interoperability testing

[Eric Brunner-Williams]

A side-conversation returns to the main IDNAbis list ...

Erik van der Poel wrote:
> Yes, it's OK with me if Eric would like to take any of my comments
> back to this list.
>
> Erik
>
> On Tue, Jul 14, 2009 at 1:13 PM, Shawn Steele<Shawn.Steele at microsoft.com> wrote:
>   
>> Can we take this back to the list?
>>
>>
>>
>> -Shawn
>>
>> -----Original Message-----
>> From: Erik van der Poel [mailto:erikv at google.com]
>> Sent: Tuesday, July 14,  2009 9:57
>> To: Eric Brunner-Williams
>> Cc: Shawn Steele; Gervase Markham
>> Subject: Re: interoperability testing
>>
>> Hi Eric,
>>
>> Yes, that is a bit clearer. Depending on how much detail we want, it
>> could be quite a lot of work for each app developer to come up with an
>> "interoperability story" that is comprehensive.
>>
>> For browsers, I believe we should consider a couple of different
>> inputs. One is HTML, specifically the domain names inside URLs inside
>> tags like <a href="...">, <img src="..."> and so on. Given this input,
>> a browser emits a couple of outputs, namely, DNS packets and HTTP
>> packets. To achieve interoperability, therefore, all browsers (MSIE,
>> Firefox, etc) must produce the same output for all inputs.
>>
>> Another type of browser input is the URL text field where users can
>> type a URL on the keyboard. I believe most browsers use the same IDNA
>> API for this text field as they do for URLs in HTML. In theory,
>> browsers could do special things for certain users, such as the 4
>> different versions of the letter i in Turkish. It would be nice if a
>> Turkish user could type the same thing in MSIE and in Firefox to get a
>> particular result. So this would be a different kind of
>> interoperability.
>>
>> Email apps may well use the same APIs as browsers do, either directly
>> or indirectly via HTML. The same could be said of IM and other apps.
>> So, sometimes the availability of an API determines what each app will
>> end up doing.
>>
>> Erik
>>
>> On Tue, Jul 14, 2009 at 9:15 AM, Eric
>> Brunner-Williams<ebw at abenaki.wabanaki.net> wrote:
>>     
>>> Shawn, Gerv,
>>>
>>> First, I think the thread should stay on the main IDNAbis mailing list.
>>> Second, I'd like to make the references to "interoperability" a little more
>>> specific than they have been in the past. If I'm happy, as a registry
>>> operator, to say "Oh, that works for me, I don't care about ...", and each
>>> of us is happy, as a <insert application here>, to say "Oh, that works for
>>> me, I don't care about ...", we don't actually end up knowing if each of our
>>> happy-making assumptions and use cases for which a sufficiency of something
>>> works, are all mutually intelligible and consistent across all use cases and
>>> applications.
>>>
>>> We should not end up with "interoperability for DIGITSbis" meaning I'm happy
>>> with primes, Gerv is happy with non-negative evens, Erik is partial to zero,
>>> and Shawn is good with non-positive odds.
>>>
>>> I'm not asking for "how do you test X", I'm asking for how does Application
>>> A, using X, interoperate with Application B, where A and B could be
>>> different versions of the same application, or as different as XMPP and
>>> SMTP. If both consume X, but there is no way to meaningfully "interoperate"
>>> an instance of A and an instance of B (worst case A==B==/dev/null, hard to
>>> have a sink talk to a sink), then we shouldn't talk about "interoperate" in
>>> the context of A and B, but rather something about the determinism (or its
>>> lack) of A and B as state machines when processing the same stream of X
>>> inputs.
>>>
>>> Clearer?
>>>
>>> Eric
>>>
>>>
>>> Shawn Steele wrote:
>>>       
>>>> I missed this, but I don’t know what question Eric's asking :)
>>>>
>>>> The discussion makes me think of at least 3 possibilities:
>>>>
>>>> 1) IDNA2003/2008
>>>> 2) API level behavior consistent?  (Are similar APIs available on all
>>>> platforms)
>>>> 3) Does it do the same thing in Firefox as in IE?
>>>> 4) Does the link go where it's "supposed to"?
>>>>
>>>> -Shawn
>>>>
>>>> -----Original Message-----
>>>> From: Gervase Markham [mailto:gerv at mozilla.org]
>>>> Sent: Monday, July 13,  2009 10:48
>>>> To: Eric Brunner-Williams
>>>> Cc: Erik van der Poel; Shawn Steele
>>>> Subject: Re: interoperability testing
>>>>
>>>> On 11/07/09 23:24, Eric Brunner-Williams wrote:
>>>>
>>>>         
>>>>> If you, or Shawn, or Gerv, or Pete, could state what the
>>>>> interoperability story is for your apps, something that doesn't end up
>>>>> with "and then there was phishing, badness and thrashing", it would help
>>>>> me at least.
>>>>>
>>>>>           
>>>> [off-list] Are you asking about how our apps will manage the
>>>> interoperability of IDNA2003 and IDNA2008? If so, I have no idea, as a)
>>>> I haven't yet got my head around the current proposals, as you may have
>>>> noticed and b) even if I had, it wouldn't be me implementing them.
>>>>
>>>> I'm attempting to summon reinforcements.
>>>>
>>>> Gerv
>>>>
>>>>
>>>>         
>>>
>>>       
>>     
>
>
>