15616 - Korean related links incorrect

Status: VERIFIED FIXED

(Core :: Internationalization, defect, P3)

Description

[Allan Masri]

with MacOS-J1 8.5, M10(this bug may happen on other platforms but testing is
blocked):

1. Set Charset in View menu to EUC-KR
2. Load www.naver.com (or www.yahoo.co.kr)
3. Open "What's related" folder.

RESULT:

Some links in "What's related" are displayed correctly. Others are not.

Comment 1

[msanz]

Should this be assigned to [e-mail]? Frank, do so if needed.

Comment 2

[blee]

This is a known problem that's been fixed for JA.
I tried this with both Naver and Yahoo (kr). Some links are displayed correctly
and some don't. I suspect ones that show up in garbage come in incorrect data
type when it arrived from database. If that's the case, this is not a client bug.

Comment 3

[msanz]

Still a problem in the server side; reassigning to kristif

Comment 4

[Kristi Fredrickson]

Young, please evaluate as to whether this is related to the UTF-8 conversion
problems which you've been working on a tool to solve.

Comment 5

[young]

This is same category bug(#14222) like latin1.
Yes. We need to contact server person.

Comment 6

[Kristi Fredrickson]

Per Bob Jung, this bug may be related to double-byte file problems that
existed in 4.0, and for which the international engineering team developed a
work-around for Japanese, but not Korean.

Reassigning to Bob Jung for evaluation.

Comment 7

[bobj]

As blee noted, this is a known what's related server problem and it also occurs
with 4.x clients.

The client I18N team provided a tool to what's related vendor to
heuristically convert Japanese web page titles to UTF-8 on the
server-side.  This tool did not handle anything other than Japanese,
so all other languages would be converted as if encoded in Latin1.
It was left up to the server folks and/or the vendor to extend the
tool for other languages/encodings.

Please contact me if you need the source to the tool that we sent
to the vendor.

Comment 8

[Kristi Fredrickson]

(per cherim, forwarding this to Alexa engineer for evaluation)

Q,
   Per my voicemail, here is the bug regarding UTF encoding of What's Related
Korean web page titles. Any information or contacts regarding whether Alexa will
be extending the the UTF-8 conversion tool (currently for Japanese only) to
include Korean?

Thanks,
Kristi

Comment 9

[Erik van der Poel]

Note that the tool that Alexa used to convert Japanese titles to UTF-8 was
written by me. I don't remember whether I sent them the source. They may not
be able to extend it for Korean.

Comment 10

[Kristi Fredrickson]

Reassigning to Jared Linley, the Netcenter producer for What's Related. He has 
access to the Netcenter engineers who are working with Alexa.

Comment 11

[Jared Linley]

This bug should be fixed with the new Alex database.

Comment 12

[Allan Masri]

verified