new TLDs and extended domain names.

[Roedy Green]

ICANN just voted to drastically open up what constitutes a legit TLD
(top level domain name). For example there may soon be a .ebay. I
gather that creating a new TLD will be not much more of a deal than
creating an new domain.

Also have a look at http://idn.icann.org/#Things_to_test
for some of the new sorts of names we will have to deal with in our
Java programs:

You can have Arabic, Chinese, Indian, Russian, Thai, Hebrew --
languages that don't use the Roman Alphabet, and read right to left.

It uses an unfamiliar encoding e.g.
href="//xn--fdbk5d8ap9b8a8d.xn--deba0ad/"

I wonder how you are supposed to key these into a browser when you
don't have a link to click.

 

[Arne Vajhøj]

ICANN just voted to drastically open up what constitutes a legit TLD
(top level domain name). For example there may soon be a .ebay. I
gather that creating a new TLD will be not much more of a deal than
creating an new domain.

Also have a look at http://idn.icann.org/#Things_to_test
for some of the new sorts of names we will have to deal with in our
Java programs:

You can have Arabic, Chinese, Indian, Russian, Thai, Hebrew --
languages that don't use the Roman Alphabet, and read right to left.

It uses an unfamiliar encoding e.g.
href="//xn--fdbk5d8ap9b8a8d.xn--deba0ad/"

I wonder how you are supposed to key these into a browser when you
don't have a link to click.

It is IDN.

(which Java BTW has supported since 1.6)

A good browser and operating system should allow the users to enter
the URL using their alphabet and then the software converted the URL
to IDN "behind the scene".

Arne

 

[Mark Space]

ICANN just voted to drastically open up what constitutes a legit TLD
(top level domain name). For example there may soon be a .ebay. I
gather that creating a new TLD will be not much more of a deal than
creating an new domain.

Oh look! A new wave of revenue for registrars and a new addition to the
list of FLDs to register when you register a .com FLD! How did that happen?

:rollseyes:

 

[Roedy Green]

A good browser and operating system should allow the users to enter
the URL using their alphabet and then the software converted the URL
to IDN "behind the scene".

I discovered that dropping http://xn--fdbk5d8ap9b8a8d.xn--deba0ad/
into Opera 9.5 takes you to an ICAN Hebrew page. It echoes the url in
Hebrew.

The web is pushing us toward tools that can handle documents and
programs with any combination of languages.

 

[Daniel Pitts]

I discovered that dropping http://xn--fdbk5d8ap9b8a8d.xn--deba0ad/
into Opera 9.5 takes you to an ICAN Hebrew page. It echoes the url in
Hebrew.

The web is pushing us toward tools that can handle documents and
programs with any combination of languages.

Firefox 3.0 also does so. Its pretty nifty if you ask me

 

[Roedy Green]

I discovered that dropping http://xn--fdbk5d8ap9b8a8d.xn--deba0ad/
into Opera 9.5 takes you to an ICAN Hebrew page. It echoes the url in
Hebrew.

I have learned a bit more. The Java encode/decode is java.net.IDN

RFC 3490 describes the algorithm.

Opera works and echos the URL in Hebrew. Firefox, Sea Monkey, Flock
and IE all work, but echo the URL as gibberish hex peppered with
percent signs.

 

[Roedy Green]

And Safari.

Unlike Firefox, Flock, Seamonkey and IE,
Safari echoes the URL both in the original encoding as in Hebrew.
Opera echoes the Hebrew.

I am in the middle of a giant upload, but when it finishes, my page on
what I have learned about IDN will be ready:

http://mindprod.com/jgloss/idn.html

 

[Roedy Green]

t's probably too much to hope that IE gets this right

same as firefox, at least IE 8 beta.

 

[Daniel Pitts]

I have learned a bit more. The Java encode/decode is java.net.IDN

RFC 3490 describes the algorithm.

Opera works and echos the URL in Hebrew. Firefox, Sea Monkey, Flock
and IE all work, but echo the URL as gibberish hex peppered with
percent signs.

Which version of firefox? Firefox 3 seems to do the same as opera.

 

[Joshua Cranmer]

No one can predict what you will finally see. And I don't know what you
mean about the "debating...hasn't made much sense". The nature of the
Internet makes it impossible that content can be controlled by domain
name. Inasmuch as there is a debate about trying to construct a system
for compartmentalizing content for easier filtering, it seems to me that
having the debate makes sense. The solutions, if they even exist, are
non-obvious.

Oh come on, you've forgotten the best solution: ban all images since
they might contain pr0n. While you're at it, ban the text too, since
people could be using ASCII art.

Heck, it's what my ISP did with Usenet.... /me goes off into a corner
and writes some angry letters.

 

[Roedy Green]

RFC 3490 describes the algorithm.

There are now at least 3-different URL-encoding schemes. There should
be some mechanism to retire formats and gradually corral everyone to
the latest and greatest.

 

[David Segall]

have a look at http://idn.icann.org/#Things_to_test
for some of the new sorts of names

OK, I went there. How can I tell if my browser is displaying the sites
correctly. "It's all Greek to me".

 

[Andreas Leitgeb]

OK, I went there. How can I tell if my browser is displaying the sites
correctly. "It's all Greek to me".

Άλλα μόνο ένα άπο αυτά ήταν Ελληνικά!
(But only one of them was Greek!

 

[Roedy Green]

OK, I went there. How can I tell if my browser is displaying the sites
correctly. "It's all Greek to me".

see http://mindprod.com/jgloss/unicode.html
to learn what the various glyphs look like for various languages.

 

[Christian]

ICANN just voted to drastically open up what constitutes a legit TLD
(top level domain name). For example there may soon be a .ebay. I
gather that creating a new TLD will be not much more of a deal than
creating an new domain.

Also have a look at http://idn.icann.org/#Things_to_test
for some of the new sorts of names we will have to deal with in our
Java programs:

You can have Arabic, Chinese, Indian, Russian, Thai, Hebrew --
languages that don't use the Roman Alphabet, and read right to left.

It uses an unfamiliar encoding e.g.
href="//xn--fdbk5d8ap9b8a8d.xn--deba0ad/"

I wonder how you are supposed to key these into a browser when you
don't have a link to click.

I must say that I think that these new Unicode domains will give
developers and the rest of the world a lot of headaches in future.

Its nothing to unite the internet and give us all equal chances.
The main thing that will come is that you will now have
tens of diffrent unicode characters that look like an "a".

What this means for me is that several fraudulent sites will open up on
domains looking exactly like ebay.com/paypal.com but actually have
some obscure unicode character that is just rendered like an "a".

And as we see this will happen .. firefox doesn't show the punycode in
the address line but the real unicode.

This introduction of unicode will give us lots of woe in the future.

sry about being pessimistic about this
Christian

 

[Roedy Green]

What this means for me is that several fraudulent sites will open up on
domains looking exactly like ebay.com/paypal.com but actually have
some obscure unicode character that is just rendered like an "a".

I saw a few sentences on the problem. I think they hope to avoid that
by vetting cheating names.

 

[Martin Gregorie]

The .sex and .xxx domains would work for much of the content. They are
advantageous to the purveyors of smut (no offense intended to said
purveyors) and advantageous to parents trying to keep their kids out of
it. So for a voluntary system, the kind of system we really want, these
domain names represent a real step forward.

Of course this is my opinion only.

Sounds like a reasonable idea to me. Its really odd why the US
Dept of Commerce, apparently prompted by a bunch of bone-headed so-called
Christians, unilaterally killed off the .xxx TLD when it was proposed a
year or two back.

 

[Roedy Green]

Top-level domains will cost six figures, and will be subjected to a review
process of the name itself as well as of the organization who wants to own
it to ensure that they actually have the resources to manage a TLD.

the CBC said no prices have been announced, but that experts estimate
it will be in the order of $100,000.

I could see names of the form cityname.hotel would be snapped up.

 

[Roland de Ruiter]

ICANN just voted to drastically open up what constitutes a legit TLD
(top level domain name). For example there may soon be a .ebay. I
gather that creating a new TLD will be not much more of a deal than
creating an new domain.

I think this subject is off topic for c.l.j.programmer.

[...]
It uses an unfamiliar encoding e.g.
href="//xn--fdbk5d8ap9b8a8d.xn--deba0ad/"

That's puny code: <http://en.wikipedia.org/wiki/Punycode>

 

[Roedy Green]

I think this subject is off topic for c.l.j.programmer.

I defend my post on the following grounds.

1. it is news relevant to all Java programmers.

2. there is very little traffic in comp.lang.java.*. It will die
completely if there is not a steady trickle of interesting bits to
discuss. At this point we should be encouraging discussion, not trying
to stomp it out. There was a time when overwhelm was the problem, but
not today. We have a problem with Miss Grundy's lecturing the newbies
severely on every tiny infraction of their personal rules, scaring
them off. That is no way to encourage the spread of Java. That
bullying is the where we have to focus energy stopping.

3.. The Java-specific question is, what does this mean for Java coding
and existing coding?

4. The key feature of Java is multi-platform, multi-national,
multi-lingual. . That is Java's niche. The new domain encoding is a
piece of information crucial to that core functionality.

5. If I didn't mention it here, it is highly unlikely the majority of
Java coders would stumble on it elsewhere within the next month or so.
I am speaking to a community. Spamming is telling them things of no
relevance. That is not what I did.

6. There is so much true spam on the newsgroups, it seems funny to
make MORE fuss about something in the grey area.