Hello Ken
On 13-Jul-01, you wrote:
> But CSS2 is another matter. I was talking about
> HTML4 + CSS2 + Unicode all three, and I stand by my claim that a HTML4
> viewer would have to be written with CSS2 (or any other style system) in
> mind, that style sheets can't just be plugged in. They can be plugged in to
> the HTML4* writer* - of course! That's what style sheets are for - but *not* the
> HTML *reader*. I know enough about programming to know HTML4+CSS2
> reader would be* very* messy to program. Especially with the tight memory
> constraints that Amigas present.
V is ready for CSS - it's been designed in as object oriented a manner as
possible such that adding stuff like DOM properties, and parsing
stylesheets such that they are set properly is quite easy: as long as you
know what you're doing.
V 3.2 was not ready for this: which is why there is a nice new layouter
etc.
>> would be the difficulty in adding a Unicode conversion step from some
>> web codepage to some standard one, UTF8 or UTF16 let's say, usable by
>> a new font API?
>>
>> About 10 lines of code in various places? Less than that, even..
>
> 10!! - are you serious? more like >10,000. Unicode doesn't come in images,
> it comes as text, and it is very complex and intricate changing it into
> images, especially when bearing in mind the needs of CSS2. Take it from
> one who knows.
If you have a Unicode font (say, a Microsoft Truetype webfont) and the OS
or some library supports extracting glyphs and rendering them then the
process really is as simple as:
* Get codepage used from web page meta tag or server response
* Convert text in page from codepage to some Unicode format (UCS2 or
UTF8 or UTF16 or whatever) used by the OS API
* Pass UCS/UTF text to OS API
V is entirely capable of doing the first. Linux/BSD etc. offers a function
called iconv() which converts between codepages and unicode text
formats - either libiconv or hacked into glibc 2.1 - and this *has* been
ported to the Amiga as a link library.
All that is missing is this mystical OS API - all it needs is to be able to
extract glyph images from fonts as per the codes it gets from the
data you pass (caching them if necessary so that it doesn't pull out
standard Latin characters every single time) and render them to the
screen. Most unicode support is actually built into the font engines:
for instance, ask for a "e with acute accent" and it'll pull the two
glyphs ("e" and the accent) out of the font and merge them for you.
You don't need to store a shitload of graphics to draw text to a screen
if you have the vector information to store them and the ability to pick
out arbitrary codepoints.
I'm intrigued as to how you've implemented your system if you require
that you send a bunch of glyphs to people.. can I have a copy so that
I can see? :P
> When the next Amiga OS comes out, I expepct that Voyager will need a
> major re-write anyway.
Total crap, I'm afraid - when the next Amiga OS comes out, if it's not
compatible with original Amiga apps such that V would require such
a major rewrite as you suggest, then they really will have broken the
AmigaOS something scarily.
In fact, due to the required design of a native PPC AmigaOS, and if
they aren't screwing things sideways to do it, porting to PPC for it
would be no more difficult than to MorphOS. All that is required is
a little motivation.
I really wish people wouldn't assume that just because Voyager is
an Amiga application, that it doesn't employ modern techniques to
make it work like it does; and that suitably "advanced" things like
Unicode, CSS or whatever would be inherently impossible. I assure
you - everything you can do in Internet Explorer on Windows, I'm
sure Voyager could do with the greatest of ease. But Microsoft has
had a larger team of full-time paid professionals on the job - and
in the same amount of time (7 years :) has managed to produce
something as comprehensive as IE6 that a part-time one-man show
and a little help from friends could never hope to achieve.
But.. that said.. all it needs is time.
> So, hopefully, we will see it provide Unicode functionality.
Voyager already has perfectly good support for Unicode, just not the
facilities to use it (i.e. no standard implementation of a text conversion
routine, and no way of rendering to the screen)
Thanks
--
Matt <[EMAIL PROTECTED]>
_____________________________________________________________________
Voyager Mailing List - http://v3.vapor.com/
Voyager FAQ....: http://faq.vapor.com/voyager/
Listserver Help: mailto:[EMAIL PROTECTED]?Subject=HELP
Unsubscribe....: mailto:[EMAIL PROTECTED]?Subject=UNSUBSCRIBE
