On 12/12/2017 18:19, Ryan Sleevi wrote:
On Tue, Dec 12, 2017 at 8:36 AM, Jakob Bohm via dev-security-policy <
[email protected]> wrote:

On 12/12/2017 01:08, Adam Caudill wrote:

Even if it is, someone filed the paperwork.  Court houses have clerks,
guards, video cameras, etc...  It still may present a real physical

point

from which to bootstrap an investigation.


Court houses also have online systems. I think if you read both Ian and

James' work, you'll see the issues they're raising address this
hypothetical.


I shall certainly read their work closely on that matter.  In my

experience, these generally don't allow filings for new businesses from
those not previously known to the court/registrar in real life.

I can say from my own experience, in some states in the US, it's a trivial
matter to create a company online, with no validation of identity or other
information. It takes about 10 minutes, and you'll have all the paperwork
the next day. When I did this (in a state I had never done business in
before), there was absolutely no identity checks, no identity documents,
nothing at all that would tie the business to me if I had lied. Creating a
business with no connection to the people behind it is a very, very simple
thing to do.


A lot of people have posed suggestions for countermeasures so extreme
they should not be taken seriously.  This includes discontinuing EV,
requiring that companies cannot get EV certs during their first year of
existence, or suggesting that only "famous" companies can get EV
certificates.

Here is a more reasonable suggestion:

1. In the Fx UI, display the actual jurisdictionOfIncorporation instead
   of just the country, especially where those differ (For example
   Kentucky versus all-of-US).


I will note that suggestions of this nature are not particularly productive
or in line with usability research. Beyond the simple fact that "UI is not
a democracy" - it's a product decision - let's look at what we're saying:

Users are expected to know and understand these distinguishing factors.
Either they're expected to understand that the "Stripe, Inc" they want to
use is not in Kentucky (an external piece of knowledge few have) or that
the "Stripe, Inc" they normally deal with is something they have to
memorize (in addition to the URL and other distinguishing factors)

As such, this does not stand the basic "Is this a reasonable thing to ask
all users" test.


2. Add a rule that if there is a big national or international company
   with a name, other companies cannot get certificates for the same
   name in related jurisdictions.  For example if there is a company
   listed on NYSE or NASDAQ, no similarly named US company can get an
   EV or OV certificate for that name.  Ditto for a reasonable list of
   national registries in each country.  CAs should be required to
   publicly state which "big-status" lists beat local
   company/organization registrations in each country, and similar for
   any special lists of major global organizations, such as Google or
   The Red Cross.


I realize your intent is well, but this is a huge and tremendous challenge.
If you don't believe me, understand why
http://www.wipo.int/portal/en/index.html exists.

There is not a 'reasonable list' here nor is there a good definition of
'big national or international company'. Or look at
http://www.wipo.int/amc/en/domains/decisions/html/2000/d2000-1015.html

5. It is worth noting that some countries do use a national company
   registry which ensures uniqueness directly.  Denmark is one such
   country (though the uniqueness checking is probably limited to exact
   matches).


There are a number of such countries.

I think it's great you're engaging, but it's not for lack of consideration
of your proposals that the result is to remove EV. Rather, it's because the
impracticality of them - as already demonstrated - that they shouldn't be
on the table.


The overall thing is that the current thread seems to be a major case of throwing the baby out with the bathwater.

The entire problem boils down to this:

  Some jurisdictions (such as Kentucky) allow people to set up
  businesses with official legal names that are significantly
  misleading.  This failure by those governments naturally allows such
  misleading names to appear in documents (such as X.509 certificates)
  that are visible far outside those jurisdictions.

  Insisting that no businesses at all get to show proof of their legal
  names to Internet users just because some businesses have been allowed
  by their governments to hold misleading names, is doing significantly
  more damage to the overall community of Internet users than the
  current risk of such government-authorized situations being used to
  deceive users.

The simple solution, seemingly rejected by some here, is to state "This
is a vulnerability in the relevant government entity (such as whatever
Kentucky institution allowed the name), it is not misissuance to put
such official legal names in EV certificates". If this would be accepted, no changes are needed in the EV rules, case closed.

As an alternative to this simple solution, I have therefore been
suggesting *minimal* restrictions on EV certificates that could mitigate
the problem while still allowing most businesses to get EV certificates
for their non-misleading names.

And since these would be restrictions only for EV certificates, rather
than restrictions on having those names at all, the community can
reasonably afford to impose rules more strict than what governments,
trademark authorities etc. can practically impose.

My "lists of big companies" would be much more ad-hoc and much less
formal than what is used for e.g. trademark enforcement.  The intention
being that there will be a small proportion of businesses (less than
0.5%) that will be unable to get EV certificates for their trademarked
and government registered names due to conflicts with other businesses
by the same name.  Sometimes this will be mutual (two similarly named
companies can't get EV certificates because of each other), at other
times one entity gets the opportunity semi-arbitrarily (based on lists
that are not arbitrary CA or Browser decisions, but were arbitrarily
chosen without directly targeting those businesses).

The "Spring Inc" not-of-Kentucky versus "Spring Inc" (Kentucky) case is
the one that would be imperfectly captured by the lists of US-wide and
world-wide companies.  There is no certainty that "Spring Inc" will
be on those lists at all relevant times, though it should be predictable based on things at least partially in the companies control (such as if they are on the "Fortune 500 list" and if they are publicly traded on
their big national stock exchange(s)).

The "display the jurisdictionOfIncorporation" rule captures the more
common case of "First National Bank" (Kentucky) vs. "First National
Bank" (Arizona), where neither has any reason to be considered the only
"First National Bank" on a wider scale, but the user has a reasonable
expectation to connect to the one in a specific location.

Making it attractive for such locally operating businesses to register
in their actual location rather than in a "haven" such as Delaware or
Luxemburg is good on a society level.  It is similar to how passengers
on Caribean cruise-liners may have a slight preference for ships that
are labeled "M/S Example, Fort Lauderdale" (with a US flag) rather than
"M/S Example, Freetown" (with a Liberian flag).


Enjoy

Jakob
--
Jakob Bohm, CIO, Partner, WiseMo A/S.  https://www.wisemo.com
Transformervej 29, 2860 Søborg, Denmark.  Direct +45 31 13 16 10
This public discussion message is non-binding and may contain errors.
WiseMo - Remote Service Management for PCs, Phones and Embedded
_______________________________________________
dev-security-policy mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-security-policy

Reply via email to