Do we need new dns system based on UTF-16?

Question

0.00/5 (No votes)

See more:

I have written own dns system based on utf-16 codepage names.
Server runs windows service, that answers client requests.
Client sends request "GET /national.words.in.utf-16.codepage STL 1.0" using utf-16 codepage.
Server looks for dns name "national.words.in.utf-16.codepage" in local file and gives answer "::1" for ip 6 or "127.0.0.1" for ip 4.

Here "STL" is my company name, "1.0" is protocol version.

I consider all of you to be The Committee on Telecommunications and the developers of browsers.

Now i ask you: "Do we need such system?"

Copied to http://www.codeproject.com/Messages/3829076/Do-we-need-new-dns-system-based-on-UTF-16.aspx[^]

Posted 25-Mar-11 3:40am

Kozlov_Sergey

Updated 25-Mar-11 6:55am

v5

Add a Solution

3 solutions

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Emilio Garavaglia · Answer 1 · 2011-03-25T21:44:00

Solution 3

It seems there are some misconceptions about what DNS is and is for
DNS is not "for browers" or for "this or that app".
Is a communication protocol used to map an "application level address" into a "network level address".
The form of such "addresses" must be the same in all DNS system around the universe.

The way an host-name is represented in a DNS string was ASCII, and had been extended to include all Unicode sets by means of an encoding named "punycode[^]". And that's how DNS strings are supposed to be transmitted around the world, and -hopfully- stored in zone files.

Whatever your opinion is about the efficiency of that encoding, changing it requires a change in all DNS around the world (nobody can actually know how many of them are) in a way that the new coding does not disrupt the existing ones.

Posted 25-Mar-11 21:44pm

Emilio Garavaglia

Comments

[no name] 29-Mar-11 2:35am

I think, that this answer means "Yes, we need new protocol."
So, how can i make this protocol to be used world wide?

Emilio Garavaglia 29-Mar-11 14:58pm

If you think you have a good idea, goto http://www.rfc-editor.org/ and publish an RFC describing your protocol (that is: how different system should interact to resolve domain name according to your specification and how you manage the compatibility with what existent). At that point, if at least other two independent developers develope something different that interoparate properly with your implementation your RFC will become an "Internet draft". The adoption will depend of the wish of the other developers in the Internet community.

[no name] 3-Apr-11 1:11am

How can i add my rfc document to http://www.rfc-editor.org/ ?

Emilio Garavaglia 3-Apr-11 14:06pm

By reading the site submission guideline and style guides.

#realJSOP · Answer 2 · 2011-03-25T03:48:00

Solution 1

The Q/A section isn't really intended for "opinion" threads. Find an appropriate forum. I think even the Lounge would be more appropriate than posting it in Q/A.

Posted 25-Mar-11 3:48am

#realJSOP

Comments

Sergey Alexandrovich Kryukov 25-Mar-11 13:38pm

John, it's a bit late to discuss on this level. Unicode names are already introduced in Russia, not sure about Belorus. There is a lot of criticism (which I kind of share), but this is about how the standard was conceived.
--SA

Sergey Alexandrovich Kryukov · Answer 3 · 2011-03-25T07:36:00

Solution 2

Sergey.
You should be aware of the fact that Unicode domain names are already introduced in Russia.
Did you read this: http://en.wikipedia.org/wiki/Internationalized_domain_name[^]?

—SA

Posted 25-Mar-11 7:36am

Sergey Alexandrovich Kryukov

Comments

[no name] 25-Mar-11 16:29pm

I know, that existing system based on utf-8. But i think, that new system based on utf-16 is better.
Any way i will use new systemin my on network.

Emilio Garavaglia 29-Mar-11 15:00pm

Nope, is based on punycode. That's not UTF8

Sergey Alexandrovich Kryukov 29-Mar-11 15:14pm

Thank is correct, it is not UTF-8. The point of criticism is that punycode makes underlying Unicode name pretty much unrecognizable. The initial intent did make some sense, but what was done appeared to be... as always :-)
--SA

Emilio Garavaglia 29-Mar-11 15:34pm

Agree. But if I have to choose from UTF8 and 16 ... consider endianess and related issues, I would prefer UTF8.

But consider also that punycode was necessary because there are systems around the world, that - according with the preexistent standard - pretend an URL to have only alphanumerical [0..9,A..Z,-] characters.
I understand punycode cannot be easy to read. But UTFx cannot be processed by many old system. There are names out there that are registered from more than 20 years, that service providers must still grant (until someone will pay for, or until certain names are required to exist from the authorities)

That's why I'm not of the idea to change the protocol (let the string go from DNS servers and clients as punicode), but to make some API available to the OS so that you can invoke DNS passing a UTF string (something like gethostbyUTF8name(const char* str), or gethostbyUTF16name(const wchar_t* str) ) letting the OS to convert to punycode before sending it on net.

Sergey Alexandrovich Kryukov 29-Mar-11 16:04pm

I mainly agree with all what you say here.
Your preference of UTF-8 is well motivated, but there are more reasons to prefer it to other UTFs, for example, it is less sensitive to data loss (did you know that? imaging a data loss with a sh*t by 1 byte... UTF-8 will be suffered only locally).

Apparently, your explanation about legacy is perfectly valid. I still thing punicode is ill-conceived. This is just yet another case of weak standard, such as... do you know how many inconsistencies are there in XML standard, for example. People just used to it...

--SA