Hello -

I'm working on a team that is planning to add Welsh language support to a
large existing IT system which is partially web-based and
English-language-only so far. I've heard that 2 characters in Welsh
(w-circumflex and y-circumflex) are not supported in our default ISO-8859-1
character set, so a partial move to Unicode for internal storage of text
might be required.

I haven't yet found a Welsh-language website that uses these 2 characters,
so are they actually used much in Welsh? Is not supporting them likely to
cause problems?

Thanks

"Simon"  wrote in message
news:4861281d$0$78076$bed64819@news.gradwell.net...
> Hello -
>
> I'm working on a team that is planning to add Welsh language support to a
> large existing IT system which is partially web-based and
> English-language-only so far. I've heard that 2 characters in Welsh
> (w-circumflex and y-circumflex) are not supported in our default
ISO-8859-1
> character set, so a partial move to Unicode for internal storage of text
> might be required.
>
> I haven't yet found a Welsh-language website that uses these 2 characters,
> so are they actually used much in Welsh? Is not supporting them likely to
> cause problems?
>
> Thanks
>

I've just found a webpage that uses y-circumflex at the end of the third
paragraph, so it can't be that uncommon:
http://news.bbc.co.uk/welsh/hi/newsid_7460000/newsid_7462500/7462534.stm

This webpage uses ISO-8859-1 with entities for the y-circumflex. Using
entities would be very messy in my application, so if support for these
characters is needed, I would have to go for Unicode.
I guess my question still is: would not supporting these 2 characters be
considered bad practice for a Welsh-language business application?

Scripsit Simon:

> I'm working on a team that is planning to add Welsh language support
> to a large existing IT system which is partially web-based and
> English-language-only so far.

Do you plan to add other languages later? Is this about names only or 
also about prose texts? After all, ISO-8859-1 is insufficient even for 
normal English prose; think about dashes and proper quotations marks.

> I've heard that 2 characters in Welsh
> (w-circumflex and y-circumflex) are not supported in our default
> ISO-8859-1 character set,

Right. They are included in ISO-8859-14 (a.k.a. ISO Latin 8, or 
"Celtic"), but that’s not a feasible option on the WWW (IE does not 
recognize that encoding).

> so a partial move to Unicode for internal
> storage of text might be required.

That might be easy, or it might be extremely complicated. But that's 
really beyond the scope of these groups. As far as WWW authoring is 
concerned, Unicode - specifically UTF-8 - is a good option, but you 
could keep using ISO-8859-1 and represent those letters using character 
references like ŵ for w with circumflex. But you might have to deal 
with the encoding problem of the data bases involved, for example, and 
with data entry.

> I haven't yet found a Welsh-language website that uses these 2
> characters, so are they actually used much in Welsh?

I don't know Welsh, but I expect those characters to be so rare that 
using some clumsy notation like character references for them wouldn't 
be a major problem.

> Is not supporting them likely to cause problems?

Some people might say that it is tolerable to omit the circumflex, but 
it may be distinctive (i.e. the only difference between otherwise 
identical words, thought the context usually resolves the issue). And in 
2008, I think it is inappropriate to add support to languages to IT 
systems without supporting them properly, with all the characters needed 
for their correct writing.

-- 
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

"Jukka K. Korpela"  wrote in message
news:lha8k.22674$_03.7947@reader1.news.saunalahti.fi...
> Scripsit Simon:
>
> > I'm working on a team that is planning to add Welsh language support
> > to a large existing IT system which is partially web-based and
> > English-language-only so far.
>
> Do you plan to add other languages later? Is this about names only or
> also about prose texts? After all, ISO-8859-1 is insufficient even for
> normal English prose; think about dashes and proper quotations marks.
>
> > I've heard that 2 characters in Welsh
> > (w-circumflex and y-circumflex) are not supported in our default
> > ISO-8859-1 character set,
>
> Right. They are included in ISO-8859-14 (a.k.a. ISO Latin 8, or
> "Celtic"), but that’s not a feasible option on the WWW (IE does not
> recognize that encoding).
>
> > so a partial move to Unicode for internal
> > storage of text might be required.
>
> That might be easy, or it might be extremely complicated. But that's
> really beyond the scope of these groups. As far as WWW authoring is
> concerned, Unicode - specifically UTF-8 - is a good option, but you
> could keep using ISO-8859-1 and represent those letters using character
> references like ŵ for w with circumflex. But you might have to deal
> with the encoding problem of the data bases involved, for example, and
> with data entry.
>
> > I haven't yet found a Welsh-language website that uses these 2
> > characters, so are they actually used much in Welsh?
>
> I don't know Welsh, but I expect those characters to be so rare that
> using some clumsy notation like character references for them wouldn't
> be a major problem.
>
> > Is not supporting them likely to cause problems?
>
> Some people might say that it is tolerable to omit the circumflex, but
> it may be distinctive (i.e. the only difference between otherwise
> identical words, thought the context usually resolves the issue). And in
> 2008, I think it is inappropriate to add support to languages to IT
> systems without supporting them properly, with all the characters needed
> for their correct writing.
>
> -- 
> Jukka K. Korpela ("Yucca")
> http://www.cs.tut.fi/~jkorpela/
>

Thanks for your reply.

Unfortunately multi-lingual support has not really been a priority in the
system design up to now,
although it has always been a possible future requirement. The system is a
complex mixture of
databases, Windows applications and web applications. I believe all the
databases and programming
languages we use already support Unicode , so I would aim to use that
support, rather than character
references which would be clumsy as you say.

Message-ID: <48612a47$0$78077$bed64819@news.gradwell.net> from Simon
contained the following:

>> (w-circumflex and y-circumflex) are not supported in our default
>ISO-8859-1
>> character set, so a partial move to Unicode for internal storage of text
>> might be required.
>>
>> I haven't yet found a Welsh-language website that uses these 2 characters,
>> so are they actually used much in Welsh? Is not supporting them likely to
>> cause problems?
>>
>> Thanks
>>
>
>I've just found a webpage that uses y-circumflex at the end of the third
>paragraph, so it can't be that uncommon:

and it's text/html;charset=iso-8859-1
-- 
Geoff Berrow  0110001001101100010000000110
001101101011011001000110111101100111001011
100110001101101111001011100111010101101011

"Geoff Berrow"  wrote in message
news:dnd264ducvnqg9gbsqu510mqgsdpl9r837@4ax.com...
> Message-ID: <48612a47$0$78077$bed64819@news.gradwell.net> from Simon
> contained the following:
>
> >> (w-circumflex and y-circumflex) are not supported in our default
> >ISO-8859-1
> >> character set, so a partial move to Unicode for internal storage of
text
> >> might be required.
> >>
> >> I haven't yet found a Welsh-language website that uses these 2
characters,
> >> so are they actually used much in Welsh? Is not supporting them likely
to
> >> cause problems?
> >>
> >> Thanks
> >>
> >
> >I've just found a webpage that uses y-circumflex at the end of the third
> >paragraph, so it can't be that uncommon:
>
> and it's text/html;charset=iso-8859-1
> -- 
> Geoff Berrow  0110001001101100010000000110
> 001101101011011001000110111101100111001011
> 100110001101101111001011100111010101101011

Hello -

Unfortunately (for me) that webpage uses character entities to represent the
characters outside ISO-8859-1. This isn't really a workable approach for me,
because the text I'm displaying will be stored and processed in various
databases and applications (web and non-web). I will probably end up storing
and processing the data using UCS-2 or similar and generating webpages in
UTF-8.

Scripsit Simon:

> I believe all
> the databases and programming
> languages we use already support Unicode , so I would aim to use that
> support, rather than character
> references which would be clumsy as you say.

Sounds like a simple way to go then. It is surely simplest to use 
Unicode throughout, especially if character data needs to be transferred 
between applications as plain text (where no character references or 
markup can be used). It's also simplest in data entry if people 
immediately see what they have typed, and entering characters with 
circumflex should not be a problem; you can e.g. use the keyboard layout 
outlined at
http://en.wikipedia.org/wiki/Keyboard_layout#United_Kingdom_extended

Yet, it's always possible that some software component doesn't grok 
Unicode. Let's hope such problems are solvable. The web-related 
components shouldn't be a problem.

-- 
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Simon wrote:
> Hello -
> 
> I'm working on a team that is planning to add Welsh language support to a
> large existing IT system which is partially web-based and
> English-language-only so far. I've heard that 2 characters in Welsh
> (w-circumflex and y-circumflex) are not supported in our default ISO-8859-1
> character set, so a partial move to Unicode for internal storage of text
> might be required.
> 
> I haven't yet found a Welsh-language website that uses these 2 characters,
> so are they actually used much in Welsh? Is not supporting them likely to
> cause problems?

It could be a support problem (though I don't know why, given the 
availability of UTF-8 as well as the option of numeric character 
references): see the note at the bottom of

http://www.menai.ac.uk/clicclic/

As made clear at

http://www.cs.cf.ac.uk/fun/welsh/Lesson01.html

the circumflex really is supposed to appear in these locations. (Note 
that even on this page, section 1.2 explains that because of support 
issues, they are using their own ugly work-around for accented 
characters.) Examples are given: "ty^" = "house", along with the pair 
"gw^ydd" = "goose" and "gwy^dd" = "trees", which are pronounced differently.

Message-ID: <486139da$0$78081$bed64819@news.gradwell.net> from Simon
contained the following:

>
>Unfortunately (for me) that webpage uses character entities to represent the
>characters outside ISO-8859-1. This isn't really a workable approach for me,
>because the text I'm displaying will be stored and processed in various
>databases and applications (web and non-web). I will probably end up storing
>and processing the data using UCS-2 or similar and generating webpages in
>UTF-8.


Surely you can add the character entities using a script when the pages
are generated?
-- 
Geoff Berrow  0110001001101100010000000110
001101101011011001000110111101100111001011
100110001101101111001011100111010101101011

On Tue, 24 Jun 2008 19:15:48 +0100, Simon wrote:

> 
>  I will probably
> end up storing and processing the data using UCS-2 or similar and
> generating webpages in UTF-8.

I'll add my vote for UTF-8 as the way to go if there's a choice. Either 
way there'll be some problems but UTF-8 is likely to be more future-proof 
in the long run.

Oh, and I have encountered the need for a w with circumflex, but that was 
an old song title so it might have been an archaic Welsh form.

-- 
Anahata
anahata@treewind.co.uk  ==//== 01638 720444
http://www.treewind.co.uk ==//== http://www.myspace.com/maryanahata

Simon wrote:

> I've just found a webpage that uses y-circumflex at the end
 > of the third paragraph, so it can't be that uncommon:
> http://news.bbc.co.uk/welsh/hi/newsid_7460000/newsid_7462500/7462534.stm

That’s really odd! They have a “y with circumflex” (ŷ)
but write “vowel with ASCII apostrophe” instead of
“vowel with acute” throughout.

Btw:
You should start with your newsreader^W Outlook Express
in supporting special, non-ASCII characters:

Tools → Options → Send
   Mail Sending Format → Plain Text Settings → Message format MIME
   News Sending Format → Plain Text Settings → Message format MIME
                                               Encode text using: None

Otherwise you cannot even write or quote

    1 € = 100 ¢

Simon wrote:

> I've heard that 2 characters in Welsh (w-circumflex and y-circumflex)
> are not supported in our default ISO-8859-1 character set, so a
> partial move to Unicode for internal storage of text might be required.
> 
> I haven't yet found a Welsh-language website that uses these 2 characters,
> so are they actually used much in Welsh? Is not supporting them likely to
> cause problems?

ISO-8859-1 does not even contain a euro sign (€), which seems to be
an even stronger argument to move to Unicode asap than the missing
Ŵ ŵ Ŷ ŷ for Welsh.

Holy crap.  I'm looking at two of your posts, and both in the body and in
the article's line in the headers pane, your name is not in the font I
have configured.  And it's a *different* not-configured-by-me font in the
body than in the headers pane.


-- 
Blinky
Killing all posts from Google Groups
The Usenet Improvement Project -->  http://improve-usenet.org
Found: a free GG-blocking news *feed* -->  http://usenet4all.se

Blinky the Shark wrote:
> Holy crap.  I'm looking at two of your posts, and both in the body and in
> the article's line in the headers pane, your name is not in the font I
> have configured.  And it's a *different* not-configured-by-me font in the
> body than in the headers pane.

I noticed the same thing, in Thunderbird.

Harlan Messinger wrote:
> Blinky the Shark wrote:
>> Holy crap.  I'm looking at two of your posts, and both in the body and in
>> the article's line in the headers pane, your name is not in the font I
>> have configured.  And it's a *different* not-configured-by-me font in the
>> body than in the headers pane.
> 
> I noticed the same thing, in Thunderbird.

His FROM line reads

From: =?UTF-8?B?77yh772O772E772S772F772B772T44CA77yw772S772J772M772P772Q?=

I don't know what to make of this.

Harlan Messinger  writes:

> Harlan Messinger wrote:
>> Blinky the Shark wrote:
>>> Holy crap.  I'm looking at two of your posts, and both in the body and in
>>> the article's line in the headers pane, your name is not in the font I
>>> have configured.  And it's a *different* not-configured-by-me font in the
>>> body than in the headers pane.
>>
>> I noticed the same thing, in Thunderbird.
>
> His FROM line reads
>
> From: =?UTF-8?B?77yh772O772E772S772F772B772T44CA77yw772S772J772M772P772Q?=
>
> I don't know what to make of this.

If I cut and paste to my utf-8-dump program:

$ utf-8-dump -f '[%u] %n\n'
Ａｎｄｒｅａｓ　
[U+FF21] FULLWIDTH LATIN CAPITAL LETTER A
[U+FF4E] FULLWIDTH LATIN SMALL LETTER N
[U+FF44] FULLWIDTH LATIN SMALL LETTER D
[U+FF52] FULLWIDTH LATIN SMALL LETTER R
[U+FF45] FULLWIDTH LATIN SMALL LETTER E
[U+FF41] FULLWIDTH LATIN SMALL LETTER A
[U+FF53] FULLWIDTH LATIN SMALL LETTER S
[U+3000] IDEOGRAPHIC SPACE
[U+000A] <control>

Presumably your newsreader thinks it needs a separate font to find
suitable gyphs for these characters (mine does too).

-- 
Ben.

Scripsit Ａｎｄｒｅａｓ　Ｐｒｉｌｏｐ:

> ISO-8859-1 does not even contain a euro sign (€), which seems to be
> an even stronger argument to move to Unicode asap than the missing
> Ŵ ŵ Ŷ ŷ for Welsh.

Not really, because
a) the UK does not use the euro currency
b) the euro sign can conveniently be written using the entity reference 
€
c) the euro sign should not be used in normal text, according to 
reputable language authorities; instead, the currency name should be 
written, except perhaps in tables and other contexts where saving space 
is crucial.

For commercial pages oriented towards countries using the euro, the euro 
sign is needed, but it’s not really comparable to the issue of letters 
needed for proper writing of a language.

-- 
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

On Wed, 25 Jun 2008, Jukka K. Korpela wrote:

>> ISO-8859-1 does not even contain a euro sign (€), which seems
>> to be an even stronger argument to move to Unicode asap
>
> Not really, because
> a) the UK does not use the euro currency

By that logic, they won't need a dollar sign on their keyboards.

> b) the euro sign can conveniently be written using the entity
> reference €

In HTML. But IIRC, the OP wrote of some "large existing IT system"
with internal ISO-8859-1 character set. I wonder if one could
write  €  there.

Message-ID:  from Ben Bacarisse contained the
following:

>Presumably your newsreader thinks it needs a separate font to find
>suitable gyphs for these characters (mine does too).

Yup, just get a load of question marks in Agent.

-- 
Geoff Berrow  0110001001101100010000000110
001101101011011001000110111101100111001011
100110001101101111001011100111010101101011

In comp.infosystems.www.authoring.html message <kv6dnRvtYfLOa_zVnZ2dnUVZ
8hednZ2d@posted.plusnet>, Wed, 25 Jun 2008 02:40:03, anahata
 posted:
>
>Oh, and I have encountered the need for a w with circumflex, but that was
>an old song title so it might have been an archaic Welsh form.

A search for "Welsh Water" rapidly locates <http://www.welshwater.com/>,
in the foot of which is "Dwr Cymru Cyf 2008."  Copy'n'paste into here
has not reproduced the circumflex over the w; but graphic copy'n'paste
into Paint, zoomed, reveals it well.

See also <http://www.dwrcymru.co.uk/Welsh/Contactus/index.asp>, or
<http://cy.wikipedia.org/wiki/Hafan>.

ISTM unlikely that the Welsh could manage without their word for water.

-- 
 (c) John Stockton, nr London UK. replyYYWW merlyn demon co uk Turnpike 6.05.
 Web <URL:http://www.uwasa.fi/~ts/http/tsfaq.html> -> Timo Salmi: Usenet Q&A.
 Web <URL:http://www.merlyn.demon.co.uk/news-use.htm> :  about usage of News.
 No Encoding. Quotes precede replies. Snip well. Write clearly. Mail no News.

Andreas Prilop wrote:

<snip>

For the record, I see no unusual font behavior in the From field from that
post.


-- 
Blinky
        Is your ISP dropping Usenet?
        Need a new feed?
        http://blinkynet.net/comp/newfeed.html

On Wed, 25 Jun 2008 21:48:14 +0200, Dr J R Stockton  
 wrote:

> In comp.infosystems.www.authoring.html message <kv6dnRvtYfLOa_zVnZ2dnUVZ
> 8hednZ2d@posted.plusnet>, Wed, 25 Jun 2008 02:40:03, anahata
>  posted:
>>
>> Oh, and I have encountered the need for a w with circumflex, but that  
>> was
>> an old song title so it might have been an archaic Welsh form.
>
> A search for "Welsh Water" rapidly locates <http://www.welshwater.com/>,
> in the foot of which is "Dwr Cymru Cyf 2008."  Copy'n'paste into here
> has not reproduced the circumflex over the w;

Hmmm. I would find that hard to believe (probably just a newsreader  
thing), just a test (UTF-8 posting):
Dŵr Cymru Welsh Water
-- 
Rik Wasmus

On Thu, 26 Jun 2008 08:08:36 +0200, Rik Wasmus  
 wrote:
> just a test (UTF-8 posting):
> Dŵr Cymru Welsh Water

Well, at least here it's clearly there :)
-- 
Rik Wasmus
...spamrun finished

Op 25-06-08 18:01 heeft Ben Bacarisse als volgt van zich laten horen:
> Harlan Messinger  writes:
> 
>> Harlan Messinger wrote:
>>> Blinky the Shark wrote:
>>>> Holy crap.  I'm looking at two of your posts, and both in the body and in
>>>> the article's line in the headers pane, your name is not in the font I
>>>> have configured.  And it's a *different* not-configured-by-me font in the
>>>> body than in the headers pane.
>>> I noticed the same thing, in Thunderbird.
>> His FROM line reads
>>
>> From: =?UTF-8?B?77yh772O772E772S772F772B772T44CA77yw772S772J772M772P772Q?=
>>
>> I don't know what to make of this.
> 
> If I cut and paste to my utf-8-dump program:
> 
> $ utf-8-dump -f '[%u] %n\n'
> Ａｎｄｒｅａｓ　
> [U+FF21] FULLWIDTH LATIN CAPITAL LETTER A
> [U+FF4E] FULLWIDTH LATIN SMALL LETTER N
> [U+FF44] FULLWIDTH LATIN SMALL LETTER D
> [U+FF52] FULLWIDTH LATIN SMALL LETTER R
> [U+FF45] FULLWIDTH LATIN SMALL LETTER E
> [U+FF41] FULLWIDTH LATIN SMALL LETTER A
> [U+FF53] FULLWIDTH LATIN SMALL LETTER S
> [U+3000] IDEOGRAPHIC SPACE
> [U+000A] <control>

Those are actually fake Latin letters, which are used in Japanese and 
Chinese systems, since the CJK symbols are broader, therefore they have 
broader latin letters as well.  It's a mean trick to make your name look 
different without using html.

> Presumably your newsreader thinks it needs a separate font to find
> suitable gyphs for these characters (mine does too).

That's very probable, since most fonts won't contain those glyphs. 
You'd need a Chinese/Japanese font which contains them.

H.

In uk.net.web.authoring message , Thu,
26 Jun 2008 08:08:36, Rik Wasmus  posted:
>> A search for "Welsh Water" rapidly locates <http://www.welshwater.com/>,
>> in the foot of which is "Dwr Cymru Cyf 2008."  Copy'n'paste into here
>> has not reproduced the circumflex over the w;
>
>Hmmm. I would find that hard to believe (probably just a newsreader
>thing),

Then you should read my headers.

-- 
(c) John Stockton, nr London UK.  ???@merlyn.demon.co.uk  Turnpike v6.05  MIME.
 Web  <URL:http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
     Check boilerplate spelling -- error is a public sign of incompetence.
    Never fully trust an article from a poster who gives no full real name.

Andreas Prilop wrote:
> On Wed, 25 Jun 2008, Jukka K. Korpela wrote:
> 
>> a) the UK does not use the euro currency
> 
> By that logic, they won't need a dollar sign on their keyboards.

The dollar key is kinda handy for some programming languages (Perl, PHP, 
XSLT, others too). 

The missing '#' on UK Mac keyboards is very annoying. Although Alt+3 does 
make it appear, it's easier to just remap the '§' key. (And why they felt 
the need to have '§' when they don't have a '#' or '€' remains a mystery!)

-- 
Toby A Inkster BSc (Hons) ARCS
[Geek of HTML/SQL/Perl/PHP/Python/Apache/Linux]
[OS: Linux 2.6.24.4-1mnbcustom-g5n1, up 43 days, 19:44.]

                              Olympics Monkey
          http://tobyinkster.co.uk/blog/2008/07/31/olympic-monkey/

> Dwr Cymru Cyf
> Dwr Cymru Welsh Water

The Windows Courier font does not seem to support UTF-8. Outlook
Express changes the font from Courier to a variety of Arial when the
Encoding settings get changed to Unicode (UTF-8).

-- 
Jim Carlock
More Than Five Senses... We All Know More.
http://www.associatedcontent.com/article/381163/more_than_five_senses.html

On Mon, 25 Aug 2008, JC wrote:

> The Windows Courier font does not seem to support UTF-8.

The *bitmap font* Courier, which comes with MS Windows and
which has nothing to do with other PostScript/TrueType fonts
of the same name, has a *very* restricted character repertoire.
You should use Courier New (TrueType) instead.

-- 
http://groups.google.com/groups/search?q=author:Alan.J.Flavell