt-out mechanism should be implemented with two opposed points of view: 1/ this security about the protection of the password must be for everyone else it is unuseful (which is true in a perfect world), no matter if China and other HTTPS-unlucky people cannot login (and hence must edit under IP or not edit); 2/ although security is very important, not to allow HTTP logins in China (and other HTTPS-unlucky people) will destroy etablished parts of the community and should be avoided, so implementation of work-arounds is needed. And this last discussion had not to be on wikitech-l because it is political, and was only a few raised elsewhere (where HTTPS is technical and should be discussed on wikitech-l.)
Finally some work-arounds were implemented; first it was a list of wikis where HTTP login will be allowed (this decision became public on Monday [5]) and yesterday (sic) it was announced a geolocalised solution [6]. Secondly there will be a preference for the users, although until yesterday it was not clear for everybody how exactly it was implemented. In parallel the central notice was set up two days ago with an English-only page, pywikipediabot was announced to be ready some hours ago. And in some hours there should be the deployment target.
[1] http://blog.wikimedia.org/2013/08/01/future-https-wikimedia-projects/ [2] http://lists.wikimedia.org/pipermail/wikitech-l/2013-July/070981.html [3] http://meta.wikimedia.org/wiki/Requests_for_comment/Petition_of_HTTPS_defaul... [4] https://bugzilla.wikimedia.org/show_bug.cgi?id=52846 [5] https://meta.wikimedia.org/wiki/HTTPS?diff=5731209&oldid=5728786 [6] http://lists.wikimedia.org/pipermail/wikitech-l/2013-August/071348.html
Conclusion:
I know the fact we now know we are spied is disturbing, but…
Why the hell HTTPS is so truly *urgent* we cannot spent more than three weeks (at all) to think about the problem, investigate related problems (including political and communitical here), think about solutions and user interfaces/interactions, implement solutions, widely avertize the problem and solutions, and peacefully deploy the patches?
I would have loved some RFC and some discussion elsewhere than on wikitech-l with structured problems and solutions, and more time allowed for discussing all that with the community -- because I guess it was widely discussed internally in technical and operations teams, but the community discovered these plans and had to report potential problems in a time frame of 3 weeks.
More generally speaking, I would love the WMF share more their internal plans long before rollout -- even if I concede writing and discussion is more time-consuming than oral speak and introduce latencies -- and probably in some digest and expanded forms (I know there are already both, it’s probably to be improved and perhaps more targeted to avoid everyone’s burnout). And perhaps slow the rhythm of the technical changes to have a more stable environment (I understand this is personal and there are other PoV).
Thanks, ~ Seb35
Le Wed, 21 Aug 2013 11:37:35 +0200, Pierre-Selim pierre-selim@huard.info a écrit:
First of all, I'm sorry If my tone was not appropriate (keep in mind I'm not a native speaker).
2013/8/21 Terry Chay tchay@wikimedia.org
On Aug 21, 2013, at 1:39 AM, Pierre-Selim pierre-selim@huard.info wrote:
Just a question: Why imposing HTTPS ? Really, it will be damaging
The reason why is outlined in Ryan's blog post as well as his previous post and the Wikipedia entry on https linked from that post.
The short answer is the current state is known to present a number of privacy and security vulnerabilities further emphasized by the now-known existence of software designed to deliberaty target these vulnerabilities in Wikipedia specifically.
https://blog.wikimedia.org/2013/08/01/future-https-wikimedia-projects/
I just think the user should be informed of this and should have the choice (so the user can make an enlightened choice). And that is mostly my point. All the explanation you have given are good, and the work of the WMF is good IMO..
Thank you for all the time you spent on this feature, however I'm not convinced at all.
Luckily, the standard for the Movement is consensus, not catering to every extremist view with 100% buy-in. The latter standard is impossible as people would be affected either way. The technical component is informing the decision and helps to hash out some of the details, but this is a case where parts of the Vision are being compromised today, and a different (hopefully better) compromise is being reached through this rollout.
Off course, I was just giving my opinion, I'm one user and do not represent more than that. We will see how it works out, and I would be happy to owe you a drink if everything goes smooth.
Take care,
terry
Thank you for your answer and have a nice roll out.
2013/8/21 Ryan Lane rlane@wikimedia.org
On Wed, Aug 21, 2013 at 4:38 AM, Brion Vibber bvibber@wikimedia.org wrote:
On Tue, Aug 20, 2013 at 1:33 PM, Nathan nawrich@gmail.com wrote:
Hi, context please?
Continuation of this thread from wikitech-l:
http://lists.wikimedia.org/pipermail/wikitech-l/2013-August/thread.html#7128...
tl;dr summary:
- ops plans to switch logins to HTTPS
- switching all logins to HTTPS is known to break access for
logged-in
users in countries where Wikimedia's HTTPS servers are blocked by government censorship
- there are some plans to mitigate this by excluding some languages
from
the requirement
- this is controversial for several reasons, one of which is that it
will
break access for users in those countries on language projects that
are
not
excepted (eg English Wikipedia in mainland China)
The last point isn't accurate. The original plan was to exempt
certain
languages from the login redirection, and those projects would be
"home"
wikis. When someone logged-in there, they'd also be logged-in
everywhere
else via central auth. The current plan is to disable the HTTPS
redirect
using geolocation for countries that have a > 5% error rate for HTTPS requests.
This discussion is technical, so I'm going to move back to
wikitech-l, now.
- Ryan
Wikimedia-l mailing list Wikimedia-l@lists.wikimedia.org Unsubscribe:
https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
-- Pierre-Selim _______________________________________________ Wikimedia-l mailing list Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe _______________________________________________ Wikimedia-l mailing list Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe