During recent calls, there’s been a desire for some UA Use Cases. This from a document we had developed some time ago. I would welcome people’s suggestions for which of these should be tested - as there’s a cost associated with testing each one. Don Universal Acceptance Use Cases for UA Readiness Evaluation Purpose: To provide Use Cases for developers to use when evaluating the UA Readiness of their systems. This list will provide generic examples as well as specific names that can be used for testing. Background: Background on Universal Acceptance to go here Domain Names: Examples • secondlevel.toplevel • 域名.toplevel • secondlevel.域名 • 域名.域名 • secondlevel.punycode • punycode.toplevel • punycode.punycode • Second. חלק עליון(Right-to-left TLD) • שני. חלק עליון (Right-to-left second level and top-level) Path Cases (IRIs) • secondlevel.toplevel/unicode • 域名.toplevel/unicode • secondlevel.域名/unicode • 域名.域名/Unicode • secondlevel.punycode/unicode • punycode.toplevel/unicode • punycode.punycode/unicode • Second. חלק עליון/Unicode (Right-to-left TLD) • שני. חלק עליון /Unicode (Right-to-left second level and top-level) E-mail Addresses: • localpart@domain.tld<mailto:localpart@domain.tld> • localpart@域名.tld • localpart @domain.域名 • localpart@域名.域名<mailto:localpart@域名.域名> • localpart@domain.punycode<mailto:localpart@domain.punycode> • localpart@punycode.tld<mailto:localpart@punycode.tld> • localpart@punycode.punycode<mailto:localpart@punycode.punycode> • 域名@域名.域名 • שני. חלק עליון@localpart
I haven’t seen any replies yet but as I think it is an important topic I would like to chime in. In theory all of the cases mentioned below should be tested but as you indicated this is cost prohibitive. Would it be an idea to split this list into tiers ? Ie a minimal set of tests to make it workable but not ideal and more advanced tests that have the full solution. In the minimal set there would be no IDN, but full support for punycode (a-labels) The advanced set would allow for IDN everywhere and left-right/right-left cases. Ron Geens DNS Belgium On 17 Dec 2015, at 01:15, Don Hollander <don.hollander@icann.org<mailto:don.hollander@icann.org>> wrote: During recent calls, there’s been a desire for some UA Use Cases. This from a document we had developed some time ago. I would welcome people’s suggestions for which of these should be tested - as there’s a cost associated with testing each one. Don Universal Acceptance Use Cases for UA Readiness Evaluation Purpose: To provide Use Cases for developers to use when evaluating the UA Readiness of their systems. This list will provide generic examples as well as specific names that can be used for testing. Background: Background on Universal Acceptance to go here Domain Names: Examples • secondlevel.toplevel • 域名.toplevel • secondlevel.域名 • 域名.域名 • secondlevel.punycode • punycode.toplevel • punycode.punycode • Second. חלק עליון(Right-to-left TLD) • שני. חלק עליון (Right-to-left second level and top-level) Path Cases (IRIs) • secondlevel.toplevel/unicode • 域名.toplevel/unicode • secondlevel.域名/unicode • 域名.域名/Unicode • secondlevel.punycode/unicode • punycode.toplevel/unicode • punycode.punycode/unicode • Second. חלק עליון/Unicode (Right-to-left TLD) • שני. חלק עליון /Unicode (Right-to-left second level and top-level) E-mail Addresses: • localpart@domain.tld<mailto:localpart@domain.tld> • localpart@域名.tld • localpart @domain.域名 • localpart@域名.域名<mailto:localpart@xn--eqrt2g.xn--eqrt2g> • localpart@domain.punycode<mailto:localpart@domain.punycode> • localpart@punycode.tld<mailto:localpart@punycode.tld> • localpart@punycode.punycode<mailto:localpart@punycode.punycode> • 域名@域名.域名 • שני. חלק עליון@localpart
Em 30 de dez de 2015, à(s) 09:28:000, Ronald Geens <ronald.geens@dnsbelgium.be> escreveu:
I haven’t seen any replies yet but as I think it is an important topic I would like to chime in.
In theory all of the cases mentioned below should be tested but as you indicated this is cost prohibitive.
I disagree that cost is linked to the number of test cases. Test processes today are usually automated, so lifecycle management is usually the main cost driver, no matter testing 10 or 20 use cases. Rubens
Rubens, I fully agree that test processes should indeed be fully automated … in reality there are still a lot of cases, especially in front-end and website development I am afraid, where this is not the case. Anyhow, it is not only the testing cost that will determine which usecases to support for a certain development. Splitting the usecases into two levels would already allow a working solution at minimal level without having "to bother" with IDN development and testing. This is of course a very much ASCII-world driven viewpoint which is not necessarily in line with real Universal Acceptance but using the agile principle of “good enough” this still has value according to me. Ron Geens DNS Belgium
On 30 Dec 2015, at 13:09, Rubens Kuhl <rubensk@nic.br> wrote:
Em 30 de dez de 2015, à(s) 09:28:000, Ronald Geens <ronald.geens@dnsbelgium.be> escreveu:
I haven’t seen any replies yet but as I think it is an important topic I would like to chime in.
In theory all of the cases mentioned below should be tested but as you indicated this is cost prohibitive.
I disagree that cost is linked to the number of test cases. Test processes today are usually automated, so lifecycle management is usually the main cost driver, no matter testing 10 or 20 use cases.
Rubens
It's true that the test cases can be split into an ASCII portion and an IDN portion, and our guidance can show people how to start with one and work to the other. But we should be very careful not to imply that ASCII-only is "good enough". -----Original Message----- From: ua-discuss-bounces@icann.org [mailto:ua-discuss-bounces@icann.org] On Behalf Of Ronald Geens Sent: Wednesday, December 30, 2015 6:15 AM To: UA-discuss@icann.org Subject: Re: [UA-discuss] Use Cases Rubens, I fully agree that test processes should indeed be fully automated … in reality there are still a lot of cases, especially in front-end and website development I am afraid, where this is not the case. Anyhow, it is not only the testing cost that will determine which usecases to support for a certain development. Splitting the usecases into two levels would already allow a working solution at minimal level without having "to bother" with IDN development and testing. This is of course a very much ASCII-world driven viewpoint which is not necessarily in line with real Universal Acceptance but using the agile principle of “good enough” this still has value according to me. Ron Geens DNS Belgium
On 30 Dec 2015, at 13:09, Rubens Kuhl <rubensk@nic.br> wrote:
Em 30 de dez de 2015, à(s) 09:28:000, Ronald Geens <ronald.geens@dnsbelgium.be> escreveu:
I haven’t seen any replies yet but as I think it is an important topic I would like to chime in.
In theory all of the cases mentioned below should be tested but as you indicated this is cost prohibitive.
I disagree that cost is linked to the number of test cases. Test processes today are usually automated, so lifecycle management is usually the main cost driver, no matter testing 10 or 20 use cases.
Rubens
participants (4)
-
Don Hollander -
Mark Svancarek -
Ronald Geens -
Rubens Kuhl