Is anyone interested in $15 backorders for IDNs?

[jacksonm]

Originally Posted by ThisDot

ASCII domains are currently available. I am going to get everything IDN compatible and I will let everyone know when it is operational for IDNs.

IDNs _are_ ascii domains. DNS does not function with other than ascii domain names.

The only place where IDNs are translated to the representation of multi-byte characters is locally on your computer, via the use of idn libraries (libidn for linux, similar for windows). When you type pää.com into your web browser, the browser will convert it to "xn--p-0faa.com" before sending the name over the wire to your DNS server to do an IP address lookup.

For all intents and purposes, there is zero difference between catching foobar.com and xn--wba.com.

Getting "set up" for IDN catching means nothing more than figuring out the list of prime words which you want to keep for yourself... Honestly...

.

[Fka200]

Originally Posted by jacksonm

Getting "set up" for IDN catching means nothing more than figuring out the list of prime words which you want to keep for yourself... Honestly...

.

Exactly.

[Rubber Duck]

Wrong, there is the little matter of setting up the drop registration software to input the Language Tag into the Registry, otherwise the whole thing goes belly up.

Originally Posted by jacksonm

IDNs _are_ ascii domains. DNS does not function with other than ascii domain names.

The only place where IDNs are translated to the representation of multi-byte characters is locally on your computer, via the use of idn libraries (libidn for linux, similar for windows). When you type pää.com into your web browser, the browser will convert it to "xn--p-0faa.com" before sending the name over the wire to your DNS server to do an IP address lookup.

For all intents and purposes, there is zero difference between catching foobar.com and xn--wba.com.

Getting "set up" for IDN catching means nothing more than figuring out the list of prime words which you want to keep for yourself... Honestly...

.

[jacksonm]

Originally Posted by Rubber Duck

Wrong, there is the little matter of setting up the drop registration software to input the Language Tag into the Registry, otherwise the whole thing goes belly up.

I hear you, but I really doubt this matters. I believe that there are hundreds of thousands of xn-- domains registered without any language tags, they are still unique, and they will still show the same characters in your browser's address bar.

.

[Rubber Duck]

Originally Posted by jacksonm

I hear you, but I really doubt this matters. I believe that there are hundreds of thousands of xn-- domains registered without any language tags, they are still unique, and they will still show the same characters in your browser's address bar.

.

Look I have sort of been through this process. Trust me you will need for the script to input the language tag.

[jacksonm]

Originally Posted by Rubber Duck

Look I have sort of been through this process. Trust me you will need for the script to input the language tag.

Yes, I agree that a tag should be there. But I'm saying that it doesn't matter which tag you input. The language tag is completely useless. There are no punycode collisions. The browser does not query the language tag from a registry to determine which script to render. You can register chinese punycode with an arabic tag, and it still renders as chinese characters in your browser...

Or can you explain in concrete detail why you believe the language tag matters? I'm interested in hearing.

.

[Alphamale]

Originally Posted by jacksonm

Yes, I agree that a tag should be there. But I'm saying that it doesn't matter which tag you input. The language tag is completely useless. There are no punycode collisions. The browser does not query the language tag from a registry to determine which script to render. You can register chinese punycode with an arabic tag, and it still renders as chinese characters in your browser...

Or can you explain in concrete detail why you believe the language tag matters? I'm interested in hearing.

.

i think i registered all of my idns at Domainsite with the Afrikaans tag - i think thats the default.

[jacksonm]

Originally Posted by Alphamale

i think i registered all of my idns at Domainsite with the Afrikaans tag - i think thats the default.

Same here. When I'm not lazy, I'll drop down the list, but for the most part, Afrikaans...

.

[Rubber Duck]

I agree that the registry will not generally distinguish between languages, but it does check things like mixed Latin and Cyrillic Scripts, and there then there is the complexity of the Chinese Block. But you are correct in general it makes no difference which code you input, as long as you do input one.

ASCII scripts don't input any language tag. For ASCII domains you don't need to do this, there is effectively no field to fill in. With IDN it is different. If your script doesn't fill in the field then the registration fails. It may well be that you can input random rubbish and get away with it most of the time. If you input nothing, that is precisely what will happen.

Don't ask me how to code it. I don't know. I worked with Dan Ruby when he was converting his script for IDN last year. In the end we basically abandoned the whole thing as we got out gunned, even after upping the anti a couple of times.
Eventually Snap came in and the game was up. Fighting them is a bit like squaring up in the street, stubbing your cigar and throwing your pancho over your shoulder to reveal your pistols only to be lassoed from behind and dragged down the street. Basically it not about drops it about contractual tie ups with the registrars.

Originally Posted by Alphamale

i think i registered all of my idns at Domainsite with the Afrikaans tag - i think thats the default.

The Tag is not recorded, but it is essential for the process.

All registered domains have no tag as far as I am aware, but you cannot register one without a valid tag.

[blastfromthepast]

Originally Posted by ThisDot

ASCII domains are currently available. I am going to get everything IDN compatible and I will let everyone know when it is operational for IDNs.

From a technical point of view, IDNs are ascii domains so I don't see what your problem is.

[Rubber Duck]

Originally Posted by blastfromthepast

From a technical point of view, IDNs are ascii domains so I don't see what your problem is.

His problem is that his script won't snag them because it is not putting in the language tag.

There is a technical difference between ASCII and IDN. The xn-- hack flags to the Registry that they are distinct. Only registrars that approved for IDN can sell IDN. Go to a non-approved registry and you should not be able to register them even as punycode.

[ThisDot]

Originally Posted by Rubber Duck

Wrong, there is the little matter of setting up the drop registration software to input the Language Tag into the Registry, otherwise the whole thing goes belly up.

That is correct Duck, and if I am going to be registering domains for other users, I would prefer to enter the correct language tag and not just take the lazy way out. It should be implemented very shortly as there are just a few changes that need to be made to get this right and not interfere with the standard ascii domains that I am currently catching. I will post here when it is all setup and operational.

Thanks,
David

[jacksonm]

Originally Posted by ThisDot

That is correct Duck, and if I am going to be registering domains for other users, I would prefer to enter the correct language tag and not just take the lazy way out. It should be implemented very shortly as there are just a few changes that need to be made to get this right and not interfere with the standard ascii domains that I am currently catching. I will post here when it is all setup and operational.

Thanks,
David

This is impossible. Japanese use chinese characters. Persian use arabic characters. On and on. There is no reliable way to detect the language a name belongs to.

.

[blastfromthepast]

Originally Posted by jacksonm

This is impossible. Japanese use chinese characters. Persian use arabic characters. On and on. There is no reliable way to detect the language a name belongs to.

But it is not necessary to be so precise, as long as you get a valid tag to register with.

[Rubber Duck]

Originally Posted by jacksonm

This is impossible. Japanese use chinese characters. Persian use arabic characters. On and on. There is no reliable way to detect the language a name belongs to.

.

It doesn't. All it does is check that the domain conforms to rules set for the selected language. Each language is limited to sets of Unicode characters. ICANN have the facility to make these more or less restrictive according to their needs. There is a test run at registration to see if the domain complies with the rules for that the requested language, if it passes the name is registered. The tag that was used to check conformance is not recorded.

Originally Posted by blastfromthepast

But it is not necessary to be so precise, as long as you get a valid tag to register with.

That is correct. If you put Arabic in as Persian it would probably pass, but not necessarily the other way around.

I had problems with Azeri, so I registered them as Turkish which uses the same scripts.

[ThisDot]

Originally Posted by jacksonm

This is impossible. Japanese use chinese characters. Persian use arabic characters. On and on. There is no reliable way to detect the language a name belongs to.

.

That is probably one very good reason that the language tag is required.

[Alphamale]

Originally Posted by ThisDot

That is probably one very good reason that the language tag is required.

that is certainly why it was devised, but it was soon realised to be imperfect and redundant.

yes it is required, but it serves no purpose. It's a historical hangup

[mulligan]

As long as you can recognise which language it is when you load it up and apply the relevant tag then you should be good to go. I doubt you will need every language tag anyway as most IDNs people are looking for fall into a fairly small group of languages

[Rubber Duck]

Originally Posted by Alphamale

that is certainly why it was devised, but it was soon realised to be imperfect and redundant.

yes it is required, but it serves no purpose. It's a historical hangup

Sorry but that is simply not true.

Language tags are used to detemine which mixtures of scripts are permitted and which are not. If you put in Russian, all Latin characters are automatically blocked. Cyrillics are blocked for all Latin Languages. At one time Latin was blocked for Japanese but that has been corrected at Verisign. I think there may still be an outstanding issue with Thai, where Latin was also blocked for no apparent reason.