Re: [Slashcode-development] towerofbabel.com

[Malcolm L. <ma...@to...>]

"Slash doesn't need to deal with any character set, it only needs to deal
with one - Unicode. The modern browser should/will make conversions to/from
Unicode for both inbound and outbound data. If slash is rewritten to support
Unicode instead of whatever it is now (ISO-8859-1 probably?), then that is
all there is to  it as far as the database is concerned. Even templates
could then be in many languages, even within a template if so desired."

Rightio. Any other voices like to chime in about slash and Unicode?


Malcolm Lawrence
Editor-in-Chief
Babel: The multilingual, multicultural
online journal and community of arts and ideas.
http://www.towerofbabel.com
------------------------------------------------------------
Babel knows: People who don't work with their hands are parasites.
------------------------------------------------------------

----- Original Message ----- 
From: "Barry Caplan" <bc...@i1...>
To: "Malcolm Lawrence" <ma...@to...>;
<sla...@li...>
Sent: Thursday, July 01, 2004 7:06 PM
Subject: Re: [Slashcode-development] towerofbabel.com


> At 06:23 PM 7/1/2004, Malcolm Lawrence wrote:
>
> >"I can say quite comfortably that using flags as a ui device to indicate
> >language or any other locale issue is a bad bad idea."
> >
> >True enough. I've been building the site for 8 years and am well aware of
> >all the arguments against. But until a more suitable design element can
be
> >implemented (not to mention appropriate icons for topics) they'll have to
> >do. They're pretty, too.
>
>
> Hmm. I guess I wasn't clear enough on this point.
>
> There is not one single professional localizer on the face of the earth
that would recommend flags as an UI element indicating locale.
>
> I hope that is plain enough. If not, at least it will turn up in google
for someone someday who will read why and decide to not use flags. )
>
> The reasons, as I enumerated some of them, fall in the general category of
"They don't have a one to one relationship with what is being described" and
"users have visceral, political feelings about them, usually rightfully so".
>
> Example: I am an English speaker in the US. I sometimes see a UK Union
Jack flag which is meant to indicate English text. But does it? Or is it
something more specifically about the UK that lurks in that site? What about
people in other countries? Do I have to know their flags too? Australia's
first language is English - what flag should they use on their sites to
indicate English localization?
>
> Example: I live in Silicon Valley where > 30% of people do not speak
English as a first language at home (maybe > 50%! I forget so I will go with
the conservative value for now). So if I have content that is US locale
based, but localized into various languages such as Chinese, Spanish,
Vietnamese, etc., what flags shall I use? If not US, then what one to use
again falls into the problem just mentioned - e.g. Spanish is spoken many
places, and is not the same Spanish everywhere anymore then English is
universal.
>
> And even if I somehow choose a flag, then what if the same language is
used in a different place on the site? I surely won't be using the Taiwan
flag everywhere there is Chinese, nor the PRC flag. Except in those 2
countries, it is hard to see which would be appropriate without offending.
>
> Finally, take a look at Canada. With 2 official languages (English and
French), what does a Maple Leaf flag say about the language of  the material
behind it? Nothing at all! But it might say something very serious about
political issues you don't mean to say!
>
> The reason there are no generally accepted icons to represent locale is,
quite frankly, because there are none to be had.
>
> It is not as though people have not put a lot of thought and hard earned
experience into this, looking for a good way. They have. A lot of people and
a lot of time.
>
> If it is pretty you want, then make whatever you do pretty, whatever that
means to you. You can do that and have plain text too. Photoshop works
wonders for that :)
>
> For an example of a pretty good effort, look at how yahoo.com indicates
locales on their pages. That basic technical approach, coupled with whatever
UI beautification is needed for your site, is a pretty good way to go.
>
> As for if slashcode supports Unicode, I don't know as it has been well
over a year since I looked at it. How complicated that would be to do is a
matter of conjecture - there are technical, testing, and management issues
to consider. I have done just that with closed source code that was far more
complex then slashcode, so I am confident it could be done. That it hasn't
been done until now (if it hasn't been done) strikes me a a combination of
all three factors.
>
> AFIK Unicode support is de rigeur for any new project that hopes to scale.
For any existing project that hopes to stick around with a worldwide user
base, then the switch is going to need to be made. I think I may be
volunteering to work with the developers to understand what the effort level
and tasks should be, so that people can sign onto them in a coherent
fashion. But my experience is that this sort of conversion does need to be
coordinated very closely with any other ongoing development, otherwise it is
just a fork in the code and no one wants that. Managing the codelines so
they don't fork (or they do but they merge back together at a defined point
more likely) is something I can definitely bring to the table.
>
> >"In a perfect world, the browser would accept HTML in Unicode and display
> >properly from there. For folks whose users all have modern browsers, that
is
> >possible. The browser will make the conversion to the right character
> >encoding locally, or it will have Unicode fonts enabled. When there are
> >Unicode fonts available, then you get the advantage of displaying
multiple
> >languages on a single page, which, frankly, is what I would expect of a
site
> >called "towerofbabel.com", slashcode or not :)"
> >
> >Well, the browser isn't the problem. It's the ability of slash to be able
to
> >deal with any character set when a story is submitted or a comment
posted.
> >Would those problems go away as soon as slash were dealing with all
internal
> >processing in Unicode?
>
> Slash doesn't need to deal with any character set, it only needs to deal
with one - Unicode. The modern browser should/will make conversions to/from
Unicode for both inbound and outbound data.
>
> If slash is rewritten to support Unicode instead of whatever it is now
(ISO-8859-1 probably?), then that is all there is to  it as far as the
database is concerned. Even templates could then be in many languages, even
within a template if so desired.
>
> In your case, you probably would not need to go to subdomains unless you
wanted to - you could just have slash sections for each language. And if
someone posted Chinese on the French page, so what? slash won't care so
neither should you. Trust me you don't want a case statement for every
codeset dependent feature in the code. Your domain name pretty much sums up
the reason why that is the case :)
>
> Best,
>
> Barry
>
>