Re: Graphics on the net?

[nathan@cco.caltech.edu (Nathan Mates)]

In article <35vk5v$8dd@crl3.crl.com>, Joe Kohn <joko@crl.com> wrote:
>[Various IIGS GIF decoder dicussion]

>I've never seen Prism to see how it works. Is it any good?

   If you want to make a 256 or 3200 color pic, it's quite nice. Some
of the coloring isn't quite as good as Todd Whitesel's "Lord High
Giffer", but it's desktop based, and handles various other formats.


>My frustration started about a week ago when I visited a friend with
>a Mac, and I got to see with my own eyes graphics that I've displayed
>on the GS. There was no comparisons.

   Tell the friend to drop the Mac down to 16 colors (most Mac GIF
viewers force the monitor into 256 color mode or better because they
don't want to go through reducing the number of colors) and lower the
resolution, and it'll be just the same as the GS. Pixel resolution
isn't bad on the GS if you look at something in grayscale, (IBMs tend
to have these really large and chunky 320x200 modes; the GS kicks
that) but when things get colored, the 16 colors/line limit of the GS
is the main determinant.


>Source code for GIF is avaailable, and I'd be happy to find it and
>make it available if anyone thinks they can come up with a superior
>decoder.

   Won't do much. Buying us all Second Sight boards (you _do_
subscribe to GS+ and saw the ad, didn't you? ;) would be the biggest
help. The algorithm for translating a picture from 256 down to 16
colors is fairly simple (you can look it up in most graphics textbooks
if you'd rather not find it out for yourself); I can't believe that
many different programmers could all muck the same thing up to
produce roughly the same results.

   The GS monitor can display things well (aka VOC); it's mostly
the darn limitation on colors per line (Prism 3200s aren't as bad,
but still not as good as other machines) that's killing us.


>On the other hand, Jpeg's seem to produce great graphics, but the
>only Jpeg IIGS Jpeg decoder that I'm aware of is lacking due to 2
>reasons...it's only grayscale, and there are no conversion possible.

   There are unix programs to go from jpeg to gif. (The conversion
is mostly a lot of floating-point number crunching, which is why
you do NOT want to do this on a GS if at all possible). But, as I
said earlier, jpegs tend to have a higher quality source to start
with, so you get better results out on the other end.


>I'd love to be able to download GIF and Jpegs, and then convert them
>to a program that can print them. With some of the large size GIF's,
>ShowMe will let you save them as AFP, and those can be loaded into
>Platinum Paint, and scrolled, but when it comes to printing, you miss
>a lot.

   DreamGrafix recommends that you convert pictures to grayscale before
trying to print. Trying to print color graphics on a B&W printer
only comes up with horrible dithering patterns that could be easily
avoided if you let the printer do the dithering.   


>Since no real information is being made available about Second Sight,
>who knows, maybe when it's released, it'll let me view phenomenal
>color images from Hubble Space Telescope.

   The screen shots on the back of the GS+ magazine were quite nice.
I personally think that the results from the GS monitor could be
improved quite a bit; a VOC pumps out quite a nice display (sharper
than most TVs) to mine.


>If not, maybe we should organize a group of net programmers to give
>us some tools that will let GS owners have some better tools. I'm not
>a programmer, but I'd be willing to help to organize the project. Or,
>maybe this is perfect for Phoenix project.

   Use unix boxes for the jpeg number crunching, and a Second Sight
card to handle the display. That'll do 90% of what you want while
not tying up programmers with already-doable projects.

   I don't think you've ever studied groups of programmers, or their
practical uses. The thing to remember is "Adding programmers to an
already late project makes it later." Software teams have their
advantages for really large programming projects, but trying to
coordinate very separated part-time programmers for a project will
waste too much time in communication as opposed to real programming.

   One good asm programmer working full-time can produce 10,000 lines
of code a month. You'd probably have to resort to cloning to find 30
programmers who could collectively do 10,000 lines a day on the same
program.


>Whatcha think? After all, if an Amiga can have Mosaic, why can't we?
>If MacNewbies can show me net graphics that cause my jaw to fall, why
>can't one of the GS programming gurus come up with something as
>impressive?

   You could be a GS programming deity, but you have to work with the
limitations of the built-in hardware, or make new hardware. The GS
has a lot of limits (no memory protection, 2.5/1 Mhz 8-bit bus), and
we're stuck with them. We have to fight the hardware to pull off
what we've done so far; MacNewbies have things handed to them on
a silver platter.

   Amiga programmers have the benefit of a decent set of development
tools (Orca/C doesn't cut it for porting programs from other
platforms), the computer has been a graphically-oriented platform
(with the resolution and colors to match), and a lot more diehard
programmers. I'd love to be able to program full-time, but there's
this darn thing known as college, grades, and grad school on the
horizon.


>You many say I'm a dreamer, but I'm not the only one.
>                                   --- John Lennon

My response... Change one word, and everything takes on a different
tone:
 "You can call me an axe murderer, but I'm not the only one." [Kidding! ;]

   Just because others do it doesn't mean it's acceptible. There are
far better motivations to do things than being part of a crowd.


Nathan
-- 
* Nathan Mates   http://www.ugcs.caltech.edu/~nathan/               *
* Ftp humor archiver: ftp to cco.caltech.edu, look in pub/humor     *
* "Always listen to experts. They'll tell what can't be done, and   *
*  why. Then do it."  -- Robert A. Heinlein, Lazarus Long           *