> > only a person with a graphics card and a little memory (we're
> > talking machines with 24MB or more..) would have.
>
> Hmm.. I take it this isn't something that could be switched on or off
> with preferences?  Or even seperate executables?  (like the versions
> optimised for different processors)?

Sure it is, but turning off something like "International Language" and
"decent font support" by default is a little lame, don't you think?

The greatest feature of using outline fonts directly (or semi-directly)
is that you can set the DPI of the "target" - MacOS IE, Mozilla and
Opera let you set the default of 72 or 96 dpi setting and so on so
that incredibly tiny font sizes (7pt would be unreadable on MacOS
and possibly AmigaOS) would be "scaled" up slightly.

> > making a few glyphs suck up a few megabytes. Even Ken's system
> > will suffer this.
>
> There is going to be a price to pay with any system of dealing with
> glyphs.  On the other hand, how many people have minimum specced
> Amigas for daily use?

Well we deal with glyphs right now anyway: just in 10k bitmap fonts
and poorly handled outline-sourced bitmap fonts.

When your entire font set is in a normal style, and your italics and
bold characters are automatically generated, that's a little space
saved there.

I can implement real font family support now, which would leverage
ttf.library features to load real Italic glyphs, via standard ordinary
font APIs (i.e. load the font as "Helvetica" with italic set in the ta_Style
flags and see if it comes back with the flag still set..)

That feature alone would increase memory requirements threefold
the moment you set a bold or italic text on a page, since all 256
iso-8859-1 glyphs are loaded.

For an outline font your memory requirements are, simply: the size
of every glyph you open, plus 16 bytes per glyph overhead on
structures, plus possibly the size of the font if kept in memory
(ttf.library
will spool from disk if it is low on memory, or the font is ridiculously
sized.. but a 1MB font would go into memory and stay there so
that glyph generation is almost instant)

The difference between using outlines, and using graphics.library etc.
is that possibility of keeping the outline font in memory. Opening a
normal font will load the outline to spool off 256 glyphs and then close
it and free the memory. It's not nice to keep pushing and pulling it out
of memory unless you're sure you're not going to use it anymore.

> support those users as best as possible.  If CSS requires glyphs, and
> glyphs require extra memory and gfx-cards--then why not "ship" V with
> that turned off?  Then the user who has the extra required memory can
> turn the switch.

CSS doesn't require it, but we already saw that we're missing important
unicode character literals &blah; and so on, and people want UTF8
support (an English page in UTF8 wouldn't be so bad.. you just end up
with standard readable text and then couplets of accented characters
where there is a two-byte character instead of one..) and you can't do
that with plain ordinary Amiga fonts.

Not even H&P's insane switch to iso-8859-15 will help you.

> > have to cull your userbase for a certain feature. So much so
> > that I wouldn't do it.
>
> For the time being, MorphOS would count me out.  I don't have PPC.
> This may change in the future, when MOS is a "released" product.  I'm
> not anti-MOS or anti-Name.  As far as hardware, I hope that I find the
> h/w solution to let me run both.

Well MorphOS is the one I have access to. AmigaOS 4.x isn't even
tentative right now in my eyes: I have yet to see proof that those
screenshots they have online are actually generated by a PPC processor
even. The moment I get a developer kit for it (which will probably be a
few months after it gets released.. and who knows when that is?)

Amithlon is a possibility since it does have huge amounts of proc
power at it's disposal, and PC's generally have a lot of memory. But
I don't have a copy. But if you whined for it enough, I don't see why
I couldn't implement it, albeit in 68k and slow on a real Amiga, for
the sole purpose of Amithlon users, or insane Amiga users.

> (note, the mui smart refresh that bogs a real amiga--KILLS Voyager
> under amithlon...  is it using a trick of the custom chips?
> Otherwise, things like FLASH move as fast as they do on the
> Windows/Linux side.)

Smart Refresh is simply a waste of processor time and memory.

> In what context?  Will it break Voyager's back, or will it squish
> low-end support?

Probably squish someone's brain.

>  I thought that Ollie rewrote V because of CSS and
> other stuff down the line that would break the old V's back?

I mean V is rewritten to be properly object oriented and is a highly
advanced little system internally, and very easy to add CSS support to
the objects we already have (even easier when the CSS attributes do
the same thing as HTML 4 stuff) - but parsing CSS looks like a
nightmare once you get to CSS2 and things. Where's Tom Hurst?
He can give you a few special CSS tricks that are a pain to even
code onto a website :P

I just fear you'd get something which allows setting colours and
borders and not much else for a while. I'd love to have something
like changing the text inside a <span> tag dynamically. But.. it's a
pain ;)

Hell, we don't even enumerate <span> tags in the DOM yet. We
left it out so we could see where CSS is really used ;)

Dynamic menus would be possible to implement. Layers. But
there be dragons in the CSS spec. I hope what you get is enough
and you don't end up with something akin to the JS support on
the current browsers right now ;)

> certainly hope you mean old miggy support...  Besides, the Vapor team
> could always release V2.95/6 as the suggested browser for low-end
> amigas..

But it sucks :)

--
Matt


Reply via email to