ftp.xmission.com/2014.06.ftp.xmission.com.tar/ftp.xmission.com/pub/lists/fractint/archive/v01.n023

Internet Message Format | 1997-10-05 | 42KB

From: [e-mail] (fractint Digest)
To: [e-mail]
Subject: fractint Digest V1 #23
Reply-To: [e-mail]
Sender: [e-mail]
Errors-To: [e-mail]
Precedence:


fractint Digest        Sunday, October 5 1997        Volume 01 : Number 023



In this issue:

    Re: (fractint) Serious Fractint 19.6 bug?
    Re: (fractint) SnowFlakeColor
    Re: (fractint) Serious Fractint 19.6 bug?
    Re: (fractint) SnowFlakeColor
    Re: (fractint) Serious Fractint 19.6 bug?
    Re: (fractint) Serious Fractint 19.6 bug?
    Re: (fractint) SnowFlakeColor
    Re: (fractint) Howdy!
    Re: (fractint) Howdy!
    Re: (fractint) SnowFlakeColor
    Re: (fractint) Boundary is partially right!
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor
    Re: (fractint) Howdy!
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor
    Re: Re: (fractint) SnowFlakeColor
    Re: (fractint) dmj-pub.frm
    Re: (fractint) Serious Fractint 19.6 bug!
    Re: (fractint) SnowFlakeColor
    Re: (fractint) SnowFlakeColor

See the end of the digest for information on subscribing to the fractint
or fractint-digest mailing lists and on how to retrieve back issues.

----------------------------------------------------------------------

Date: Sat, 4 Oct 1997 23:09:03 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) Serious Fractint 19.6 bug?

Yo Paul,

> Of course if it hangs on startup he won't be able to reach the "common
> problems" help screen, except by going reading it or using makedoc on a
> machine where it runs ok.

Of course, but our friend is able to run Fractint on other machines
so can read the docs.

If you think I'm being silly, I hereby assign you to the mythical
"Fractint Support team" and answer the questions I get. <g!> The docs
may not be the best, but 99% of the questions I spend time answering
are already answered in  the docs. Many issues are discussed in the
"Common Problems" section.

Example: one of the most common questions I get is "why does going
to text mode and back corrupt my image". I'm about ready to make
textsafe=save the default just to avoid seeing this question. This
would mean that every time you go to text mode, your graphics screen
would be saved to disk. Might be better to make the default slower
and safer, and have a textsafe=no or some such for folks who know our
trick for leaving the graphics image intact in video memory works on
their machine.

Tim




- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sat, 4 Oct 1997 23:09:03 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Arrrgghhh! My "somewhwat technical note" that retracted my earlier
theory was itself wrong. I'm not having a good time with this Lsys
question. It turns out I have more than one sstools.ini on my system,
and my developer version was picking up float=yes but my version 19.6
wasn't!!! So all that theorizing about stack was wrong.

Jonathan Osuch pointed out to me that the Lsystems "SnowFlakeColor"
was never broken in float mode. It's broken in integer mode.

Could someone with a fast computer or a lot of patience try
SnowFlakeColor at high orders such as 7 and 8, and tell me if it's
OK with float=yes? I tested 6 and it worked with float=yes. It gets
*really* slow at higher orders.

Tim


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 00:16:44 -0400 (EDT)
From: [e-mail] (Paul Derbyshire)
Subject: Re: (fractint) Serious Fractint 19.6 bug?

textsafe=save is the most user friendly default. And, it will run
perfectly fast if it saves to extended memory instead of disk. (c'mon
NOBODY runs Fractint on machines with actually only 640K anymore! Or if
they do, then it can always use disk as a fail safe if it doesn't detect
enough ram. (Malloc returns null.)

- --
    .*.  Where feelings are concerned, answers are rarely simple [GeneDeWeese]
 -()  <  When I go to the theater, I always go straight to the "bag and mix"
    `*'  bulk candy section...because variety is the spice of life... [me]
Paul Derbyshire [e-mail], http://chat.carleton.ca/~pderbysh

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sat, 4 Oct 1997 23:20:58 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Paul wrote:

 > Glad the problem has been resolved.

See my other emabaressed  note on lsys. Actually, the situation is
good if it's what I now think. Just use float=yes and all is well.
Snowflakecolor is only broken in integer mode, which doesn't bother
me. Lots of things don't work in integer mode.

> That's what I THOUGHT you said. SYCHRONOUS ORBITS?! WOOHOO!

I have Synchronous Orbits working just fine in a long double version
in my version. But synchronous orbits is only useful for very deep
zooms right at the limit of double precision. Using long double gives
an extra three orders of magnitude, but the fun won't really start
until we port Synchronous Orbits to arbitrary precsion. I'm really
looking forward to that.

For those who don't know, the Synchronous Orbits algorithm is a huge
speedup for certain deep zoomed fractals. It works by flying a number
of orbits "in formation" and detecting when they start to get out of
formation. The algorithm then subdivides and continues with a smaller
formation, continuing until the subdivision process reaches the pixel
level. This can save an enormous amount of time, because up to the
point of subdivsion, the iterations of all the points inside the
formation have been saved. For deep zooms this can be thousands of
iterations for thousands of pixels.

It's not clear how well Synchronous orbits will work with other than
mandelbrot and Julia, but I'm planning on implementing it generally,
at least for the types supporting arbitrary precision.

Tim



>
>
> --
>     .*.  Where feelings are concerned, answers are rarely simple [GeneDeWeese]
>  -()  <  When I go to the theater, I always go straight to the "bag and mix"
>     `*'  bulk candy section...because variety is the spice of life... [me]
> Paul Derbyshire [e-mail], http://chat.carleton.ca/~pderbysh
>
> ------------------------------------------------------------
> Thanks for using Fractint, The Fractals and Fractint Discussion List
> Post Message:   [e-mail]
> Get Commands:   [e-mail] "help"
> Administrator:  [e-mail]
> Unsubscribe:    [e-mail] "unsubscribe fractint"
>
>

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sat, 4 Oct 1997 23:25:07 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) Serious Fractint 19.6 bug?

Paul suggested:
Paul wrote:

> textsafe=save is the most user friendly default. And, it will run
> perfectly fast if it saves to extended memory instead of disk.

Actually, that's a good idea. Jonathan Osuch added the needed
functions to use extended memory in just that way. That's a good
application of his functions, I'll look into it.

Tim


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 00:31:34 -0400 (EDT)
From: [e-mail] (Paul Derbyshire)
Subject: Re: (fractint) Serious Fractint 19.6 bug?

>
>Paul suggested:
>Paul wrote:
>
>> textsafe=save is the most user friendly default. And, it will run
>> perfectly fast if it saves to extended memory instead of disk.
>
>Actually, that's a good idea. Jonathan Osuch added the needed
>functions to use extended memory in just that way. That's a good
>application of his functions, I'll look into it.
>
>Tim

Don't forget disk-video modes. They use disk or ram as optimal. Probably
textsafe=save can be implemented essentially the same way. For that matter
it's a fact that RAM accesses faster than video, so normal video fractals
perhaps could compute to RAM and be updated to screen on a timer, say
every half second? The speedup might be noticeable. Faster fractals on
faster machines the bottleneck might be video writes and bank switching,
although for slower fractals or slower machines, where you see a pixel...
twenty seconds later a pixel... it won't matter a whit.


- --
    .*.  Where feelings are concerned, answers are rarely simple [GeneDeWeese]
 -()  <  When I go to the theater, I always go straight to the "bag and mix"
    `*'  bulk candy section...because variety is the spice of life... [me]
Paul Derbyshire [e-mail], http://chat.carleton.ca/~pderbysh

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun,  5 Oct 1997 13:50:37, -0500
From: [e-mail] (MR CHARLES F CROCKER)
Subject: Re: (fractint) SnowFlakeColor

Here is the result of my survey of the Lsystem SnowFlakeColor with a
Pentium 90.

Res        Math  Order
640X480     FP     5       0:34.00
640X480     INT    5       0:33.56    Few missing pieces
640X480     INT    6       4:05.36    Considerable missing
640X480     FP     6       7:11.11
1280X1024   FP     6       7:05.73
1280X1024   FP     5      34.44
1280X1024   INT    5      33.78       Missing pieces. Doesn't fit screen.
1280X1024   INT    4       3.46       Top off screen by 4 or 5 pixels.
1280X1024   INT    3       1.16       Complete.
1280X1024   FP     7    1:32:13.58

It doesn't seem that keeping integer is worthwhile. Exactly what is
happening isn't clear to me. As the order goes up the figure gets
progressively bigger missing pieces and more outside of the screen. The
plotting seems to progress normally but random pieces get skipped. If you
want to see the structure order 5 is about as high as necessary with
ordinary displays.

Changing the pallet to Rainbow and making the background black pleases me.
The order 7 figure then is then colored pink mush.

Charles

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sat, 4 Oct 1997 15:47:02 -0500
From: [e-mail]
Subject: Re: (fractint) Howdy!

Hi, Thierry (pronounced th-ear-ee?)

I guess I could put my e-mail message in the formula, but why?

As far as the zooming feature is concerned: I'm certainly no expert but
if you look at a L-system fractal carefully, you will probably be able to
predict what the "zoomed" result will look like: exactly what the normal
100% view will be!
To see what I mean, pick any L-system fractal and start with 1itereation
and look carefully at the result.
Then use a 2 iteration, then 3, then 4 and so on. I think that you will
see why a zooming feature in these types of fractals is unneccesary.

On Sat, 4 Oct 1997 15:48:16 -0400 Thierry Boudet
<[e-mail]> writes:
>        Hello Nuke.
>
>I have tested your "nice flower". I like it !
>I have never played with L-system, but I'm going now
>to try it. Only one note: may be you can put your
>e-mail in a comment in the formule ?
>
>And now _my_ question: why zoom fonctions are
>disabled in L-system ? I cant' see a technical
>reason, because I think that the complete image
>is computed in the first pass (twirlling baton)
>an displayed in the second pass. Is there a
>guru here for an explanation ?
>
>        Best regards from Toulouse, France
>
>        Thierry Boudet
>
>------------------------------------
>Fltree(with color) { ;author Nuke
>        Angle 32
>        Axiom +++++++++X
>        X=<01F[@.64+++X]-F[@.29---X]@.5X
>        }
>------------------------------------
>
>
>------------------------------------------------------------
>Thanks for using Fractint, The Fractals and Fractint Discussion List
>Post Message:   [e-mail]
>Get Commands:   [e-mail] "help"
>Administrator:  [e-mail]
>Unsubscribe:    [e-mail] "unsubscribe fractint"
>

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 14:47:36 -0400
From: Thierry Boudet <[e-mail]>
Subject: Re: (fractint) Howdy!

> I'm certainly no expert but
> if you look at a L-system fractal carefully, you will probably
> be able to predict what the "zoomed" result will look like:
> exactly what the normal 100% view will be!

        May be, may be, me to, I'm not an expert. But in
Fractint, the zoom fonctions are not only for "deep zzzzoooom",
but can rotating, panning, modifying aspect ratio and 'skewing'
your image. It work great for IFS. Why not for L-systems ?

        Think a little about it.
        Best regards.
        Thierry.


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 15:03:36 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Charles Crocker wrote:

> Here is the result of my survey of the Lsystem SnowFlakeColor with a
> Pentium 90.

Thanks for testing. I know SnowFlakecolor is broken in integer mode.
Looking at your results, all the float=yes cases were OK. All the
remains might be for some patient soul to try orders 7 and 8.

I'm not surprised, nor does it bother me, that the integer math
SnowFlakeColor is broken. Anyone with a pentium should use float=yes
in sstools.ini and make it the default. It's true that fractint made
a name for itself with integer math, but that dates back to a trime
when most PCs did not have coprocessors, and coprocessors were
slower. With each chip generation, floating point performance inches
closer to integer math.

Thanks again for your benchmarks. Silly of me to have forgotten that
SnowFlakeColor works fine with float=yes.

Tim








- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 14:28:53 +100
From: "Benno Schmid" <[e-mail]>
Subject: Re: (fractint) Boundary is partially right!

Justin wrote:
> Note:  I tried to send this directly to you, Benno, but the mail got
> returned with
> some strange error message, which follows:

I have checked  out all of my adresses as soon I returned from Verona
and they worked fine, so I can't imagine what has happened.
>
> Yes, I got the files.  I checked them out, and IMHO, they actually do
> resemble
> the border of the twindragon curve,  except that the resulting curve is
> actually
> like one straight segment of the border of the twindragon instead of
> being the
> entire closed border.

Justin,
You're right, and although I could not believe it first, Mandelbrot's
construction _is_ a part of the twin dragon border. It is in
particular the line binding the two smaller copies of the dragon
together. So the whole border is made up upon copies of the "skin"
curve. I have not found out how many, perhaps even infinite, but the
dimensions are certainly identical.
> I've tried to manipulate the L-system version to make
> it close, but it is difficult when I don't acually have the picture of
> the original twindragon in front of me. <g>

Why don't you just print it out?
>
> Anyway, if I can get it to work, I'll
> post it and anyone else who has any interest can check it out for
> themselves.
Please keep me noticed what you have found out.

Benno


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 05 Oct 1997 16:38:39 -0400
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Tim,

 - Thanks for testing. I know SnowFlakecolor is broken in integer mode.
 - Looking at your results, all the float=yes cases were OK. All the
 - remains might be for some patient soul to try orders 7 and 8.

I can run these tonight on my Pentium-166.  (Finally replaced my Cyrix with
an Intel... hoo boy does FractInt like having a fast FPU!)  Order 6 worked
fine in float mode here, too.

 - With each chip generation, floating point performance inches
 - closer to integer math.

Now there's an understatement. :)  Pentiums are faster at floating-point
than integer for anything that is multiply-intensive (like fractals).  22.3
for float vs. 26.8 for integer mode, M-set zoom near the top circle,
maxiter=1000.  As the iterations increase, so does the speed benefit for
floating-point.  Even the initial M-set image is slightly faster on the FPU.

Damien M. Jones  /  temporary sanity designs  /  http://www.emi.net/~dmj/
   [e-mail]  /  my gallery: http://www.geocities.com/SoHo/Lofts/2605/


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 15:55:10 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Damien said:

> Now there's an understatement. :)  Pentiums are faster at floating-point
> than integer for anything that is multiply-intensive (like fractals).  22.3
> for float vs. 26.8 for integer mode, M-set zoom near the top circle,
> maxiter=1000.  As the iterations increase, so does the speed benefit for
> floating-point.  Even the initial M-set image is slightly faster on the FPU.

For this I reason I have the following "modest proposal" (don't
worry, I'm not going to do this instantly, still thinking <g!>)

I want to rip out all integer code from Fractint. My reasons are

1. It is getting harder and harder to add new features to Fractint
because of resource limitations. Removing all the integer code would
greatly simplify things by suddenly reducing the size and complexity
of Fractint.

2. When we do move to a 32 bit environment, we will have to remove
integer code anyway. The integer math code is written in assembler,
and not just any assembler, but medium model assembler. This code
won't port easily to any other environment. Note that Xfract already
supports only floating point.

There are only two arguments I can think of against this.

1. There are still millions of older machines around the world
without FPUs. The most recent use the 486SX chips.

(I think that is a weak argument as time goes by.)

2. Artists have a huge number of images that depend on artifacts of
integer math.

There's no good answer to this except "keep your old copy of
Fractint".

I'd like comments on this. Robin Bussell's "Fractint Wish List" web
page has a number of pleas to "please don't kill the DOS version of
Fractint". If removing integer math extended the life of the DOS
fractint would folks be for it? Removing integer math would not
necessarily slow getting ports of Fractint for newer environments.

What do folks think?

Tim


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 16:04:41 -0500
From: [e-mail]
Subject: Re: (fractint) Howdy!

Hey Thierry,
I'm not familiar with any of the other possiblities for the zooming
features. I'll have to check it out.

Can somebody help with this one: How do you set the default file in
L=-system, and Formula types?
Plus- Is there a way to have Fractint start at a particular fractal type,
and automatically set the video, selct the file etc.. And if you tell me
to use the batch mode please explain a little to get me started.
(is batch mode in Factint like a batch file in DOS?)
On Sun, 5 Oct 1997 14:47:36 -0400 Thierry Boudet
<[e-mail]> writes:
>> I'm certainly no expert but
>> if you look at a L-system fractal carefully, you will probably
>> be able to predict what the "zoomed" result will look like:
>> exactly what the normal 100% view will be!
>
>        May be, may be, me to, I'm not an expert. But in
>Fractint, the zoom fonctions are not only for "deep zzzzoooom",
>but can rotating, panning, modifying aspect ratio and 'skewing'
>your image. It work great for IFS. Why not for L-systems ?
>
>        Think a little about it.
>        Best regards.
>        Thierry.
>
>
>------------------------------------------------------------
>Thanks for using Fractint, The Fractals and Fractint Discussion List
>Post Message:   [e-mail]
>Get Commands:   [e-mail] "help"
>Administrator:  [e-mail]
>Unsubscribe:    [e-mail] "unsubscribe fractint"
>

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 17:10:11 -0400 (EDT)
From: [e-mail] (Ian J Kaplan)
Subject: Re: (fractint) SnowFlakeColor

Tim Wegner wrote:
>
>
> I want to rip out all integer code from Fractint. My reasons are
>
> 1. It is getting harder and harder to add new features to Fractint
> because of resource limitations. Removing all the integer code would
> greatly simplify things by suddenly reducing the size and complexity
> of Fractint.
>

   Having not looked at the source I don't know how complicated this
would prove, but would it be possible and feasible to isolate out the
integer optimizations? Maybe there's an int=yes flag that redirects
things to a giant block of integer code... presumably this would give a
less tightly optimized result when running integer math, but it would let
the world's 486sx owners continue, and preserve the usefullness of old
artifact images. It's worth remembering that while few serious fractal
lovers are still stuck on a non-FPU processor in this country, that's
less true in many other parts of the world, at least judging by what I've
seen on this listserv. On the other hand, it would be relatively trivial
to keep the last integer version availible. I'd have to say that if
there's a relatively simple way to preserve integer math as a sort of
module, great, but otherwise we may as well move ahead... becuase it will
make the rest of fractint development much faster and much more portable,
really.


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 17:13:54 -0400 (EDT)
From: [e-mail] (Michael Traynor)
Subject: Re: (fractint) SnowFlakeColor

Tim Wegner wrote:

>I want to rip out all integer code from Fractint. My reasons are
>

Go for it!
...
>
>There are only two arguments I can think of against this.
>
>1. There are still millions of older machines around the world
>without FPUs. The most recent use the 486SX chips.

This makes me feel better about my 486DX. (I have a pentium 166 at work, a
government job.  I don't need it for anything I do at work, and it attracks
thieves - one of them, on my machine alone, has already been stolen - but
it sure is nice for fractals.  I need to justify the expenditure of tax
dollars on the damn thing somehow.)

>
>2. Artists have a huge number of images that depend on artifacts of
>integer math.
>
>There's no good answer to this except "keep your old copy of
>Fractint".

With the recent efforts to make sure all the old versions are archived,
this is not really a flippant answer at all.



- --
Mike Traynor

People who like this sort of thing will find this the sort of thing they like.
    Abraham Lincoln

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 05 Oct 1997 15:04:07 -0600
From: Rich Thomson <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

I say ditch the integer math and let people know that the DOS-only
version of fractint has a limited lifespan.  Perhaps one more major
release and then that's it.  If people are stuck with old machines (some
of my friends are) or don't want to use a windowing environment
(again, some of my friends are like that), then they should use the
last DOS release of fractint.  Which I would imagine would be
something like rev 20.x.  Trying to support both DOS and windowed
environments from the same code base is a problem we've discussed
often and it is just such a huge pain.  I think that 20 releases under
DOS is pretty damn impressive!  Retiring fractint from DOS would make
adding new features to the code SO much easier that people might
actually do it more often :)
- --
  ``Between stimulus and response is the will to choose.''  -- Steven Covey
 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
     3D Paint: The Power to Create in 3D;        Rich Thomson
     email me for more info                [e-mail]

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 05 Oct 1997 18:07:17 -0500
From: Justin Kolodziej <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

>Could someone with a fast computer or a lot of patience try
>SnowFlakeColor at high orders such as 7 and 8, and tell me if it's
>OK with float=yes? I tested 6 and it worked with float=yes. It gets
>*really* slow at higher orders.

I've tried order 7, and it worked (at least as far as I ran it- I
accidentally pushed a fn key as I was running it, and wiped it out most of
the way through -- GRRR! )  Anyway, I'll run order 8 overnight tonight, and
let you know the results tomorrow ( if it's finished by then on my Pentium
Pro 180 ;) )

Justin

- --no cool signature :( --

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 20:07:10 -0400 (EDT)
From: henry birdseye <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

On Sun, 5 Oct 1997, Rich Thomson wrote:

>
> I say ditch the integer math and let people know that the DOS-only
> version of fractint has a limited lifespan.  Perhaps one more major

If Fractint for Win could do batch animations then go for it. But my
experience is that DOSFract is extremely fast and I use it for batch
zooming animation all the time.

- -----------------
Henry S. Birdseye
Video Compositing Artist, Fractal Zoomer, Raytracer, Film Collector,
  Techno Head, .net addict

www.mich.com/~ozymand
www.prodcolor.com


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 05 Oct 1997 19:19:52 -0500
From: Justin Kolodziej <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

>What do folks think?

>Tim
I think as long as DOS lives, there should be a DOS version of Fractint, for
the main reason that it seems a lot easier to develop for DOS than for
Windows 95.  I could be wrong, but I have been able to produce a Mandelbrot
program for DOS that doesn't even put Fractint 2 to shame (no interactive
zooming, no speedups, but automatic 1024x768 resolution) using Visual C++
1.52, but I don't even know how to get started using Windows 3.1 with
version 1.52 or Windows 95 using 4.0!  You guys must have been super
programmers to even produce a Windows version in the first place--or is it
somehow easier with non-Visual C?
BTW, the source code for Fractint won't compile with VC++1.52.  I haven't
tried Winfract yet.  I wanted to add colors for truecolor mode, but 1) I
don't know where to find the limit for the maximum colors and 2) It wouldn't
compile anyway.

Justin

- -- no cool signature :( --


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 21:51:36 +0800
From: Francois Blais <[e-mail]>
Subject: Re: Re: (fractint) SnowFlakeColor

At 10/5/97 3:03:00 PM, Tim Wegner wrote:

>Thanks for testing. I know SnowFlakecolor is broken in integer mode.
>Looking at your results, all the float=yes cases were OK. All the
>remains might be for some patient soul to try orders 7 and 8.

On my 486 DX-50, at 640x480, (float=yes) SnowFlake took about 5 minutes at order
7, and about 20 minutes at order 8. (about 10 minutes to compute and the same
time to render)
Hope this helps.

- --
           La voix de ma contrebasse     *    Quebec City - Canada


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 20:53:46 -0500
From: [e-mail] (Evan J Hall)
Subject: Re: (fractint) dmj-pub.frm

Since my e-mail service doesn't have a file send/recieve function, I
wasn't able to use this par file for a while, but I now have a mime
decoder, and I decoded it and attempted to look at the fractals.
Unfortunately, I got errors such as "Undefined Function", "Need More ')',
and "')' needs a matching '('.  I haven't been able to fix these errors
and I'm beginning to wonder if I need a different version of Fractint.
Please help!

+
|  :)Evan Hall
|  [e-mail]
|  My other computer is a Commodore 64
+---------------------------------------+

On Wed, 17 Sep 1997 15:32:08 -0400 "Damien M. Jones" <[e-mail]>
writes:
>--=====================_874539128==_
>Content-Type: text/plain; charset="us-ascii"
>
>Howdy folks,
>
>Attached to this message is a formula file containing all the new
>coloring
>algorithms I have made for FractInt, for both the Mandelbrot set and
>the
>NovaM fractal.  Many of these are variations on orbit trap types; a
>few are
>based on formulae Kerry Mitchell posted here recently.  There are over
>100
>different formulae in this file.  Details as to what each formula does
>are
>contained in the "dmj--Read-Me-First" formula which will be the first
>one
>listed.  Although I have only provided versions of the algorithms for
>Mandelbrot and NovaM types, instructions are included on how to adapt
>the
>techniques to other types.  Each formula is extensively commented as
>well.
>
>Lee Skinner asked for example PARs to show some of what is possible.
>I
>have forty or so images produced with these formulae (and a few other
>formulae that aren't ready for "prime time") so I don't really want to
>post
>PARs for all of them here.  I hope to have most of them up on my web
>gallery later today.  Here are six, though, to show you what is
>possible
>with these techniques:
 (Files snipped)

- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 20:20:39 -0600 (MDT)
From: "Sean (and/or) Jaqueline" <[e-mail]>
Subject: Re: (fractint) Serious Fractint 19.6 bug!

On Tue, 30 Sep 1997, NOEL_GIFFIN wrote:

>     On the other hand, is it possible that when you unzipped the
> new version of fractint,  you might have used the option not to overwrite
> existing versions of certain files, or as my previous note suggested, were
> you one of the eager ones that picked up fractint from spanky in the
> first day or two before the official release announcement?

    Nope.  Until I downloaded 19.6 from Spanky, I hadn't had a copy of
Fractint on this particular machine.
    And nope.  I downloaded 19.6 (the first time) in July or August.

    But I have a theory.  If one converts the version number "19.6" to
alphabetic characters in the following manner (1=A, 2=B, 3=C...) the
program is then renamed "Fractint SF."  The reference to Science Fiction
in the version number is an obvious clue.  I hesitate to say any more,
lest I'm silenced.

- ---
* Brought to you by Sean and/or Jaq, and their 18 cats:
    * Crystal, Sputnik, Venus, Berkeley
    * Tinker, Evers, Chance,
    * Crosby, Stills, Nash,
    * Tigger, Pooh, Piglet,
    * Orion, Cursa, Spica, Polaris, and Atria.
 [And yes, we have children, too.  Can't remember their names, though.]


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 21:29:15 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Justin said:

> I think as long as DOS lives, there should be a DOS version of Fractint, for
> the main reason that it seems a lot easier to develop for DOS than for
> Windows 95.

I wasn't proposing killing the DOS version; quite the contrary, I was
proposing to extend its life by eliminating integer math, thereby
making room for growth.

Think of fractint for DOS as a turtle that grows inside a shell that
doesn't. Get's kinda tight inside the shell <g!> I'm proposing to
surgically remove half the turtle to create some room for new stuff
inside the fixed shell.

Kinda a bad metaphor, but I guess it will do <g!>


> BTW, the source code for Fractint won't compile with VC++1.52.

I've lost track of version numbers, but the newest VC++'s don't even
come with a conventional memory compiler. Fractint does compile fine
with the last conventional memory compiler Microsoft made. It won't
compile with any compiler that targets only Windows.

Tim


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

Date: Sun, 5 Oct 1997 21:29:15 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) SnowFlakeColor

Just a word folks. Remember that nontechnical messages are most
welcome on this list along with technical ones. This note is
technical, but don't be scared away if it's over your head. You're
still welcome here.

Rich said:

> I say ditch the integer math and let people know that the DOS-only
> version of fractint has a limited lifespan.
> ...
> Trying to support both DOS and windowed
> environments from the same code base is a problem we've discussed
> often and it is just such a huge pain.

Let me suggest an alternative. What is a pain is *not* supporting
DOS and Windowing environments together, but supporting the medium
memory model and flat memory models together. If the integer math
were removed from Fractint, and Fractint were ported to djgpp (the
extended DOS GNU C compiler) then the DOS version could live for as
long as DOS lives. That's because djgpp is a 32 bit flat memory model
environment. Most code that compiles under djgpp will also compile
under Linux or Unix or Win95. We could easily maintain a core of
portable code and everyone would be happy.

A djgpp-compiled Fractint would run on a 386 with plenty of memory.
It would be slower than the old Fractint until we got the floating
point assembler ported. Very possibly the assembler could work under
Linux also. It's not hard to bolt a windowing environment on top of
the DOS Fractint (Bert Tyler proved that). The problem is that
current windowing environments are 32 bit. Of course I'm assuming we
could do a major reorgaization of the underlying engine to isolate it
from the GUI.

The one sad thing about this is that most of Bert Tyler's original
contribution would be gone if we did this - the integer math and the
video. However we might be able to port the video to djgpp (maybe
Bert could even be dragged out of retirement to do this), but it
might be easier to use existing video libraries.

Tim


- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

------------------------------

End of fractint Digest V1 #23
*****************************

To subscribe to fractint Digest, send the command:

    subscribe fractint-digest

in the body of a message to "[e-mail]".  If you want to
subscribe something other than the account the mail is coming from, such
as a local redistribution list, then append that address to the
"subscribe" command; for example, to subscribe "local-fractint":

    subscribe fractint-digest [e-mail]

A non-digest (direct mail) version of this list is also available; to
subscribe to that instead, replace all instances of "fractint-digest"
in the commands above with "fractint".

Back issues are available for anonymous FTP from ftp.xmission.com, in
pub/lists/fractint/archive.  These are organized by date.